ISSN: 2252-8814
IJAAS
International Journal of
Advances in Applied Sciences
Advances in Applied Sciences (IJAAS) is a peer-reviewed and open access journal dedicated to publish significant research findings in the field of applied and theoretical sciences. The journal is designed to serve researchers, developers, professionals, graduate students and others interested in state-of-the art research activities in applied science areas, which cover topics including: chemistry, physics, materials, nanoscience and nanotechnology, mathematics, statistics, geology and earth sciences.
Editor-in-Chief: Qing Wang, National Institute of Advanced Industrial Science and Technology (AIST), Japan Co-Editor-in-Chief: Chen-Yuan Chen, National Pingtung University of Education, Taiwan, Province of China Bensafi Abd-El-Hamid, Abou Bekr Belkaid University of Tlemcen, Algeria Guangming Yao, Clarkson University, United States Habibolla Latifizadeh, Shiraz (SUTECH) University, Iran, Islamic Republic of EL Mahdi Ahmed Haroun, University of Bahri, Sudan
Published by:
Institute of Advanced Engineering and Science (IAES) Website: http://iaescore.com/journals/index.php/IJAAS/ Email: info@iaesjournal.com, IJAAS@iaesjournal.com
Information for Authors International Journal of Advances in Applied Sciences (IJAAS) is an interdisciplinary journal that publishes material on all aspects of applied and theoretical sciences. The journal encompasses a variety of topics, including chemistry, physics, materials, nanoscience and nanotechnology, mathematics, statistics, geology and earth sciences. Submission of a manuscript implies that it contains original work and has not been published or submitted for publication elsewhere. It also implies the transfer of the copyright from the author to the publisher. Authors should include permission to reproduce any previously published material.
Ethics in publishing For information on Ethics in publishing and Ethical guidelines for journal publication (including the necessity to avoid plagiarism and duplicate publication) see http://www.iaescore.com/journals/index.php/IJAAS/about/editorialPolicies#sectionPolicies
Paper Submission You must prepare and submit your papers as word document (DOC or DOCX). For more detailed instructions and IJAAS template please take a look and download at: http://www.iaescore.com/journals/index.php/IJAAS/about/submissions#onlineSubmissions The manuscript will be subjected to a full review procedure and the decision whether to accept it will be taken by the Editor based on the reviews. Manuscript must be submitted through our on-line system: http://www.iaescore.com/journals/index.php/IJAAS/ Once a manuscript has successfully been submitted via the online submission system authors may track the status of their manuscript using the online submission system.
IJAAS
International Journal of
Advances in Applied Sciences Study and optimization of a renewable system of small power generation Mohammed Salim Hadjidj, Bibi Triki Faouzi, Didi Faouzi
199-211
CP-NR Distributed Range Free Localization Algorithm in WSN Deepak Prashar, Kiran Jyoti, Dilip Kumar
212-219
Depth Estimation from Defocused Images: A Survey Jyoti B. Kulkarni, C. M. SheelaRani
220-225
Cost Allocation of Reactive Power Using Matrix Methodology in Transmission Network Gaurav Gupta, Manisha Dubey, Anoop Ayra
226-232
Development of Russian Driverless Electric Vehicle Andrey Mikhailovich Saykin, Sergey Evgenievich Buznikov, Denis Vladimirovich Endachev, Kirill vgenievich Karpukhin, Alexey Stanislavovich Terenchenko
233-239
Design of an IOT Based Online Monitoring Digital Stethoscope B. Revanth Reddy, S. Roji Marjorie, P. Ramakrishna
240-244
Analysis and Implementation of Unipolar PWM Strategies for Three Phase Cascade Multilevel Inverter Fed Induction Motor Drive Ravikumar Bhukya, P. Satish kumar
245-254
Energy and Load Aware Routing Protocol for Internet of Things S. Sankar, P. Srinivasan
255-264
Angular Symmetric Axis Constellation Model for off-line Odia Handwritten Characters Recognition Pyari Mohan Jena, Soumya Ranjan Nayak
265-272
An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid Message Encryption Over Wsns Y .M. Wazery, Mona A. S. Ali
273-285
Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM Communications Renuka Bhandari, Sangeeta Jadhav
286-297
(Continued on next page)
Responsibility of the contents rests upon the authors and not upon the publisher or editors.
IJAAS
Vol. 7
No. 3
pp. 199-308
September 2018
ISSN 2252-8814
Mobile Learning Technologies Khalil Alsaadat
298-302
Requirement Elicitation Model (REM) in the Context of Global Software Development Muhammad Yaseen, Umar Farooq
303-308
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 199~211 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp199-211
199
Study and Optimization of a Renewable System of Small Power Generation Mohammed Salim Hadjidj, Nacer-Eddine Bibi-Triki, Faouzi Didi Department de physique Unite de Recherche Materiaux et Energies Renouvelables (URMER), Tlemcen, Algerie
Article Info
ABSTRACT
Article history:
In this paper, a study was conducted on the sustainable development of solar and wind energy sources. The approach adopted is to exploit the two renewable resources by arriving to determine optimal configurations of photovoltaic and / or wind energy system with storage to provide electricity to a self-contained residential apartment located in the city of Tlemcen , in Algeria. The Tlemcen site showed a more favourable trend to use the photovoltaic system alone on the hybrid PV / wind system because of the low wind speeds of this site. The calculation method used is based on the monthly averages for ten consecutive years, data collected by the Tlemcen Zenâta weather station in order to have a better reliability analysis of an electric power generation system. In addition, the methods used in this study can be used to determine the optimal size of the most economical hybrid system that corresponds to any site in the world and for any requested load.
Received May 6, 2018 Revised May 17, 2018 Accepted May 27, 2018 Keyword: Hybrid Photovoltaic Optimization. Photovoltaic System Sizing Storage System Wind Wind System
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Mohammed Salim Hadjidj, Department de physique Unite de Recherche Materiaux et Energies Renouvelables (URMER), Tlemcen, Algerie. Email: m.salimhadjidj@gmail.com
1.
INTRODUCTION Man has always had increasing energy needs all over the world. Consequently, energy consumption is increasing as a result of economic growth and an increase in electricity consumption per inhabitant. To this end, the developing countries will be in need of energy to ensure their development although the use of fossil energy will have a negative impact on their environment thus making it necessary to use renewable energies for a healthy and sustainable energetic transition. Algeria has available one of the most important solar potential in the world. Sunshine duration in almost all the territory exceeds 2000 hours a year reaching 3900 hours in the high plateaus and the Sahara. The energy received daily on a horizontal surface of 1 m² is in the order of 5kwh in most parts of the territory, being about 17 kWh/m²/year in the north and 2650kWh/m²/year in the south of the country [1]. Algeria also has available a considerable wind energy whose potential is particularly important in the south with average speeds of 4 to 6 m/s; a priceless resource that can meet domestic needs in remote areas while the north is less windy with the exception of microclimates in the coastal region of Oran and Bedjaia and the areas of Tiaret, Biskra and Setif [1], [3]. The exploitation of the potential of renewables (photovoltaic, wind) of isolated places distant from conventional networks of power supply must stand as a priority due to the financial cost caused by the installation of a conventional electric network. Yet, one must not lose sight of the weather and topographic variation which must be taken into account. The photovoltaic energy cannot be a continuous source of energy because of its low availability in winter. Likewise, wind energy is greatly irregular in time and cannot produce energy constantly. Therefore, the separate use of these two sources presents problems with regard to energy requirements. To solve the previously mentioned problems, and to study an energy system that will supply electricity to a self-contained
Journal homepage: http://iaescore.com/online/index.php/IJAAS
200
ISSN: 2252-8814
apartment located in Tlemcen in Algeria for a reference year, a photovoltaic and / or wind energy system with storage is suggested. This system generates simultaneously energy from the absorbed solar energy and the captured wind energy that can be consumed directly or stored in batteries whose role is to ensure the continuity of use. Nevertheless, developing a cost-effective system includes design issues such as the dimensioning of the correct size of each component and the economic optimisation of the cost of the generated kWh. The purpose of this article is to show the reliability of using the electrical energy of solar and / or wind resources to meet the needs of a requested electrical load. A thorough study is used to find the best energy system configuration while presenting the results of sunray measurements, wind speeds, and energy data generated by the modules photovoltaic or wind turbines and storage that varies according to days of autonomy. The system obtained aims to supply a charge for domestic use (apartment) where consumption is about 3,260 kWh per day (September-May) and 9,150 a day (July-August).
2.
SITE DESCRIPTION AND ASSESSMENT OF SOLAR AND WIND RESOURCES AND REQUIRED CHARGE 2.1. Studied Site The selected site is located in Tlemcen in a region of Zenâta in Algeria. The characteristics of the site are displayed on the Table 1.
Table 1. Characteristics of the Zenâta Site [2] Site
Latitude
Longitude
Altitude
Albédo
Tlemcen
35.02°N
1.18°E
247 m
0.20
2.2. Assessment of the Solar and Wind Resources It is crucial to define and specify the data of local resources (solar irradiation and wind speed) of the selected region. This study is based on a daily data source of two sources, solar and wind energy, measured at a height of 10 m above ground level for 12 months and year-round for 10 years ( 2000 - 2010). Data were obtained from the Tlemcen Zenata meteorological station METAR / SYNOP [3]. a) Solar resource Figure 1 shows the monthly average of each year of the solar radiation index of the region of Tlemcen.
Figure 1. Monthly sun radiation of the tlemcen site The solar radiation in Tlemcen reaches its minimum of 2.3 kWh/m²/day in December and its maximum of 7.5 kWh/m²/day in June, and the annual average is 4.8 kWh/m² /day. b) Wind resource Figure 2 represents the monthly average of each year of wind speeds of the region of Tlemcen.
IJAAS Vol. 7, No. 3, September 2018: 199 – 211
IJAAS
ISSN: 2252-8814
201
Figure 2. Monthly wind speeds of the tlemcen site
The wind speeds are considered as relatively weak throughout the year. January being the most windy month with a wind speed of 3, 2 m/s and October the least windy of the year with a speed of 1,3 m/s. The annual wind speed average for the site in Tlemcen over the study period. c) Temperature Table (2) shows the monthly average if each year of temperature for the region of Tlemcen.
Table 2. Monthly Temperature Average Month
Jan
Feb
Mar
Apr
Mai
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Annual average
Temperature (c°)
6.27
8.23
11.6
14.8
19.9
25.4
28.9
28.0
23.1
17.6
11.7
7.65
17.0
The monthly temperature average reaches its minimum in January estimated to 6.27°c and its maximum to 28.9°c in July and the annual average is 17°c. These temperatures will not affect the operation of convertors in the photovoltaic installation or wind turbines. 2.3. Characteristics of the Selected Apartment and its Energy Balance The apartment chosen for the study is of type not connected to the conventional power distribution network and equipped with all the devices to provide comfort to the occupants. In addition, it is permanently occupied throughout the year and the domestic equipment operates under a standard voltage 220V-50Hz (mains voltage). The daily consumption is supposed constant during the nine following months (SeptemberMay) in the order of 3,260 KWh per day and another constant value during the summer season (June, July, and August) estimated at 9,150 KWh per day, the daily energy needs of the apartment by Wh/day are shown on the Table (3)[4].
Table 3. Assessment of Daily Energy Needs of Apartment Duration of use (hours) Adults’ room 11 4 Children’s room 22 5 Living room 22 6 Corridor 22 2 Bathroom 22 2 Toilets 11 1 Kitchen 11 7 Refrigerator 120 8 (winter), 12 (summer) Television LCD 72 5h30 Air conditioner 3663 0 (winter), 6 (summer) Others 100 2 Total lighting of apartment Power(w)
Lighting
Appliances
Daily consumption (Wh) 44 110 132 44 44 11 77 96 (winter), 1440 (summer) 397.26 0 (winter), 6600 (summer) 250 3260/9150
Figure 3 shows the total consumption of the apartment in one typical day in the winter season (September-May): 3,260 kWh in 24 hours. Figure 4 shows the total consumption of the apartment in one day in the summer season,(June-August): 9,150KWh in 24 hours.
Study and Optimization of a Renewable System of Small Power ....(Mohammed Salim Hadjidj)
202
ISSN: 2252-8814
Figure 3. Daily profile (winter season)
Figure 4. Daily profile (summer season)
3.
METHODS The crucial stage in the conception of a power generation system is determining its optimal size which depends essentially on the climatic data of the site and the characteristics of the parameters contained in this system. This part discusses the models used in the study to determine the optimal size of the electric power generation system that can meet the electrical needs of the apartment. 3.1. Global Incoming Solar Radiation and Energy Produced by the Photovoltaic Generator The global incoming solar radiation in a photovoltaic panel according to the HDKR (Hay, Davies, Klucher, Reindl) model [5]:
1 + cos b GT =(Gb + Gd Ai ) Rb + Gd (1 − Ai ) 2 1 − cos b 3 b 1 + f sin 2 + G ρ g 2
(1)
The energy produced by a photovoltaic panel is estimated from the values of the global irradiation on an inclined plane, the ambient temperature and the values of the photovoltaic panel manufacturer’s data. It is given by [6]:
E pv = R pv .S pv .Pf .H .N
(2)
The performance of the photovoltaic generator is represented by the following equation:
R pv = ηr {1 − γ (Tc − TSTC }
TC= Ta + Ginc (
NOCT − 20 ) 800
(3)
(4)
3.2. Distribution of wind Speeds and Energy Yielded by Wind Generator The Weibull function is used to characterise the distribution of wind frequencies during the study period and is defined as [7]:
k V f (V ) = A A
k −1
V k exp A
(5)
Wind speeds Vm can be calculated according to Weibull’s K and A parameters as indicated below:
1 Vm = AΓ 1 + k IJAAS Vol. 7, No. 3, September 2018: 199 – 211
(6)
IJAAS
ISSN: 2252-8814
203
Wind power density of a site according to Weibull’s probability density function can be expressed as follows [8]:
= P
1 3 Seol ρ A3Γ 1 + 2 k
(7)
where Seol is the area swept by the wind turbine blade (m²) and p is air density (1.225kg/m³). Once the wind power density is given, the energy yielded by the wind generator for a desired period can be calculated by [9]:
E 1 3 = ρ A3Γ 1 + .T Seol 2 k
(8)
3.3. Battery Size The storage capacity of batteries is determined according to the maximum required load (maximum monthly load) that is expressed by [10]:
Cbatt ,tot =
EL ,m ,max
ηbat .U bat .Pdd .N m
.N ja
(9)
The number of required batteries is determined according to the capacity of a battery unit Cbatt,u as in the case of the surfaces of photovoltaic generators by taking by rounding up the full ratio value.
C N batt = ENT batt ,tot Cbatt ,u
(10)
Spv and Seol are the respective surfaces of the panel and the turbine able to produce a 100% coverage of the load during the least favourable month [11]:
E S pv = max d E pv
(11)
E Seol = max d Eeol
(12)
The total energy produced by the photovoltaic modules and wind turbines which supply the whole charge is expressed by:
= Ed E pv .S pv + Eeol .Seol
(13)
By using the two renewable sources, the charge is divided into two parts. If the fraction of the charge supplied by the photovoltaic system is f, then the complement of the charge which is (1-f) must be fed by the wind system. The limit values of f correspond to pure systems. In fact, f=1 corresponds to a full utilization of the photovoltaic system and f=0 represents a full utilization of the wind system. The equations (11 , 12 ) become:
E pv .S pv = f .Ed
(14)
Eeol .Seol= (1 − f ).Ed
(15)
Study and Optimization of a Renewable System of Small Power ....(Mohammed Salim Hadjidj)
204
ISSN: 2252-8814
The calculations are based on the monthly averages of each year respectively. The surfaces of the photovoltaic panels and rotor of wind turbine are determined from the monthly average values of each year calculated from E pv and
Eeol . The full charge is marked Ed and the surfaces of the photovoltaic and rotor of
wind are expressed through the following equations: S =1 = f. S PV
Ed E pv
S =1 = (1 − f ). S EOL
(16)
Ed EEol
(17)
the retained component of the surface Si,u (i=1) for the photovoltaic component and (i=2) for the rotor of turbine component. The surface of the component unit Si,u (Spv,u = 0.34m², Seol,u = 0.56m²) The retained surface is calculated according to the following equation:
SiS,u = Cte.Si ,u
(18)
With: Cte is a whole number close to a high degree of the ratio
SiS and 'S' is stands for the scenario. S i ,u
4.
SYSTEM DESCRIPTION The system under study includes photovoltaic panels and/or wind turbines connected to the direct current bus (DC) and storage batteries. Each storage battery is connected in series to the 120V (DC) bus. A convertor connected into the alternative current bus (AC) is used to convert into alternative current the energy yielded by the photovoltaic panels, the wind turbines and the energy stored in batteries. The electricity generated by our system supplies the apartment. The energy not consumed after serving the charge is stored in the batteries. The photovoltaic system produces a direct voltage which is stored in the battery after crossing a charge controller of the photovoltaic system. The wind turbine yields an alternative current which is converted into direct current and stored in the battery. A discharge load is also connected to the battery to deviate the extra load when the battery is fully loaded. The diagram that groups together each possible component of the system is illustrated in Figure 5 [12].
Figure 5. Flow diagram of the system
The technical specifications of the main components of the system are indicated in Table 4.
IJAAS Vol. 7, No. 3, September 2018: 199 – 211
IJAAS
ISSN: 2252-8814
205
Table 4. Technical Details of the System Parameters
Values
Photovoltaic panels
Price:65.00€/panel
Nominal capacity (kW)
0,05
Panel Performance (%)
13
Voltage Mpp (V)
18
Intensity Mpp (A)
2,78
Short-circuit current (A)
3,16
Open-circuit voltage (V)
22,2
Warranty (years)
10
Size (length/width/height)(mm)
630 x 545 x 25
Weight (Kg)
4
Life time (years)
20
Wind turbine
Price: 2000.00 € /turbine
rated capacity (W)
1000
Maximum power (W)
1500
Start speed (m/s)
2
Nominal speed (m/s)
10
Stop speed (m/s)
55
Wind turbine efficiency (%)
96
Noise level (dB)
45
Warranty (years)
5
Life time (years)
25
Weight (Kg)
78
Rotor length (m)
2,8
Rotor width (m)
2
Battery
Prix: 305.00 € /battery
Nominal voltage (V)
12
Capacity (Ah)
230
Maximum voltage (V)
14,4
Starting current (A)
1150
Charge voltage (%)
10
Series-connected batteries
10
Life at 50% of discharge (cycle)
200
Size (length/width/height)(mm)
518 x 276 x 242
Weight (Kg)
56,75
Warranty (years)
1
Convertor
Prix: 1000.00 € /convertor
Maximum power (kW)
1210
Maximum voltage (V)
400
Voltage range PV, MPPT (V)
139 - 320
Max input current (A)
10
Nominal power (kW)
1000
Output current (A)
5,6
Nominal voltage range (V)
220 - 240 / 180 - 260
Frequency range network (Hz)
50 - 60
Maximum efficiency (%)
93
Size (length/width/height)(mm)
434 x 295 x 214
Weight (Kg)
22
Noise level (dB)
39
Study and Optimization of a Renewable System of Small Power ....(Mohammed Salim Hadjidj)
206
ISSN: 2252-8814
5.
RESULTS AND DISCUSSION This part shows the influence of the characteristics of solar and wind energy resources on the sizing of the size and the efficiency of an electric power generation system. After entering the necessary data in the calculation program, he will execute several simulations by modifying parameters and determining optimal solutions. The result of the simulation shows the most feasible configuration of the energy system as well as the energy produced by each source. The results of the photovoltaic energy gathered are illustrated in Figure 6.
Figure 6. Solar photovoltaic energy received in the site on an inclined plane
The Figure 6 shows the variation profile of the solar voltaic energy collected from day to day during the year. Two maximum values can be distinguished: The energy collected on the 110th day reaches 6.87kWh and 6.80kWh on the 237th day. The wind energy collected on the basis of daily wind speeds in the region of Tlemcen is shown in Figure 7.
Figure 7. Daily wind energy collected in Tlemcen
Figure 7 shows that the months of January and December are the most profitable of the year when the recovered wind energy reaches the maximum value of: 14.2 kWℎ/d. For the rest of the year, as the wind energy is very weak, this region resorts to more photovoltaic than wind electricity. The wind rose and the distribution of wind speed frequencies are evaluated all year long to determine if the wind blows in one direction with a respective intensity all over the year. The annual results are shown in Figures 8-9. The terminal values of “k” and “A” parameters are 1 and 2.42m/s respectively. Considering value “A” who is very low, which makes it very difficult to exploit the wind potential since all wind turbines start to produce from 2 m/s. The value of "K" is estimated at 1 m / s, this low wind density is not favorable to the continuous operation of the wind system for a whole day. Therefore, Tlemcen presents very weak conditions in term of wind resource.
IJAAS Vol. 7, No. 3, September 2018: 199 – 211
IJAAS
ISSN: 2252-8814
207
As regards the of the wind rose diagram shown in figure (8), it is noted that the direction of dominant winds is from the northern side of Tlemcen.
Figure 8. Wind speeds frequencies according to the Weibull distribution
Figure 9. Diagram of the wind rose at 10m height
Figure 10. Hybrid energy (photovoltaic-wind) received Figure 10 is interpreted for the following periods: The 1st period: from the 1st day until the 90th day, the load of the apartment is fixed at 3,26kWℎ/d (See Table 3). This winter period has 8 deficit days that are: (1, 3, 5, 8, 10, 12, 15 and 16) days, this gives one day of autonomy since the days are not successive, so storage in the batteries is introduced to cover the load of the apartment. The 2nd period: from the 91st day to the 274th day, the load of the apartment is fixed at 9.15 kWℎ/d. This summer period does not include deficit days, the system operates normally without recourse to batteries, the energy produced is fairly constant in this period and reaches the maximum value of 5.6 KWh / d. The third period: from the 275th day until the end of the year, the load of the apartment is set at 3.26 kWℎ/d. This winter period has other deficit days that are: (235, 237, 239 and 242) days it gives one day of autonomy. Since the load demand is constant for each period, the results show that the most unfavourable month is when the ratio between solar irradiation and wind speed is minimum. According to the results, the worst month is the month of December. 5.1. Calculating the Number of Batteries The number of days of autonomy is evaluated at one day (24 hours). The daily energy demand during this day is set at 3260 ah. The storage system will compensate the interruption of the power generation system. The electricity that comes out of the batteries does not come entirely to electrical devices: part is lost in the wires and during the conversion of current DC to AC by the converter, the amount of energy to be returned is 3.95 kWh. In order to have a longer life of the batteries, a maximum discharge depth of 50% is set, the Study and Optimization of a Renewable System of Small Power ....(Mohammed Salim Hadjidj)
208
ISSN: 2252-8814
capacity of the storage system will have to be 7.91 kWh, this leads to a quantitative number of 3 storage batteries during this deficit day. The calculation program finds the best configurations of the power generation system that will generate enough or all of the energy at the apartment. The results of these configurations are shown in the Table 5. Table 5. Optimum configurations Parameters
Optimum Configurations Configuration A
Configuration B
Photovoltaic system(KW)
1,7
1,4
Number of panels PV
34
28
Number of wind turbines
0
1
Number of batteries
3
3
Number of convertors
1
1
Total cost (€)
4125
5735
Table 5 shows the two best retained configurations. The configuration (A) contains a purely photovoltaic system with a power of 1.7 kW and consists of 34 photovoltaic panels, 3 storage batteries, one converter and no contains wind generator. The surface of the photovoltaic panels represents 12.7 m². The net present cost of this configuration is estimated at 4125 euro. The second configuration (B) contains a hybrid wind and photovoltaic system. The number of photovoltaic panels is inferior to that of the configuration (A) and has 28 panels and represents 9.6 m², generating a power of 1.4 kW. The number of storage batteries and converters is the same as that of the configuration (A). Regarding the net present cost of this configuration, it is higher and estimated at 5735 euro. 5.2. Production Result and Electric Consumption This part shows more details on the two configurations by comparing the energy produced and consumed annually of each configuration (kWh/year). Figures 11-12 show the average monthly power production of each configuration.
Figure 11. Average monthly power production of configuration 'A'
Figure 12. Average monthly power production of configuration 'B'
From Figure 11 it has been found that the production of electrical energy comes only from the photovoltaic system estimated at 3803 kWh/year. The average hourly output is less than 0.4 kW/h from November to February, unlike the other months of the year when this average is greater than or equal to 0.5 kW/h. The electrical energy consumed directly from the bus (AC) represents 1756 kWh/year, the rest of the energy produced is stored in batteries so that it can be used during the night or in the deficit period in solar radiation and represents 1606 kWh/year is 42.2% of the production of the pure photovoltaic system. Figure 12 shows that the production of electrical energy from the photovoltaic/wind hybrid system is estimated at 3822 kWh/year where 82% is generated by the photovoltaic panels and 18% of the wind system is respectively 3132 kWh/year and 690 kWh/year. The energy generated by the photovoltaic panels is much IJAAS Vol. 7, No. 3, September 2018: 199 – 211
IJAAS
ISSN: 2252-8814
209
higher than that of the wind system during all the months of the year. The average hourly production in February, October and November is less than 0.4 kW/h However, it can reach 0.5 kW/h the other months of the year. The electrical energy consumed directly from the bus (AC) represents 1755 kWh/year, the rest is stored in the batteries so that it can be used during the night or in the deficit period of the two energy sources and represents 1730 kWh/year or 45.3% of the production of the hybrid system. 5.3. Other Results of Configuration 'A' This part will discuss the behaviour of photovoltaic panels with storage during all the days of the year. The annual energy produced by the photovoltaic system per day and throughout the year is shown in Figure 13.
Figure 13. The daily energy produced by the PV during the year According to Figure 13, the production is stable during all the days of the year and this between 6:00 and 19:00. The daily output is around 0.4 kW up to 1.2 kW continuously and can even reach 1.8 kW and very rarely 2 kW during the day. The production is zero during the night and represents 0 kW. The hours of operation of the photovoltaic panels are evaluated at 4387 h/year. Figure 14 shows the daily status of charging and discharging batteries throughout the year.
Figure 14. Daily status of charge and discharge of batteries during the year
Figure 14 shows that the batteries are always full during the year except during the summer period which lasts from June until August when it can reach a discharge depth estimated from 90% to 65%. In mid-July the battery discharge will reach the maximum of 50% for a few hours at night. The energy convertor operates fully all year long. 5.4. Other Results of Configuration 'B' This part will discuss the behaviour of the photovoltaic/wind hybrid system during all the days of the year. The annual energy produced by the photovoltaic system and the wind system each day of the year is shown respectively in Figure 15 and 16.
Study and Optimization of a Renewable System of Small Power ....(Mohammed Salim Hadjidj)
210
ISSN: 2252-8814
Figure 15. Daily energy produced by PV during the year
Figure 16. Daily energy produced by the wind turbine during the year
According to Figure 15 and 16, it has been noticed that the production of the photovoltaic system is dominant during the year whereas it is reduced during the months of January, November and December relative to the results of the configuration 'A'. this phenomenon is due to the presence of the wind generator that associates with the photovoltaic system. Nevertheless, wind generation is valid essentially only in December and January. It is estimated very low between 0.24 and 0.48 kW. It reaches its maximum value of 2.23 kW during the middle of January. The hours of operation of the wind turbine are low. The batteries will have the same behaviour of the configuration (A) and the converter works fully during the year. Therefore, the configuration "A" is better than the configuration "B" taking into account the total net cost of the investment and which equals respectively 4125 euro against 5735 euro.
6.
CONCLUSION The study conducted in this article leads to deduce optimal configurations of photovoltaic system and/or wind power generation with storage in batteries, in order to provide electricity to a residential apartment autonomous and away from the conventional power grid in Tlemcen, Algeria. The methodology and optimization models used to determine the best energy production system are valid for any site in the world and especially for small power sites. The results obtained show the inefficiency of the wind energy potential in this non windy site, from which the exploitation of the only one of a pure photovoltaic system is recommended while avoiding the exploitation of a photovoltaic / wind hybrid system.
REFERENCES [1] [2] [3] [4]
[5] [6] [7]
[8]
Ministère de l’Energie et des Mines. (2007). Guide des Energies Renouvelables , Algérie. R.MAOUEDJ. (2005). Application de l’énergie photovoltaïque au pompage hydraulique sur les sites de Tlemcen et de Bouzaréah. Mémoire de Magister, Université de Tlemcen, Algérie. http://www.infoclimat.fr D.SAHEB-KOUSSA, M. BELHAMEL et K.BENFERHAT, Contribution à l’étude théorique du comportement d’un système hybride (éolien- photovoltaïque- diesel) de production d’électricité sans interruption, Afrique SCIENCE, Vol. 05 N°1, pp. 23 – 38, ISSN 1813-548X, 2009. Reindl DT, Beckman WA, Duffie JA. (1990). Evaluation of hourly tilted surface radiation models. Sol Energy, Vol. 45, n° 1, p. 9-17. M.A. EL HADIDY. (2002). Performance Evaluation of Hybrid (Wind/Solar/Diesel) Power Systems. Renewable Energy, Vol. 26, p. 401 – 413. Khahro SF, Tabbassum K, Soomro AM, Dong L, Liao X. (2014). Evaluation of wind power production prospective and Weibull parameter estimation methods for Babaurband, Sindh Pakistan. Energy Convers Manage, Vol. 78, n° 956, p. 67. Akpinar EK, Akpinar S. (2005). An assessment on seasonal analysis of wind energy characteristics and wind turbine characteristics. Energy Convers Manage, Vol. 46. N° 1848, p. 67.
IJAAS Vol. 7, No. 3, September 2018: 199 – 211
IJAAS
ISSN: 2252-8814
211
[9]
Keyhani A. Ghasemi-Varnamkhasti M. Khanali M. Abbaszadeh R. (2010). An assessment of wind energy potential as a power generation source in the capital of Iran. Tehran. Energy, Vol. 35, p. 188-201. [10] O.GERGAUD. (2002). Modélisation énergétique et optimisation économique d’un système de production éolien et photovoltaïque couplé au réseau et associé à un Accumulateur. Thèse de Doctorat en Electrotechnique, Ecole Normale Supérieure de Cachan, France. [11] C. PROTOGEROPOUIOS, B.J. BRINKWORTH and R.H. MARSHALL. (1997). Sizing and Techno Economical Optimization for Hybrid Solar Photovoltaic/Wind Power Systems with Battery Storage . Int. J. Energy Res., Vol. 21, p. 465-479,. [12] H. BELGHITRI, Modélisation, simulation et optimisation d’un système hybride éolien photovoltaïque . Mémoire de magister, Université de Tlemcen, 2011.
Study and Optimization of a Renewable System of Small Power ....(Mohammed Salim Hadjidj)
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 212~219 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp212-219
212
CP-NR Distributed Range Free Localization Algorithm in WSN Deepak Prashar1, Kiran Jyoti2 ,Dilip Kumar3 1
CSE, IKGPTU, Punjab, India Department of IT, GNDEC, Punjab, India 3 Department of Electronics, SLIET, Punjab, India 2
Article Info
ABSTRACT
Article history:
Advancements in wireless communication technology have empowered the researchers to develop large scale wireless networks with huge number of sensor nodes. In these networks localization is very active field of research. Localization is a way to determine the physical position of sensor nodes which is useful in many aspects such as to find the origin of events, routing and network coverage. Locating nodes with GPS systems is expensive, power consuming and not applicable to indoor environments. Localization in three dimensional space and accuracy of the estimated location are two factors of major concern. In this paper, a new three dimensional Distributed range-free algorithm which is known as CP-NR is proposed. This algorithm has high localization accuracy and resolved the problem of existing NR algorithm. CP-NR (Coplanar and Projected Node Reproduction) algorithm makes use of co-planarity and projection of point on plane concepts to reduce the localization error. Results have shown that CP-NR algorithm is superior to NR algorithm and comparison is done for the localization accuracy with respect to variations in range, anchor density and node density.
Received Feb 14, 2018 Revised Apr 20, 2018 Accepted May 27, 2018 Keyword: Centroid DV-Hop GPS NR WSN
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Deepak Prashar, CSE, IKGPTU, Punjab, India. Email: deepakprash@gmail.com
1.
INTRODUCTION There are several issues in WSNs [1], [2] which are required to be considered for the construction of robust and efficient network. In Wireless Sensor Networks, sensor nodes determine their location by localization. For this purpose localization algorithms are used in WSN [3], [4], [5]. Localization is a process of estimating the position and co-ordinates of wireless sensor nodes. Localization information is important due to several factors such as the recognition of gathered data, node labeling, management and concern of localized nodes in a decided region, impact of node’s density and coverage, energy draft formation, topographical routing, entity tracking, and other geographic algorithms. All these factors make localization as one of the key mechanics for the evolution and exploitation of WSN. The nodes which are conscious of their positions either from GPS or manualy configured are known as Anchor nodes. The nodes which don’t know their positions are called dumb or unknown nodes that take the help of anchor nodes to obtain their positions. Localization can be done either manualy or by using GPS. GPS is best measurement approach that obtains location directly without any further computation. But addition of GPS to all nodes in the wireless sensor network is impractical because of its high cost, high power usage and environment suppression, which makes it infeasible for indoor applications. Self-localization is a substitute of GPS. Hop count and neighborhood measurements are the techniques that make use of radio connectivity information. The area of evaluations can use either range or connectivity information based on design of area constraints. Several techniques have been suggested for working out Localization problem [6], [7], [8] but most of them consider only for 2-dimensional network. Hence, location estimation issue in three dimensions is a Journal homepage: http://iaescore.com/online/index.php/IJAAS
IJAAS
ISSN: 2252-8814
213
demanding concern area in the research community. Analysis of localization approaches concentrating on the 3-D space is a significant task to support real applications of WSNs, because the difference between localization in 2-D and 3-D space is expressive. Three anchor nodes are required to determine the location of nodes in 2D, whereas in 3D, four anchor nodes are needed. 2-D spaces cannot be exactly changed to 3-D just by addition of one more parameter. There are several issues that can be smoothly resolved in the sensor network as 2-D but are very complex in 3-D [9], [10]. Moreover, the triangulation approach that is engaged in 2-D space, is not applicable 3-D.In 3D, quadrilateration approach can be used to localize a WSN that uses distance measurements of four non-coplanar sensors to find location of an unknown node in 3D. Therefore we need a cost effective and resource efficient localization algorithm for 3D WSNs. As the sensor networks are application peculiar it is truly difficult to conclude the algorithms for localization that fits best to all the different schemes.
2.
PROBLEM IDENTIFICATION There is a need of 3D localization [11], [12] to provide accuracy and reducing positioning error [13] in harsh as well as flat terrain spaces. The three dimensional localization is much more complicated and has more the computational complexity as a result of which it is not appropriate to extend the 2D localization algorithm to a 3D algorithm directly. The research on 3-D localization is more realistic and localization algorithms in the three-dimensional space are necessary. One of the distributed range free algorithms in 3-D based on Node Reproduction [14] is taken as the point of further research with respect to this paper. It has smaller computation overhead and provides better localization than DV-Hop and Centroid Algorithm. In NR Algorithm [10] the localization space is supposed to be in the space of a cube which is having the length of edge to be 100m. The total volume of the 3D localization space is 100×100×100 cubic meter. There are 216 anchor nodes considered in this localization technique. Form each one of a cube, every eight anchor nodes having edge length 20m in the space. All the unknown nodes are deployed randomly. Area of communication of each unknown node is another important parameter. When the packets sent by anchor nodes enter into the communication range of any unknown node, packets are detected immediately and unknown node records the corresponding anchor node’s information including anchor node ID and coordinates. Then unknown node can obtain reproduced node using location information of three different anchor nodes and at the end the unknown node’s coordinates are obtained by the use of three anchors and one reproduced node. NR algorithm [14] has following steps: a. The algorithm begins by broadcasting data packets from anchors to the whole network periodically. For this time duration (T) can be set to manual. Anchor node ID, and its co-ordinates are the main constituents of the data that is broadcasted. b. Unknown nodes just listen to these packages that are sent in the time duration (T). Unknown nodes need to record number of packets they have received from different anchors. c. Check whether the time duration (T) is arrived or not. If arrived then information to be recorded, otherwise it will go on waiting. d. After recording anchor nodes information is, unknown node computes reproduced node with the help of three anchor nodes. e. Based on recorded three anchor nodes information unknown node predict a fourth anchor node’s position on the same plane with the help of three anchor nodes by formulating a square with edge length 20 m. f. Using these four nodes, the square center point is computed. Reproduced node differs from the center node by means of one different cordinate direction and the other two are in the same. g. Reproduced node is estimated by addition of half communication range to any one of three cordinate directions. h. After generation of reproduced node, unknown nodes end the localization process. Three anchors and reproduced node forms a tetrahedron. Then by using similar way of centroid algorithm, the center of these four nodes estimates position of the unknown node. NR algorithm doesn’t require sensors having any special range-determining hardware. It depends on node-anchor communication to find the location of the unknown nodes that reduces the computational overhead of the whole network. Moreover the communication overhead is reduced which prolongs the whole network’s lifetime. In this algorithm there are some problems in which future work is considered to be done: a. Position of the Reproduced Node: After determination of reproduced node, it is not sure about its plane in which it is generated. It may lie on the same plane with unknown node or in the opposite plane. If it is in opposite plane with respect to the unknown node, localization error is produced. During localization process, the reproduced node’s position is calculated randomly which is not a best solution and cause CP-NR Distributed Range Free Localization Algorithm in WSN (Deepak Prashar)
214
ISSN: 2252-8814
number of uncertainties in the network. How to resolve this issue is a future research area in respect to NR [14] which is worked upon in the proposed algorithm below. b. Communication Range: In NR algorithm, the small communication range leads to incompletion of location estimations. Because of this low value of communication range, the covered sensor area is small which pose as an obstacle in the generation of reproduced node. Whereas large communication range provides highly accurate estimations, but in this case the packets from different anchors get conflicted and this cannot produce accurate estimation.This issue is also taken in to account in the propsed approach.
3.
PROPOSED WORK CP-NR technique has tried to reduce the problem of localization error due to wrong position of reproduced node as mentioned in the above section. In this method, the localization space is a cube of volume 100×100×100 cubic meter. There are total 400 nodes out of which 216 anchor nodes and rest are unknown nodes. A cube having edge length 20m is formed using eight anchors in which randomly deployed unknown nodes exists. The communication range in this algorithm is 30m also. This algorithm generates reproduced node same as NR algorithm. Anchor nodes broadcast messages in random threshold time duration (T). Unknown nodes listen messages and check whether time duration (T) has arrived. If time duration (T) is arrived, unknown nodes record information of three anchor nodes which have maximum beacon transmission. Using this information unknown node predict a fourth one anchor node on to make a square of side 20m in the same plane. Using these four nodes, the center point of the square is computed. Reproduced node is estimated by addition of half communication range to any one of three ordinate directions of center node. But after generation of reproduced node, check on the plane in which reproduced node lies is done. The co-planarity concept [15] of four nodes determined the position of the reproduced node. For co-planarity, suppose three points A, B, C forms a plane and determines whether the fourth point D lies on the same plane. Cross product of two vectors (Vector AB × Vector AC) is normal to plane. Prove that the last vector AD is normal to this cross product. Hence the triple product should equal to zero. �� . ( ��× ��) =0 𝐴𝐴
𝐴𝐴
𝐴𝐴
(1)
If reproduced node does not lie within the same plane that of recorded three anchors nodes then need to use projection of point technique to put this node in the appropriate plane to increase the localization accuracy. And at the end three anchors and reproduced node forms a tetrahedron. Then by using similar way of centroid algorithm, the center of these four nodes estimates position of the unknown node. Projection of a point on a plane: Let π be any plane and P be a given point, not on the plane π, then P’ the foot of perpendicular on the plane, is called orthogonal projection of P on the plane π. 4.
RESULTS AND ANALYSIS In the proposed CP-NR algorithm localization error due to incorrect position of reproduced node is reduced using the concepts of co-planarity and projection of a point on the plane which determine the plane of the reproduced node with respect to recorded anchor nodes and positions it in the appropriate plane to generate more accurate location estimation results. In these algorithms the basic node for location estimation is reproduced node. How to generate it in the appropriate plane is the main concern. Generation of the reproduced node in the opposite plane produced high localization error. Here to reduce that localization error first check the existence of generated reproduced node with respect to the anchor nodes that is whether it is in the same plane of the anchor nodes or not, with the help of co-planarity concept. If it is found in some another plane with respect to anchor nodes from which it is generated, then projection of a point on the plane concept is used to place it in the same plane of anchor nodes. All the simulation work is done in the MATLAB software, a programming language that provides interactive visualization, mathematical computations and programming environment. In the first sub section 1, steps related to deployment with respect to existing NR and proposed CP-NR algorithms is executed. These steps include design of network space, density and deployment of anchor and unknown nodes, selection of anchor nodes and generation of the reproduced node for the localization process. In the second sub section 2, experimental results of localization errors in done for the both algorithms. This section shows that how the proposed algorithm is better than the existing one in the terms of localization accuracy. In the third sub section 3, the comparison among the proposed CP-NR and the existing NR algorithm on the basis of localization error is done. Then the effect of several parameters such as anchor density, node density and range parameter on the results of these algorithms is executed. IJAAS Vol. 7, No. 3, September 2018: 212 – 219
IJAAS
ISSN: 2252-8814
215
4.1. Sub Section 1 First of all nodes are deployed in 3D environment having the cube of edge length 100m and having total nodes 400, out of which 216 is the anchor’s density. The total localization space volume is 100*100*100 cubic meter in which the unknown nodes are randomly deployed. Figure 1 represents the cubical 3D localization space in which red colored 216 anchor nodes and (400- 216) black colored unknown nodes are deployed. The anchor nodes broadcast beacon information periodically which is listened and recorded by the unknown nodes lie in their range. This information is recorded in the descending order which forms the basis of anchor node selection. Figure 2 represents the selection of anchor nodes on the basis of their maximum recorded information parameter at unknown nodes in cubical 3D localization space which having 216 red colored anchor and (400-216) black colored unknown nodes. Now selection of three anchor nodes from which unknown nodes received maximum beacon information is done.
Figure 1. Localization space in NR and CP-NR algorithms
Figure 2. Selection of three anchor nodes
The generation of the reproduced node is the crucial aspect of both NR and CP-NR algorithms. It helps in the location estimation of unknown nodes. The three selected anchor nodes and this reproduced node form a tetrahedron and use centroid concept to end the localization process of these algorithms. The generation of the reproduced node makes effect on the communication range to cover the whole network as well as the density of the anchor nodes. Figure 3 represents the generation of reproduced node using four anchor nodes: three selected anchor nodes and one determined anchor node. This fourth anchor node is selected by unknown node having same plane that of three selected anchors. These anchor nodes form a square plane from which reproduced node is generated. In the Figure 4, four anchor nodes are of green color and reproduced node is of blue color.
Figure 3. Generation of Reproduced Node
CP-NR Distributed Range Free Localization Algorithm in WSN (Deepak Prashar)
216
ISSN: 2252-8814
4.1. Sub Section 2 The results of the localization errors in NR algorithm in both the scenarios when the reproduced node lie within the same plane of anchor nodes and when it lies in the opposite plane are presented here. When the reproduced lies within the same plane of selected anchors then localization error is less. But opposite to this when it is not in the same plane it produces more inaccurate position estimations of the unknown nodes. Figure 4 represents the localization error when the reproduced node lies in the accurate plane and Figure 5 represents the localization error when reproduced node lies in the opposite plane.
Figure 4. Error with reproduced node in appropriate plane
Figure 5. Error when reproduced node in the opposite plane
The problem of high localization error in the existing NR algorithm when reproduced node in some other plane has been addressed in the proposed CP-NR algorithm. After the generation of it, its position with respect to the selected anchors is examined using co-planarity concept of four anchor nodes. If it is in some other plane then it is to be brought on to the accurate plane using projection of a point on the plane concept. This concept projects the co-ordinates of reproduced node in the accurate plane. In this way CP-NR reduces the error in localization and provides accurate position estimations of the unknown nodes. Figures 5 and 6 represent the localization error in the CP-NR algorithm.
Figure 6. Localization error in CP-NR algorithm
Figure 7. Localization error in CP-NR algorithm
4.1. Sub Section 3 The CP-NR algorithm eliminates the problem of existing NR algorithm that is high localization error due to the inaccurate plane of the reproduced node. In the proposed algorithm the position of reproduced node is always correct which leads to the high localization accuracy. The appropriate plane of reproduced node is obtained whenever it lies in another plane using projection of its co-ordinates on the plane. Table1 below contains the five values of localization error estimations randomly in the both CP-NR and NR algorithms.
IJAAS Vol. 7, No. 3, September 2018: 212 – 219
IJAAS
ISSN: 2252-8814
217
Table 1. Localization Error Estimations of NR and CP-NR Algorithms Number of Observations 1 2 3 4 5
Localization error Existing NR Proposed CP-NR 1 0.0874 14 0.2552 34 0.5503 57 0.8667 63 1.3433
Figures 8 and 9 represent the comparison among both algorithms in terms of localization accuracy. This describes high localization accuracy of CP-NR algorithm as compared to existing NR. In these figures five random numbers of observations are taken to take the localization error estimations of both proposed CP-NR and existing NR algorithm. After that these estimations are mapped into percentages of localization errors with respect to number of observations. Both the graphs are giving representation of localization accuracy. CP-NR algorithm has less localization error estimations which lead to high accuracy as compared to NR algorithm as shown in the figures below:
Figure 8. Comparison of CP-NR with existing NR
Figure 9. Comparison of CP-NR with existing NR
In this section below comparision based on the impact of several parameters such as anchor density, total node density and range on the localization accuracy on both CP-NR and existed NR algorithm is done. Here three localization error estimations are taken to compute the average localization error at each range of anchor nodes. Figures 10 and 11 described that how the localization accuracy varies as count of anchor nodes are changed in the network. From the results of these computations it has been observed that NR algorithm gives accurate estimations with medium (250-350) of anchor nodes. But CP-NR produces accurate estimations at low and high value of anchor node density. Figure 12 represents the localization accuracy comparisons among both the algorithms on the basis of anchor node density.
Figure 10. Anchor Density in NR
Figure 11. Anchor Density in CP-NR
Figure 12. Comparision of NR and CP-NR
CP-NR Distributed Range Free Localization Algorithm in WSN (Deepak Prashar)
218
ISSN: 2252-8814
To study the total node density impact on the localization error in both algorithms, fixed number of anchors i.e 216 are taken. Then varied the total number of nodes which gives the number of unknown nodes (total nodes – anchor nodes) in the network. Figures 13 and 14 represents total node density impact on the localization error in the existing NR and proposed CP-NR algorithms. From these graphs it is concluded that NR algorithm performs well in high node density but the CP-NR algorithm works well in high as well as low node density scenarios. Figure 15 represents the localization accuracy comparisons among NR and CP-NR on the basis of total node density.
Figure 13. Accuracy in NR
Figure 14. Accuracy in CP-NR
Figure 15. NR and CP-NR comparision
The generation of the reproduced node makes effect on the communication range to cover the whole network. Figures 16 and 17 described the localization error variations with different range parameters. Both of them conclude that the NR and CP-NR algorithms give best estimation results at range value 30m. In these algorithms the edge length is 20m so with range value 30m the unknown nodes can listen more beacons from the anchor nodes and large localization area is covered. Both the algorithms have one common problem of the range parameter. At lower values of range less than 10m both the algorithms can’t complete localization process. The reason is that the less network area coverage which destructs the generation of reproduced node and causes loss of beacons broadcasted by anchors. When the range is higher than 30m the beacons from anchors are conflicted and area provided by selected anchor nodes for localization cannot achieve the best. To make both the algorithms optimum how to solve the problem of range parameter is of future concern. Figure 18 represents the localization comparisons among both NR and CP-NR on the basis of range parameter
Figure 16. Range Parameter in NR
Figure 17. Range Parameter in CP-NR.
IJAAS Vol. 7, No. 3, September 2018: 212 – 219
Figure 18. NR and CP-NR range comparision
IJAAS
ISSN: 2252-8814
219
5.
CONCLUSION The design of a localization algorithm in wireless sensor networks is an important task. Because of geographic correlation of perceived data, location information is used to address nodes and regions, to identify the gathered data and to increase the performance of many geographic algorithms. Accuracy in estimated location is a key driver for localization process. In this paper a new distributed localization algorithm known as CP-NR is proposed. This algorithm provides high localization accuracy and resolves the problem related to the position of reproduced node. This study describes the concepts of co-planarity and projection of a point on the plane to generate reproduced node in the appropriate plane. The use of these methods significantly improves the location estimation accuracy for unknown nodes. The simulations showed that CP-NR algorithm outperforms the NR algorithm in terms of localization accuracy. Variations in localization accuracy of CP-NR and NR algorithms on the basis of different parameters such as anchor density, node density and range is also been done. As the energy consumption of nodes is critical in WSNs, then energy efficient design of CP-NR for mobile node scenarios with immense values of range are the research directions for the future.
REFERENCES [1] [2]
[3] [4] [5] [6] [7]
[8] [9] [10] [11] [12] [13]
[14] [15]
F. Akyildiz, W. Su. Wireless sensor Networks: A survey, Computer Networks, 2005;38(4): 393-422. Lovepreet Singh.Techniques of Node Localization in Wireless Sensor Networks:A Review, International Journal of Innovative Research in Computer and Communication Engineering, ISO 3297: 2007 Certified Organization,2014;2(5):4143-4148. Pradnya Gajbhiye, Anjali Mahajan. A Survey of Architecture and Node deployment in WSN, IEEE Conference application of digital information and Web ,2008:426-430. Amita Sharma, Yogita Wadhwa. Routing and Computing in Wireless, International Journal of Advanced Research in Computer Science and Software Engineering,2013;3(1):336-339. B. R. Stojkoska, A. P. Avramova, and P. Chatzimisios. Application of wireless sensor networks for indoor temperature regulation, International Journal of Distributed Sensor Networks,2014;10:1-10. Rania Khadim, Mohammed Erritali, Abdelhakim Maaden. Rang-Free Localization Schemes for Wireless Sensor Networks, TELKOMNIKA Indonesian Journal of Electrical Engineering, 2015; 16(2):323-332. Chandirasekaran D, T. Jayabarathi. Wireless Sensor Networks Node Localization-A Performance Comparison of Shuffled Frog Leaping and Firefly Algorithm in LabVIEW, TELKOMNIKA Indonesian Journal of Electrical Engineering, 2015; 14(3):516-524. Yang Sun, Shoulin Yin, Jie Liu. Novel DV-hop Method Based on Krill Swarm Algorithm Used for Wireless Sensor Network Localization, TELKOMNIKA, 2016; 14(4): 1438-1445. Shayon Samanta , Prof. Punesh U.Tembhare. A Survey on 3d Localization in Wireless Sensor Networks, International Journal Of Computational Engineering Research, 2013;3(1):90-94. Q.Shi, H. Huo, T. Fang. A 3D node localization scheme for wireless sensor networks, IEICE Electronics Express,2009;6(3):167-172. Quan LIU, Ping REN, Zude ZHOU. Three-dimensional Accurate Positioning Algorithm based on Wireless Sensor Networks, Journal of Computers,2011;6(12). P.K. Singh. Node Localization in Wireless Sensor Networks, International Journal of Computer Science and Information Technologies, 2011; 2(6): 2568- 2572. [13] Sudha H Thimmaiah, Mahadevan G. A Range Based Localization Error Minimization Technique for Wireless Sensor Network, Indonesian Journal of Electrical Engineering and Computer Science, 2017; 7(2):395403. Xiaoming Wu, Yang Liu, Jianping Xing. Node Reproduction Based Range-free Localization Algorithm in Wireless Sensor Networks, Journal of Computes,2014; 9(5):1047-1052. Dinesh Khattar. Coplanarity of four points, The Pearson Guide To Mathematics For The Iit-Jee, 3/E, Dorling Kindersley (India) Pvt. Ltd. licensees of Pearson Education in South Asia, 2010:4-6.
CP-NR Distributed Range Free Localization Algorithm in WSN (Deepak Prashar)
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 220~225 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp220-225
220
Depth Estimation from Defocused Images: a Survey Jyoti B. Kulkarni1, C. M. Sheela Rani2, 1
Department of Computer Science and Engineering, K. L. University, Vaddeswaram (A. P.), India 2 Department of Computer Science and Engineering, K. L. University, Vaddeswaram (A. P.), India
Article Info
ABSTRACT
Article history:
An important step in 3D data generation is the generation of depth map. Depth map is a black and white image which has exactly the same size of the original captured 2D image that indicates the relative distance of each pixel from the observer to the objects in the real world. This paper presents a survey of Depth Perception from Defocused or blurs images as well as image from motion. The change of distance of the object from the camera has direct relation with the amount of blurring of object in the image. The amount of blurring will be calculated with a comparison in front of the camera directly and can be seen with the changes at gray level around the edges of objects.
Received Jun 1, 2017 Revised Feb 5, 2018 Accepted May 11, 2018 Keyword: Artificial Intelligence Classification Clustering Data Mining Intrusion Detection System Soft Computing
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Jyoti B. Kulkarni Department Of Computer Science and Engineering, K. L. University, Vaddeswaram (A. P.), India. Email: mailjyotikulkarni@gmail.com
1.
INTRODUCTION There are many evolutions, from black-and-white Television to colored Television, and HighDefinition Television is now the most popular Television in the market, from the birth of television. The development of 3D Television system has grabbed more and more attention after that. Depth estimation or extraction is a measure of the distance of, ideally, each point of the seen scene to the Stereo Vision research field. In absence of the information for absolute depth measurements such as motion, defocus, binocular disparity, the absolute distance between the observer or camera and the objects in the scene or image cannot be measured. The depth estimation is very important aspect in various applications. So, when the distance of the cameras and some camera parameters when changed, then the depth estimation is useful to calculate exact distance of camera from image in both the cases as discussed in [8]. The information of shading, edges and junctions may provide 3-D model but it will not give the exact scale of the space. There are many ways to retrieve 3D information. The most common and straightforward method is to use “active device” i.e. active cameras, which can capture the original image of the scene and detect the depth of each of pixels of the scene simultaneously. The examples are infrared cameras, sonar cameras, etc. However, most of the active devices need additional sensor to obtain depth information, so they are generally more expensive compared to common cameras. Whereas, since they have specific designed sensors, most of the times the detected depth maps are more accurate. In addition to active device, we can also use the “passive device”, which requires some post processing after the acquisition of the scene of interest, such as binocular cameras and common cameras are with focus tuning functionality. Compared to active devices, passive devices are much cheaper, and without the limitation of sensor energy issue because of the active devices, passive devices can generate a higher resolution of depth map.
Journal homepage: http://iaescore.com/online/index.php/IJAAS
IJAAS
ISSN: 2252-8814
221
In the area of passive methods, single common camera is the most preferred camera since it can be acquired from general markets and it is easy to use for most of the people. However, the lack of information from additional sensors like the active devices or different angles of view from additional camera, the binocular cameras make the calculation of depth map from a single image very difficult. To compensate this defect, a series of images of a scene can be captured on different focus planes to gather more information than just a single image alone. The result in the previous works turns out to be better than just a single image alone and many algorithms are even suitable for future hardware implementation for real time application. Also, Depth perception from stereo vision is based on the triangulation principle [9]. We can use two cameras with projective optics and arrange them side by side, such that their view fields overlap at the desired object distance. By taking a picture with each camera, we can capture the scene from two different viewpoints. The above scope considers focused images. Depth from Defocus is the challenge in perception of accurate depth as all the objects in the scene are not focused always [10]. The camera positions, light intensity and focal lengths of camera may vary which yield in blur images. So, in this paper, survey on Depth from Defocus is done from the above motivation. The remaining part of this paper is arranged as follows: Section II describes some literature survey that has done in the field of depth from images. Section III presents some experimental results of the literature. Section IV gives the conclusion.
2.
LITERATURE SURVEY In a natural image, objects on the focus plane are sharper than those out of focus due to the depth of the objects. Those works or algorithms that analyze the sharpness or blur of an object as their information of depth are categorized as “Depth From Focusing” or “Depth From Defocusing” algorithms. Either “Depth From Focusing” or “Depth From Defocusing” algorithm needs focus measure which finds out the sharpness of an object to determine its depth. There are plenty of ways to measure the sharpness, either in spatial domain or frequency domain. All of them are just different aspects of some sort of high pass filters in different point of view. Tayebeh Rajabzadeh, Abedin Vahedian, in their paper [1], were introduced a new method which used similar characteristics of defocus blur. It was found out in paper [1] that the change in object distance from the camera has direct relation with the amount of defocus blur in the image. The proposed method [1] compared to conventional defocus and other methods, was shown to be a blind method. i.e. no focus state object image is required. Another advantage of this method [1] is that it is independent from the 3-D attributes of objects or the scene. The complexity is also less compared to similar methods.. The authors Cassandra Swain, alan Peters, and kazuhiko kawamura, in paper [2], improved the accuracy of depth from defocus using Fuzzy Logic Technique. Fuzzy Logic in [1] is combined with a depth from Defocus technique to correct for uncertainty and imprecision in depth estimation. Cassandra Swain, Alan Peters, and Kazuhiko Kawamura [2] have given two inputs to fuzzy algorithm are focus quality and focal error. Focus quality is a measure of the amount of defocus in the image. Experimental results in [2] show that fuzzy logic significantly improves depth estimation compared to nonfuzzy depth from defocus. Junlan Yang, Dan Sconfeld, in [3], presented a novel method for virtual focus and object depth estimation from defocused video captured by a moving camera. They used the term virtual focus to refer to a new approach in [3] for producing in-focus image sequences by processing blurred videos captured by outof-focus cameras. The method used in paper [3] relies on the concept of Depth-from-Defocus (DFD). The authors explored several blur models which can be used to recover arbitrary transfer functions. Sangjin Kim, Eunsung Lee, Monson H. Hayes’s research work [4] uses a novel approach to depth estimation using a multiple color-filter aperture (MCA) camera and its application to multifocusing. An image acquired by the MCA camera in [4] contains spatially varying misalignment among RGB color channels, where the direction and length of the misalignment is a function of the distance of an object from the plane of focus. Therefore, if the misalignment is estimated from the MCA output image in [4], multifocusing and depth estimation become possible using a set of image processing algorithms. The MCA camera in [4] with proposed image processing algorithms enables automatic, computationally efficient multifocusing using a three-step process that involves: (i) image segmentation for classifying clusters, (ii) color shift modelbased registration and fusion, and (iii) image restoration. More specifically, an image acquired by the MCA camera is first segmented into multiple clusters, each of which has the uniform color, and then the corresponding rectangular region is generated that encloses each cluster. The MCA camera significantly enhance the visual quality of an image containing multiple objects of different distances [4]. In the paper [5] by Wided Miled, Jean-Christophe Pesquet and Michel Parent, they presented a new method for addressing robust depth estimation from a stereo pair under varying illumination Depth Estimation from Defocused Images: a Survey (Jyoti B. Kulkarni)
222
ISSN: 2252-8814
conditions. First, a spatially varying multiplicative model is developed in [5], to account for brightness changes induced between left and right views. The depth estimation problem, based on this model [5], is then formulated as a constrained optimization problem in which an appropriate convex objective function is minimized under various convex constraints modeling prior knowledge and observed information. The resulting multiconstrained optimization problem in [5] is finally solved via a parallel block iterative algorithm which offers great flexibility in the incorporation of several constraints. Experimental results on both synthetic and real stereo pairs [5] demonstrate the good performance of our method to efficiently recover depth and illumination variation fields, simultaneously. The authors in paper [5] have proposed a convex programming approach for the problem of disparity estimation in the presence of illumination variations.In a paper by C. Paramanand and A. N. Rajagopalan [6], the objective was to recover the 3-D structure of a scene from motion blur/optical defocus. In the proposed approach [6], the difference of blur between two observations is used as a cue for recovering depth, within a recursive state estimation framework. For motion blur in [6], the authors used an unblurred (focused)–blurred image pair. Since the relationship between the observation and the scale factor of the point spread function associated with the depth at a point is nonlinear, they proposed and developed a formulation of unscented Kalman filter for depth estimation. Depth estimation from a single image is a challenging problem in computer vision research [7]. By analyzing the defocus cues produced by the depth of field of lens, the information of depth can be determined. Patrick P. K. Chan, Bing-Zhong Jing, Wing W. Y. Ng, Daniel S. Yeung in their wok [7] employed reverse heat equation, which is simple and effective, for this analysis. Because the depth map is required to be smooth in many applications, a mean shift segmentation and graph cut based method is proposed in [7] to infer the depth map of the scene. The confidence of depth estimation is incorporated into the energy function of graph cut to preserve details of the depth map [7]. Experimental results [7] show that the proposed method can produce a good depth map even from a single image.
3.
EXPERIMENTAL RESULTS OF LITERATURE SURVEY The authors in paper [3] test PTF estimation with a sequence ALARM as shown in Figure 1; (a) shows the first frame and (b) shows the blurred first frame as a result of a synthetic OTF consisting of a Gaussian MTF and an arbitrary PTF. Figure 1 (c) shows the reconstruction result using Restoration from Magnitude (RFM). It is a technique [3] based on projection onto convex set (POCS), while the two convex sets are the set of space-limited functions and the set of all functions that have a Fourier transform magnitude equal to a prescribed function. Figure 1 (d) shows the reconstruction result using only proposed OTF magnitude estimation [3] and PTF is considered to be zero. Figure 1 (e) shows the reconstruction result using both proposed OTF magnitude estimation and phase estimation [3]. It can be seen that the restoration including PTF estimation performs better than the restoration without phase and restoration using RFM [3]. Figure 2 shows the two different single-aperture models.
Figure 1. Comparison of focused image reconstruction methods for video ALARM: [3] (a) original image; (b) blurred image, (c) focused image reconstruction using restoration from magnitude (RFM), (d) focused image reconstruction using blur function magnitude estimation, (e) focused image reconstruction using blur function magnitude and phase estimation [3]
IJAAS Vol. 7, No. 3, September 2018: 220 – 225
IJAAS
ISSN: 2252-8814
(a)
223
(b)
Figure 2. Two different single-aperture models [4]. (a) Aperture is aligned on the optical axis of the camera (b) Aperture is shifted away from the optical axis, which produces various convergence positions according to the distance of an object
As shown in Figure 3 (b), the single-eye range method mostly failed to estimate depth of the defocused scene background in [4]. On the other hand, the proposed depth estimation method [4] produces the best result as shown in Figure 3 (c).
(a)
(b)
(c)
Figure 3. Comparison of depth estimation of three different algorithms [4]. (a) Input image acquired by the MCA camera (b) Result of depth estimation using the single-eye range method (c) Result of the proposed depth estimation algorithm
Figure 4. Results on the Shrub stereo pair. Estimated disparity map using [5] (a) SD with affine illumination. (b) Normalized cross correlation. (c) GC with istogram transform. (d) MRF with rank transform. (e) SGM algorithm. (f) Proposed method
From Figure 4, it is noticed that local methods give noisy results and are very sensitive to illumination changes [5], while the SGM algorithm and the proposed method in [5] allow obtaining a smooth disparity map with sharp depth discontinuities. Both GC and MRF algorithms combined with a histogram and rank transform, respectively in [5], also show good performance for this stereo pair, due to the presence of large homogeneous regions. Depth Estimation from Defocused Images: a Survey (Jyoti B. Kulkarni)
224
ISSN: 2252-8814
In Figure 5, the proposed method [6] has the additional advantage that it can be applied to DFD without constraining the PSF to be Gaussian. By enforcing sparsity constraint [6], the authors also addressed the problem of depth estimation when an observation undergoes simultaneous motion and optical blur.
Figure 5. (a) Reference image (b) Simultaneously blurred observation (c) Known optical blur (d) Estimated motion kernel (e) Estimated depth by proposed method [6]
(a)
(b)
(c)
Figure 6. (a) The images of Middlebury Stereo Datasets; (b) The ground truth depth map; (c) The depth map estimated by proposed method [7]
The experimental results in the paper [7] in Figure 6 show that this depth estimation technique is reliable. In this paper [7], authors illustrate passive depth estimation method to extract depth map from a single image captured in a narrow depth of field setting. This method [7] employs reverse heat equation for pre-process and use the proposed hierarchy mean shift segmentation and graph cut with a confidence to infer the depth map. From above literature survey, it has been observed that the defocused or blur image and the Depth estimated from it, is a big challenge. But if focal length, focal error of the lens is known and the intensity of light is also known, then by various approaches, depth map can be estimated to reconstruct 3D view from that.
4.
CONCLUSION Focus quality is a measure of the amount of defocus in the image. Depth from Defocus or Depth from Motion Blur is a challenge.
REFERENCES [1] [2]
Tayebeh Rajabzadeh, Abedin Vahedian, “Static Object Depth Estimation Using Defocus Blur Levels Features,” 978-1-4244-3709-2/2010, IEEE explorer. Cassandra Swain, alan Peters, and kazuhiko kawamura, “Depth Estimation from Image Defocus using Fuzzy Logic”, 0-7803-1896-X/94, 1994, IEEE.
IJAAS Vol. 7, No. 3, September 2018: 220 – 225
IJAAS
ISSN: 2252-8814
225
[3]
Junlan Yang, Dan Sconfeld, “Virtual Focus and depth Estimation from Defocues Video sequences”, 10577149/2010, IEEE. [4] Sangjin Kim, Eunsung Lee, Monson H. Hayes “Multifocusing and Depth Estimation Using a ColorShift ModelBased Computational Camera”, IEEE Transactions on Image Processing, VOL. 21, NO. 9, SEPTEMBER 2012. [5] Wided Miled, Jean-Christophe Pesquet and Michel Parent, “A Convex Optimization Approach for Depth Estimation Under Illumination Variation”, IEEE Transactions on Image Processing, VOL. 18, NO. 4, APRIL 2009. [6] C. Paramanand and A. N. Rajagopalan, Senior Member, IEEE, “Depth From Motion and Optical Blur With an Unscented Kalman Filter”, IEEE Transactions on Image Processing, VOL. 21, NO. 5, MAY 2012. [7] Patrick P. K. Chan, Bing-Zhong Jing, Wing W. Y. Ng, Daniel S. Yeung, “Depth Estimation from a Single Image Using Defocus Cues”, Proceedings of the 2011 International Conference on Machine Learning and Cybernetics, Guilin, 10-13 July, 2011. [8] A project on “A Polynomial Based Depth Estimation from a Single Image”, Kaushik K. Tiwari. [9] “Obtaining Depth Information from Stereo Images”-whitepaper, 2012 Ensenso und IDS Imaging Development Systems GmbH. Alle Rechte vorbehalten. [10] Lauren Beck and Toshiro Kubota, “Accuracy of Visual Depth Perception in an Open Field”, Biology Department, Susquehanna University, Selinsgrove PA 17870.
Depth Estimation from Defocused Images: a Survey (Jyoti B. Kulkarni)
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 226~232 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp226-232
226
Cost Allocation of Reactive Power Using Matrix Methodology in Transmission Network Gaurav Gupta, Manisha Dubey, Anoop Ayra Department of Electrical Engineering, Maulana Azad National Institute of Technology, Bhopal, India
Article Info
ABSTRACT
Article history:
In the deregulated market environment as generation, transmission and distribution are separate entities; reactive power flow in transmission lines is a question of great importance. Due to inductive load characteristic, reactive power is inherently flowing in transmission line. Hence under restructured market this reactive power allocation is necessary. In this work authors presents a power flow tracing based allocation method for reactive power to loads. MVAr-mile method is used for allocation of reactive power cost. A sample 6 bus and IEEE 14 bus system is used for showing the feasibility of developed method.
Received Nov 23, 2017 Revised Mar 20, 2018 Accepted May 15, 2018 Keywords: Kirchhoff matrix Reactive power allocation Tracing of reactive power flow pricing of electricity
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Gaurav Gupta, Department of Electrical Engineering, Maulana Azad National Institute of Technology, Bhopal, Madhya Pradesh, India. Email: gauravmits@gmail.com
1.
INTRODUCTION A fair transmission pricing methodology should recover all the cost of the transmission system and provide profit to the transmission utility. So many methodologies have developed in past years for transmission cost allocation in transmission system. The active power production capability of generator will reduced due to more reactive power. Hence, provision of pricing of reactive power becomes an important issue to be addressed in electricity market as similar to real power pricing [1]. The voltage of the system must be controlled as an component of reactive power supports so, the reliability can achieved but more pricing options required due to unrecovered obtained cost with inclusion of capital cost under scheme proposed in [2]. Reactive power transaction depends on indefinite sources such as susceptance of line, capacitor banks rating, generator capacity, installed FACT devices capacity and so on while real power flows depends on source and direction, Reactive power flow is continuously changing due to variable system operating condition. Further transmission of reactive power does not carry over longer distances because it needed to fulfill local requirements. So while locating reactive sinks, sources of reactive power identification became a big challenge. The scheme based on proportional sharing principal [3-4] offer an effectual computational tool but that concept neither discoverable nor verifiable for loss allocation. Power factor based reactive power costing methods are in traditional use but these methods are inappropriate for the restructured power systems, because they separately charged the cost of reactive power support. In addition, the tariff in current scenario only consider local charges and consumption of reactive power calculated with respect to those variables which do not judge the complete customer’s usage [3-6]. Relative electrical distance (RED) idea for transmission charge allocation based on nodal pricing method influences the operation condition and system variable has been discussed in [7]. The majority of the above referred solutions [3-7] show that transmission usage charge also Journal homepage: http://iaescore.com/online/index.php/IJAAS
IJAAS
ISSN: 2252-8814
227
having cost of losses in transmission line as in form of their integral part so it does not required separate calculation. The non-acceptability of these methods is due to the long computational time and nonlinearity towards convergence. The Z-bus matrix and modified Y-bus matrix methods treated as circuit-based allocation methods, all the computation in these methods are based on admittance matrix to solved power flow [8], [9].The cost allocation towards line losses based on complex power injection has been addressed in [10]. The virtual flow methodology for assessment of flow of reactive power in transmission network due to different sources and particular load involvement with consideration of counter and loop flows without any difficulty has been addressed in [11]. Flow of electrical power based on tracing approach shows their importance due to its explanatory and comprehensibility in transmission network process. A method based on tracing of electrical power has been reported in [12], [13] which, having assumption that outflow and inflow on nodes are proportionally shared. This permits one to outline power flow in meshed structure. A tracing based reactive power flow is reported by Bialek with upward and downward looking principle. The upward looking principle look at the balancing of incoming flows towards the nodes and the downward looking principle look at the balancing of outgoing flows from the nodes , then compute the power spread among different loads [15].Power flow tracing methods dominate marginal participation method as there is full recovery of cost by tracing flow [16]. It also depends on the Kirchhoff current laws and easy to implements on larger power systems. Moreover it has very less volatility as compared to marginal participation methods. It also provides uniformity and fairness in charge allocation due to depends on actual usage of system [17]. The locational marginal pricing for congestion cost with FACT controller by real power rescheduling for pool based transaction has been discussed in [18]. Determination of generator contribution can be used for congestion management as proposed in [19]. In this paper a reactive power flow allocation method has been proposed. After allocation of reactive power, the total cost to be recovered from individual participant towards transaction of reactive power is also allocated to different participant. For allocation power flow tracing technique is used while for cost allocation MVAr-Mile method is used. Results are shown for 6 bus system and IEEE 14 bus system.
2. DEVELOPED METHODOLOGY 2.1. Model for Reactive Power Flow Allocation The electrical power system network consists with different component so their behaviors towards tracing of power flow become topological, so power flow by tracing theory is based on true flows in transmission system with consideration of proportional sharing principle. It handles the common issue regarding distribution of VAR (reactive power) flows in a meshed system [12-14]. To determine electricity at the nodes, the nodal power flow based on tracing which generally use implementation of the KCL (Kirchhoff’s current law). To determine the correlation in conjunction with incoming and outgoing flows the proportional sharing and nodal method is adopt. Hence this principle is similar for the validation of true power and reactive power flows. The model proposed and implemented in this paper considered network is as lossless [15], [16]. Let ln = 1. . . . . . . . e shows entire transmission line in the power system structured, Gn = 1. . . . . . . . g is entire quantity of generating units and D = 1. . . . . . d is the entire quantity of users in the structure. Again PGG = diag (PG1 , PG2 , … . , PGg ) represents generation in diagonal matrix. Thus from [16] U = K m −1 PL
(1)
UT PGG = (PG )T or PG = PGG U
(2)
PG = PGG K m −1 PL
(3)
GPM = PGG K m −1
(4)
By combining equation (1) and (2)
Obtained matrix PGG K m −1 is called generation production matrix. The generation production matrix is indicated by GPM = �t ij �, i. e., Where, R i→j = t ij R Lj
(5)
Here t ij R Lj represent the reactive flow contribution of generator situated at bus 𝑖 to the load at bus j.
Cost Allocation of Reactive Power Using Matrix Methodology in Transmission Network (Gaurav Gupta)
228
ISSN: 2252-8814
Reactive power allocated to generator placed at bus 𝑖 share the line s − b can be calculated by, RPi→s−b = t is rfs−b
(6)
To obtaining the contribution of reactive power by loads similar procedure is repeat. Where the diagonal matrix PLL = diag (PL1 , PL2 , … . . , PLd ) and EFM = PLL (K m −1 )T is the extraction factor matrix of loads to generators [16]. 2.2. Cost Allocation Model for Reactive Flows For allocation of reactive power cost following algorithm is developed. For this purpose MVAr-mile method is used. In this model, reactive power charge is allocated with respect to the reactive power base capacity of the transmission line. If the cost of the line is denoted as TCs−b (in Rs/hr) then Reactive power cost allocated to users is given by: Gi For generator 𝐺𝑖 full transmission usage cost allocation is given by, FTRCs−b G
i FTRCs−b =
RPi→s−b
rf base s−b
× TCs−b
(7)
Total transmission Reactive power cost by TRCf Gi allocated to generator 𝐺𝑖 is given by: G
TRCf Gi = ∑eln=1 FTRClni
(8)
L
h Similarly for Load Lh full transmission usage cost allocation is given by, FTUCs−b
L
h FTRCs−b =
RPj→s−b
rf base s−b
× TCs−b
(9)
Total transmission Usage cost TRCf Lh allocated to Load Lh L
TRCf Lh = ∑eln=1 FTRClnh
(10)
2.3. Partial Recovery Model Partial recovery model provide cost recovery with respect to rated reactive power capacity of transmission line. If the cost of the line is denoted as TCs−b (in Rs/hr) then reactive power cost allocated to users is given by: Gi For generator Gi , partial transmission usage cost allocation is given by PTRCs−b G
i PTRCs−b =
RPi→s−b
rfrcap s−b
× TCs−b
(11)
Total transmission reactive power cost by partial recovery model TRCp Gi allocated to generator Gi G
TRCp Gi = ∑eln=1 PTRClni
L
h Similarly for load Lh , partial transmission reactive power cost allocation is given byPTRCs−b
L
h PTRCs−b =
RPj→s−b
rfrcap s−b
× TCs−b
Total transmission reactive power cost TRCp Lh allocated to load Lh L
TRCp Lh = ∑eln=1 PTRClnh
IJAAS Vol. 7, No. 3, September 2018: 226 – 232
(12)
(13)
(14)
IJAAS
ISSN: 2252-8814
229
The mathematical formulation in Equation 4 shows the contribution of active power of generator’s to load in network. Contribution of reactive flow in line by generator can be obtained by mathematical formulation given in Equation 5.
3.
RESULTS AND ANALYSIS The presented model is implemented on standard 14 buses IEEE network and 6 bus sample network shown in Fig.1 to test their feasibility and effectiveness. The programming code is developed in MATLAB tool and results are obtained. Under MATLAB tool firstly power system components such as generator, transmission line and loads are modeled for the test system. Newton-Raphson method is used to determine power flow. The line flow limits were also checked out during the power flow. As the restructuring process in power system going on from a decade, trading of real power in power market are carried out while as a responsibility of system operator toward maintaining system security, stability and reliable operation play an important role. To achieve system operator importance, voltage and reactive power comes to the picture. In this paper, the reactive power contributions of demands have been determined using power flow tracing methods. In this perspective, the influence of total reactive power flow through the line is taken for the analysis. The proposed model is implemented on test network with 6 buses and 14 buses to show their feasibility. First of all reactive power flows are allocated to loads at normal power flow condition by using modified Kirchhoff matrices methodology given in Table I for 6 bus system. For this purpose equation 5 is used. Allocation of the cost to be recovered from individual participant toward reactive power flow through the transmission network under normal operating condition is also done. 3.1. Sample 6 Bus System The single line diagram of the sample 6 bus system is shown in Figure 1. It contains 3 generator buses and 3 load buses. The data is at 100 MVA base. Table 1 shows about line flows and cost for 6 bus system. Table 2 shows about allocated reactive power of different loads for 6 bus system. Table 3 shows about cost allocated to different loads for 6 bus system.
Table 1. Line Flows and Cost For 6 Bus System
G3 G2
L6 G1
L5
L4
Figure 1. Single line diagram of sample 6 bus system
Line 1-2 1-4 1-5 2-3 2-4 2-5 2-6 3-5 3-6 4-5 5-6 Total
Flow(pu) 0.142 0.227 0.149 0.075 0.496 0.185 0.153 0.269 0.645 0.023 0.063
Cost (Rs/hr) 223.61 206.16 310.49 254.95 111.8 316.23 211.9 286.36 101.98 447.21 316.23 2786.92
Table 2. Allocated Reactive Power of Different Loads for 6 Bus System Allocated Reactive Power to Load4 (pu) 0.0846 0.1353 0.0888 0.0212 0.1403 0.0523 0.0433 0.0000 0.0000 0.0198 0.0000
Allocated Reactive Power to Load5 (pu) 0.0771 0.1232 0.0809 0.0130 0.0860 0.0321 0.0265 0.0793 0.1902 0.0012 0.0612
Allocated Reactive Power to Load6 (pu) 0.0169 0.0270 0.0178 0.0213 0.1412 0.0527 0.0435 0.1862 0.4465 0.0000 0.0015
Cost Allocation of Reactive Power Using Matrix Methodology in Transmission Network (Gaurav Gupta)
230
ISSN: 2252-8814 Table 3. Cost Allocated to Different Loads for 6 Bus System Charge allocated to Load4(Rs/hr) 133.2212 122.8786 185.0437 72.06587 31.62407 89.39908 59.96908 0 0 384.9895 0
Charge allocated to Load5(Rs/hr) 121.4108 111.8895 168.5815 44.19133 19.38468 54.87018 36.70163 84.41765 30.07224 23.3327 307.1949
Charge allocated to Load6(Rs/hr) 26.61274 24.52123 37.09209 72.4058 31.82694 90.08282 60.24608 198.2165 70.59546 0 7.529286
3.2. IEE 14 Bus System The single line diagram of the IEEE 14 bus system is shown in Figure 2. It contains 2 generator buses and 12 load buses. The data is at 100 MVA base. Table 4 shows about line flows and cost for 14 bus system. Table 5 shows about allocated reactive power of different loads for 14 bus system. Table 6 shows about cost allocated to different loads for 14 bus system.
Table 4. Line Flows and Cost for 14 Bus System
Load
L13
G1
L14
L12
L11 L6
L3
S.No. 1 2 3 4 5 6 7 8 9 10
L10 L9
L8
L7
L5
G2
Generator
L4
Flow (MVAR) 63.774 39.797 28.848 19.862 17.173 0.359 8.641 15.371 10.438 33.236
Cost (Rs/hr) 62.26 229.49 203.47 185.65 182.97 183.69 44.18 209.12 556.18 252.02
S.No. 11 12 13 14 15 16 17 18 19 20
Flow (MVAR) 6.4 2.918 8.821 0 13.065 1.531 1.961 4.378 1.111 3.528
Cost (Rs/hr) 220.41 283.81 146.1 176.15 110.01 90.29 298.77 208.86 297.92 387.73
Figure 2. Single line diagram of IEE 14 bus system
Table 5. Allocated Reactive Power of Different Loads for 14 Bus System L3 24.0395 15.0014 14.7957 10.1869 8.8078 0.359 1.7808 3.1677 2.1511 3.6432 0 0 0 0 0 0 0 0 0 0
L4 12.8888 8.043 5.5619 3.8294 3.3109 0 3.5804 6.369 4.325 7.325 0 0 0 0 0 0 0 0 0 0
L5 2.185 1.3635 0.4933 0.3397 0.2937 0 0 0 0 2.2603 0 0 0 0 0 0 0 0 0 0
L6 3.22 2.0094 0.727 0.5005 0.4328 0 0 0 0 3.331 1.6032 0.731 2.2097 0 0 0 0 0 0 0
L7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
L8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
IJAAS Vol. 7, No. 3, September 2018: 226 – 232
L9 7.9544 4.9638 3.4325 2.3633 2.0434 0 2.2097 3.9307 2.6692 4.5207 0 0 0 0 8.8022 1.0315 1.3212 0 0 0
L10 2.526 1.5763 0.8517 0.5864 0.507 0 0.3799 0.6758 0.4589 1.9757 0.5768 0.263 0.795 0 1.5133 0.1773 0.2271 4.378 0 0
L11 1.0205 0.6368 0.2304 0.1586 0.1372 0 0 0 0 1.0557 0.5081 0.2317 0.7003 0 0 0 0 0 0 0
L12 1.776 1.1083 0.401 0.2761 0.2387 0 0 0 0 1.8373 0.8843 0.4032 1.2188 0 0 0 0 0 0.8653 0
L13 3.9481 2.4637 0.8914 0.6137 0.5306 0 0 0 0 4.0843 1.9657 0.8963 2.7093 0 0 0 0 0 0.1708 2.4527
L14 4.2157 2.6308 1.4631 1.0073 0.8709 0 0.6902 1.2278 0.8338 3.2028 0.8619 0.393 1.1879 0 2.7496 0.3222 0.4127 0 0.0749 1.0753
IJAAS
ISSN: 2252-8814
231
Table 6. Cost Allocated to Different Loads for 14 Bus System L3 23.468 86.505 104.356 95.216 93.842 183.69 9.104 43.096 114.619 27.625 0 0 0 0 0 0
L4 12.582 46.380 39.229 35.793 35.27603 639 0 18.305 86.649 230.453 55.543 0 0 0 0 0 0
L5 2.133 7.862 3.479 3.175 3.12923 1293 0 0 0 0 17.139 0 0 0 0 0 0
L6 3.143 11.587 5.127 4.678 4.611274 442 0 0 0 0 25.258 55.212 71.098 36.598 0 0 0
L7 0 0 0 0
L8 0 0 0 0
0
0
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
L9 7.765 28.623 24.210 22.089 21.7714 3761 0 11.297 53.476 142.226 34.279 0 0 0 0 74.116 60.832
L10 2.466 9.089 6.007 5.481 5.40183 9516 0 1.942 9.194 24.452 14.981 19.864 25.579 13.167 0 12.742 10.456
L11 0.996 3.672 1.625 1.482 1.4617 99569 0 0 0 0 8.005 17.498 22.535 11.598 0 0 0
L12 1.733 6.391 2.828 2.580 2.54323 2924 0 0 0 0 13.931 30.454 39.215 20.186 0 0 0
L13 3.854 14.206 6.287 5.736 5.65328 6089 0 0 0 0 30.970 67.696 87.175 44.873 0 0 0
0
0
0
0
0
0
201.292
34.600
0
0
0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
208.86 0 0
0 0 0
0 232.034 0
0 45.800 269.553
L14 4.115 15.170 10.319 9.415 9.27901 7819 0 3.528 16.704 44.4283 24.286 29.683 38.223 19.674 0 23.152 19.001 62.8772 9679 0 20.084 118.176
In Opportunity cost method the total cost of reactive power including capacitor cost is 7.31 $/hr or 467.84 Rs/hr and in case of Triangle method total cost of reactive power including capacitor cost is 267 $/hr or 16020 Rs/hr reported in [20] while the total cost allocated obtained from proposed model as shown in Table VI for IEEE 14 bus system is 69.21$/hr or 4152.94 Rs/ hr which is more acceptable to attract the investor in deregulated power market as compared to Opportunity cost method and Triangle method. 4.
CONCLUSION The main objective of the model proposed in this paper is to allocate reactive power for each load based upon the proportion of reactive power flow through the transmission line as per demand by the load. The reactive power flow tracing is done by constructing reactive power flow matrix. The main reason behind the reactive power flow is the inductive loading at the load end; hence by using MVAr-Mile method the cost of this reactive power flowing is allocated to loads. Power system network is very large so this need to have additional information regarding reactive power injected by different sources as well as shunt admittance of the transmission line, the tracing of power flow scheme becomes as effective tool to achieve that. For reliable and stable operation of power system the reactive power economics play a vital role. By allocating the reactive power flow cost by proposed model, total embedded cost associated with the transmission line can be recovered and the size of reactive power sources installations such as capacitor bank, SVC and FACTS devices can be easily done.
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8]
Abdorreza Rabiee, Haidor Alishayanfer, Nibamjady Iee, Power And Enginering Magazine pp.18-32 January/February 2009. Peter W. Sauer. What is Reactive power? Power systems Engineering Research Center, Department of Electrical and Computer Engineering , university of Illinois at Urbana- Champaign, September 16,2003. Harinder Pal Singh, Yadwinder Singh Brar, D. P. Kothari, “Combined Active and Reactive Power Dispatch Using Particle Swarm Optimization”, Proceedings of Informing Science & IT Education Conference (InSITE) 2014. Wood, A.J., and Wollenberg, B.F., “ Power generation, operation, and control ”, (Wiley, New York, 1996, 2nd Edn.). Ferdinand Gubina, David Grgic, Ivo Bani, “ Method for Determining the Generators Share in a Consumer Load ”, IEEE Transactions on Power Systems, vol. 15, no. 4, November 2000. Y. Dai, X.D. Liu, Y.X. Ni, F.S. Wen, Z.X. Han, C.M. Shen, Felix F. Wu, “ A cost allocation method for reactive power service based on power flow tracing ” , Electric Power Systems Research, vol.64 , pp.59-65, 2003. D. Thukaram C. Vyjayanthi, “Relative electrical distance concept for evaluation of network reactive power and loss contributions in a deregulated system:, IET Gener. Transm. Distrib., vol. 3, no.11, pp. 1000–1019, 2009. Seyed Mohammad Hossein Nabavi, Somayeh Hajforoosh, Sajad Hajforosh, Nazanin Alsadat Hosseinipoor, “ Using Tracing Method for Calculation and Allocation of Reactive Power Cost ”, International Journal of Computer Applications , vol. 13, no. 2, pp.14-17, 2011.
Cost Allocation of Reactive Power Using Matrix Methodology in Transmission Network (Gaurav Gupta)
232 [9]
[10] [11] [12] [13] [14] [15] [16] [17] [18]
[19] [20]
ISSN: 2252-8814
Hussain Shareef, Mohd. Wazir Mustafa, Saifulnizam Abd Khalid, Azhar Khairuddin, Akhtar Kalam, Amanullah Maung Than Oo, “Real and Reactive Power Transfer Allocation Utilizing Modified Nodal Equations” International Journal of Emerging Electric Power Systems, vol. 9, no. 6, article 4, 2008 Hermagasantos Zein, Erwin Dermawan “Cost Allocation of Transmission Losses in Electric Market Mechanism” TELKOMNIKA, vol.10, no.2, , pp. 211-218, June 2012. Hogan, W , “ Markets in Real Electric Networks Require Reactive Prices ”, The Energy Journal, vol.14, no.3 pp.171-200, 1993. J. W. Lamont and J. Fu, “ Cost analysis of reactive power support ”, IEEE Transactions on Power Systems, vol. 14, no. 3, pp. 890–898, Aug. 1999. Baughman, M. L. and Siddiqi , S. N. , "Real Time Pricing of Reactive Power: Theory and Case Study Results", IEEE Trans. on Power Systems,vol. 6, no.1, pp. 23-29, 1991. Schweppe, F. C., M. C. Caramanis, R. D. Tabors, and R. E. Bohn , Spot Pricing of Electricity, Kluwer Academic Publishers, Boston. 1998 Bialek J.W., Kattuman P.A.,“ Proportional sharing assumption in tracing methodology”, IEE Proceedings Generation transmission Distribution, vol.151, no.4, pp.526-532, July 2004. Kaigui Xie, Jiaqi Zhou, Wenyuan Li “ Analytical model and algorithm for tracing active power flow based on extended incidence matrix ”, Electric Power Systems Research, vol.79, pp. 399–405, 2009 Enshaee A., Enshaee P., “ New reactive power flow tracing and loss allocation algorithm for grids using matrix calculation ” Electric Power and Energy System, vol.87,pp.89-98,2017 S.K. Gupta, R.Bansal, Partibha Sharma, Mukesh Saini “Power Trading and Congestion Management through Real Power Rescheduling Using Unified Power Flow Controller”, Indonesian Journal of Electrical Engineering and Informatics, vol. 2, no. 4, pp. 151-160, December 2014. Sawan Sen, Priyanka Roy, Abhijit Chakrabarti, Samarjit Sengupta “ Generator Contribution Based Congestion Management using Multiobjective Genetic Algorithm ”, TELKOMNIKA, vol.9, no.1, pp. 1-8, April 2011 Abbas Ketabi , Ahmad Alibabaee , R. Feuillet “ Application of the ant colony search algorithm to reactive power pricing in an open electricity market ”, Electrical Power and Energy Systems, vol.32, pp.622-628, 2010.
IJAAS Vol. 7, No. 3, September 2018: 226 – 232
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 233~239 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp233-239
233
Development of Russian Driverless Electric Vehicle Andrey Mikhailovich Saykin, Sergey Evgenievich Buznikov, Denis Vladimirovich Endachev, Kirill Evgenievich Karpukhin, Alexey Stanislavovich Terenchenko Central Scientific Research Automobile and Automotive Institute "NAMI", ul. Avtomotornaya 2, Moscow, 125438, Russia
Article Info
ABSTRACT
Article history:
This article overviews the history of development of driverless vehicles both in Russia and the World. Foreign experience of development of driverless vehicles, including electric traction, is analyzed. Main stages of creation of experimental NAMI driverless electric vehicle are revised. Main engineering solutions are described concerning development of advanced NAMI driverless electric vehicle, its major components and control systems. Projects aimed at environmental safety of passengers in NAMI driverless electric vehicle are exemplified. Results of bench scale and running tests of NAMI driverless electric vehicle are summarized. Major advantages of driverless energy efficient and environmentally clean transport are demonstrated.
Received Jun 1, 2017 Revised Mar 20, 2018 Accepted May 27, 2018 Keyword: Air purification system Control system Driverless vehicle Electric vehicle Energy efficiency Energy storage system Navigation system Vehicle Vision system
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Andrey Mikhailovich Saykin, Central Scientific Research Automobile and Automotive Institute "NAMI" ul. Avtomotornaya 2, Moscow, 125438, Russia. Email: saykin.a.m@mail.ru
1.
INTRODUCTION Recently on the roads of various countries of the World driverless vehicles begin to appear, they are controlled by complicated automatic pilot systems. The most known participant in this market is Google Inc. company. Not only this company attempts to eliminate the necessity of a human at the wheel. All major car manufacturers in Japan, Germany, US, Italy, China, and France actively develop driverless (autonomous) vehicles and relevant technologies. Numerous undisclosed projects are developed under orders from defense industry, hence, the results of such works are unavailable in public sources. Complex knowledge intensive engineering solutions, software, sensors of control systems are considered as double-purpose products [1-3]. According to forecasts of analytical agencies wide scale application of driverless (autonomous) vehicles is expected unless and until 2025.
2.
RESEARCH METHODS This study was based on system analysis, including decomposition and optimization of engineering solutions. In addition, path planning, motion of vehicle in space (environment) were used. Artificial intelligence, theory of sets and mathematical logic were applied.
Journal homepage: http://iaescore.com/online/index.php/IJAAS
234
ISSN: 2252-8814
3.
MAIN PART Active development of driverless vehicles by leading foreign car manufacturers started in early 1980-s for passenger cars, freight cars, agricultural machines and military purpose vehicles, interfactory vehicles, transport traffic in modern logistic centers and storage sites. Numerous leading car manufacturers, such as General Motors, Ford, Mercedes Benz, Volkswagen, Audi, BMW, Volvo, are involved in development of driverless vehicles as well as large manufacturers of components for these driverless vehicles, such as Continental, Bosch, Delphi. Activities on development of driverless vehicles are also performed in Russian Federation. Leading R^D institutes, universities and car manufacturers, such as NAMI, Moscow Automobile and Road Construction State Technical University (MADI), Moscow polytechnic university, PAO KamAZ and others. In the years 2014-2016 State research center NAMI in Russian Federation developed, manufactured and tested experimental electric driverless vehicle (Figure 1).
Figure 1. NAMI driverless electric vehicle (external view)
Table 1. Main Specifications of NAMI Driverless Electric Vehicle Vehicle model Traction module Nominal power (kW) Maximum power (kW) Recovery power (kW) Automotive battery of traction module Power intensity (kW*h) Nominal voltage (V) Maximum discharge current (A) Maximum charge current (A) Steering system Type Braking system Front brakes Rear brakes Suspension Front suspension type Rear suspension type Car body Length Width Height Front track, mm Rear track, mm Clearance Curb weight Operating specifications Acceleration time to 100 km/h Maximum speed Complete charge time (h) Accelerated charge time (h) Operating temperature (°C)
IJAAS Vol. 7, No. 3, September 2018: 233 – 239
LADA-Kalina 1117 30 70 30 13 316 250 100 with electromechanical amplifier Disc-type Drum-type Samper strut Trailing arm 4040 mm 1700 mm 1500 mm 1430 mm 1410 mm 160 mm 1200 20 s 120 km/h 6 1.5 - 30 … +50
IJAAS
ISSN: 2252-8814
235
Actuating mechanisms of the NAMI driverless electric vehicle are as follows: linear actuators; servo drive unit; steering electric amplifier. Instruments of the driverless electric vehicle are as follows: sensor of linear motions of actuators; linear potentiometer. The following components are used in the machine vision system of motion control of driverless electric vehicle: AXIS PI346 video camera; SLR 1 lidar; ultrasound sensors. Driverless electric vehicle operates in three control modes: 1. "Driver" mode In this mode the driverless electric vehicle is used as common electric vehicle without any restrictions. 2. "Remote control" mode In this mode driverless electric vehicle can be controlled by PC with specialized software or by dedicated joystick. 3. "Autonomous motion" mode This mode is driverless control of electric vehicle by means of machine vision system and actuators. Driverless vehicle can either move along corridor passing by obstacles; or reproduce preset trajectory; or perform parallel parking. Propulsion unit in the NAMI driverless electric vehicle is a Solectria Azure Dynamics 70 kW threephase asynchronous engine with air cooling equipped with a DMOC 445 invertor. The torque from traction electric motor to vehicle drive wheels is transferred by means of reducer with differential. In order to avoid overloads capable to break the reducer the reducer input shaft is connected with the shaft of traction electric motor by overload coupling. The reducer shell is rigidly connected with the shell of traction electric motor. The traction electric motor with the reducer in vehicle motor compartment is mounted using three-point mounting system of power unit. Electricity accumulation system is comprised of eight modules, each contains twelve AMP20M1HD-A batteries - 96 battery cells in total. Nominal voltage: 316.8 V, internal impedance: 57.6 µΩ. Control system of battery cell is located in the top part. It is comprised of four boards, each board controls 12 cells. The control system is intended for management and balancing of battery cells, for checking of integrity of battery circuits, voltage and temperature of cells, and for data transfer to battery management system (BMS). Special software provides connection, adjustment, and diagnostics of battery management system from PC via COM port (USART1) or САN. Environmental safety of passengers in the NAMI driverless electric vehicle is provided by two air cleaning systems in passenger compartment. They are located in luggage compartment (Figure 2). Clean air from the systems is supplied via ducts located inside the passenger compartment near front seats. In addition, gas analyzer for monitoring of impurity content in the air is placed in the compartment. Controller of air cleaning system is connected to the gas analyzer and the system itself. The latter operates in recirculation mode, that is, it intakes air from the passenger compartment, cleans the air and supplies the cleaned air to the passenger compartment. The gas analyzer transfers the data on air composition to the controller which constantly analyzes content of harmful substance in the passenger compartment and controls operation modes of air cleaning system [4-6]. External view of the NAMI driverless electric vehicle with installed battery cells of traction modules installed in luggage compartment and air cleaning and distribution systems is illustrated in Figure 2.
Figure 2. NAMI driverless electric vehicle with automotive batteries of traction module installed in luggage compartment (1) and air cleaning and distribution system in passenger compartment (2)
Autonomous motion of driverless electric vehicle is provided by motion control system including machine vision system, navigation and orientation system, information management system, and actuators. Development of Russian Driverless Electric Vehicle (Andrey Mikhailovich Saykin)
236
ISSN: 2252-8814
The machine vision system generates the image of surroundings, provides recognition of other vehicles, road obstructions, road signs, traffic lights, road marking [7-10]. The information management system is the core of motion management system of NAMI driverless electric vehicle, which is responsible for route arrangement, urgent decisions depending on surrounding traffic, and generation of control signals to actuators. In accordance with the control signals the actuators directly act upon units and systems of NAMI driverless electric vehicle, substituting driver actions. The navigation and orientation system determines the current position of NAMI driverless electric vehicle. The actuator system is comprised of control actuators of electric transmission, clutching, gear shifting, steering, brakes, starter, light and sound equipment, and others. The motion management system of NAMI driverless electric vehicle is comprised of sensors, drives, electronic control unit with components and control board. The machine vision system is comprised of AXIS PI346 video cameras, SLR 100 lidar, ARS 300 radar, eight ultrasound sensors (parktronics), SS series, LTP linear potentiometer and two SANV-10 electromechanic actuators. Two AXIS PI346 cameras are installed behind the front windscreen in upper corners, and one camera behind rear windshield in its center (Fig. 4), which provides maximum visibility in front of and behind the vehicle. Application of AXIS PI346 video cameras in the machine vision system of NAMI driverless electric vehicle provides accurate recognition of road signs and marks [11-13]. Figure 3 illustrates the video cameras and lidars installed in front part of the compartment of NAMI driverless electric vehicle. Figure 4 illustrates the video camera installed in rear part of the compartment of NAMI driverless electric vehicle. Figure 5 illustrates the radar installed in the front part of NAMI driverless electric vehicle.
Figure 3. Location of video cameras in passenger compartment behind windscreen of driverless electric vehicle
Figure 4. Illustrates the video camera installed in rear part of the compartment of nami driverless electric vehicle
Figure 4. Location of video cameras in passenger compartment behind rear windshield of nami driverless electric vehicle
Figure 5. ARS 300 radar installed on radiator grille of nami driverless electric vehicle
A PS-A-4006 ultrasound sensor facilitates detection of objects and obstructions in short-range zone, automatic parking and provides sensor feedback for accurate vehicle steering in narrow road corridors. The steering system includes LTP linear potentiometer. The LTP linear potentiometer installed on steering rack in engine room facilitates high accurate detection of wheel positions at any instant.
IJAAS Vol. 7, No. 3, September 2018: 233 – 239
IJAAS
ISSN: 2252-8814
237
The major component of the system is the central processing unit which is a multiprocessor perforimng the following main functions: 1. Connection with external wireless devices (BLUETOOTH, GPS, and other); 2. Acquiring image of surrounding near NAMI driverless electric vehicle (radars, lidars, video cameras, parktronics) and its subsequent processing; 3. Acquiring information about short range surrounding from ultrasound sensors and its subsequent processing; 4. Data processing from all sensors of the system; 5. Calculation of motion parameters of NAMI driverless electric vehicle in real time in accordance with the preset algorithms; 6. Generation of control commands to external actuators in accordance with current state and motion algorithm. Integration of the control system into vehicle is provided by another device: Mechanism control unit. Its main functions are as follows: 1. Processing of position sensors of control mechanisms of NAMI driverless electric vehicle; 2. Control actions for actuators of NAMI driverless electric vehicle; 3. Data transfer concerning state of control mechanisms of NAMI driverless electric vehicle; 4. Obtaining of control commands from central processing unit. Mechanism control unit provides integration with the following mechanisms and units of NAMI driverless electric vehicle: 1. Actuators of accelerator pedal, primary and parking braking systems, clutch; 2. Steering drive; 3. Traction invertor; 4. Lighting control (position lamps, low and high beams); 5. Wiper and washer control; 6. Window raiser control. Data exchange between the devices is performed by CAN bus, it provides high transfer reliability as well as interference protection in electrical system. The system is powered by battery cells of NAMI driverless electric vehicle. Electromechanic drive in the braking system controls application of brake pedal and hand parking brake. SANV-10 actuator enables fast application of primary brake and parking brake, hence, fast vehicle stoppage. Two SANV-10 actuators are used in the braking system. One of them is installed in engine room and the other in tunnel assembly. Motion control system performs general computations of vehicle motion: position coordinates, recognition of objects around the vehicle, detection of obstructions. Central processing unit is connected to the system of orientation and navigation in order to process and obtain global coordinates of vehicle position, it communicates with external devices via Bluetooth and is equipped with data storage unit. Image recognition unit is connected to cameras which inspect surrounding objects. High rate of image recognition is supported by parallel computations. Obstructions are detected by means of radar control unit and two lidar units. All units are connected to common CAN bus, which enables controlling actions and data acquisition. The software of motion control system of experimental NAMI driverless electric vehicle executes main functions of motion control, including maintaining of safe speed, safe distance to vehicle ahead, motion path with regard to preset one, control of turn indicators, audio and light stop alarms, as well as solution of navigation task reserved by satellite system. The software is based on Inka-NAMI proprietary algorithmic and programmed approaches aimed at preventing of vehicle collision with constrictions. Receiver of the navigation system is installed in passenger compartment. The antenna output connector is in accordance with SMA standard. The antenna is installed in the roof center. Efficient operation of antennas is supported by the important condition: shadowing of GLONASS/GPS satellites by at least 140°. Antenna is fed by high frequency coaxial cable. Powering of the navigation module from 12 V vehicle network is provided by a source of secondary supply (DC-DC converter) on the basis of DA1 TES 2N-2410 diode microcircuit with protection against reverse polarity and conductive noises. The navigation system can be powered from a battery. The interconnection between the receiver and network equipment is provided by ADM 1485 and 74VHC 04 microcircuits. A user of navigation data can directly connect to full-scale NV08C-CSM-BRD navigation receiver. All components of control system of NAMI driverless electric vehicle, except for navigation antenna, are installed in passenger compartment and protected against external impacts and mechanical damages.
Development of Russian Driverless Electric Vehicle (Andrey Mikhailovich Saykin)
238
ISSN: 2252-8814
Figure 6. Bench Scale NAMI Driverless Electric Vehicle
Fig. 7. Driving Tests of NAMI Driverless Electric Vehicle
The following performances were achieved after tests of NAMI driverless electric vehicle: onecharge driving distance - at least 50 km, maximum driving speed - at least 120 km/h, acceleration time to 100 km/h – at least 20 s, and the motion control system maintains maximum driving speed of 30 km/h; recognition of obstructions - road users; recognition of road signs - at least 30 ones; detection of road marks main, on straight ways; detection of vehicle position - with the accuracy to 5 m; maintaining of preset distance to vehicles ahead - to 3 m; maintaining of driving speed - with the accuracy to 3 km/h; audio and light alarms - with response rate up to 0.5 s. The air cleaning system eliminates harmful inclusions in the passenger compartment of electric vehicle to the maximum permitted level established by Hygienic regulations GN 2.1.6.1338.
4.
RESULTS AND DISCUSSION The obtained experimental results are aimed at development of R&D and engineering complex of Russian Federation in the fields of safety, energy efficiency, and energy saving in transport industry, at development projects aimed at improvement of traffic safety and environmental safety of passengers, at implementation into educational process of technical higher schools, at implementation into business activities of automobile companies [14-17]. Implementation of the experimental results would enable achievement of cardinally new level of energy efficiency, environmental and traffic safety of single vehicles and overall traffic system, including: 1. Improvement of traffic safety due to decrease in traffic accidents by means of application of driverless vehicle control and air cleaning in passenger compartments; 2. Improvement of environmental performances of vehicles in comparison with existing ones due to application of electric traction and driverless control; 3. Reduction of harmful impact of vehicles on environment due to elimination of harmful pollutions of electric vehicle; 4. Improvement of environmental safety of passengers by means of efficient air cleaning in passenger compartment; 5. Improvement in disposal and recirculation rate of vehicles in comparison with the existing ones.
IJAAS Vol. 7, No. 3, September 2018: 233 – 239
IJAAS
ISSN: 2252-8814
239
5.
CONCLUSIONS The achieved performances of NAMI driverless electric vehicle in total and its constituents (traction module, motion control system, electricity accumulation system) are close to those of global analogs and is superior to those of Russian analogs. The achieved performances of air cleaning and distribution system in passenger compartment exceed those of all known global and Russian analogs concerning environmental safety of passengers. The developed NAMI driverless electric vehicle is implemented in this configuration for the first time in global practice.
ACKNOWLEDGMENTS This work was supported by the agreement # 14.625.21.0006 with the Ministry of Education and Science of Russian Federation (unique project identifier RFMEFI62514X0006) to create an experimental model of a driverless environmentally friendly electric vehicle.
REFERENCES [1] M. Yguel, C. Tay, M. Keat, C. Braillon, C. Laugier and O. Ayvehicled, “Dense Mapping for telemetric sensors: efficient algorithms and sparse representation,” In Proceedings of Robotics: Science and Systems Conference, 2007. [2] B. Mihailov, “Technical vision in control systems of mobile objects,” In R.R. Nazirova (ed.), Proceedings of the Scientific Conference-Workshop. Iss. 4, Moscow, Publishing house of Lomonosov Moscow state university, pp: 191-202, 2011. [3] ISO 15765-4:2011 (E), Road vehicles - Diagnostic communication over Controller Area Network (Do CAN). Part 4: Requirements for emissions-related systems. Second edition, 2011. [4] A. Saykin, S. Bakhmutov, A. Terenchenko, D. Endachev, K. Karpukhin and V. Zarubkin, “Tendency of creation of “driverless” vehicles abroad,” Bioscience Biotechnology Research Asia, vol. 11, pp. 241-246, 2014. [5] A. Terenchenko, K. Karpukhin and R. Kurmaev, “Features of operation of electromobile transport in the conditions of Russia,” Paper of EVS 28 International Electric Vehicle Symposium and Exibition, KINTEX, Korea, 2015. [6] A. Saykin, S. Buznikov and K. Karpukhin, “The analysis of technical vision problems typical for driverless vehicles,” Research journal of pharmaceutical, biological and chemical sciences, vol. 7, n. 4, pp. 2053-2059, 2016. [7] A. Kawashima, K. Kobayashi, and K. Watanabe, “Implementation of Human-Like Driving for Autonomous Vehicle,” SAE Technical Paper 2001-01-0805, 2001. doi: 10.4271/2001-01-0805. [8] ISO 15031-5:2011 (E), Road vehicles - Communication between vehicle and external equipment for emissionsrelated diagnostics. Part 5: Emissions-related diagnostic services. Second edition, 2011. [9] S. Pala, “HUD Future in the Driverless Vehicle Society: Technology Leadership Brief,” SAE Technical Paper 201201-9022, 2012. doi: 10.4271/2012-01-9022. [10] A. Saykin, “A new concept of environmentally friendly transport,” LAP LAMBERT Academic Publishing, Мoscow, No. 978-3-659-39217-7, pp: 103, 2013. [11] A. Ivanov, V. Prikhodko and S. Shadrin, “Development of the external indirect pressure control system in pneumatic tires,” Life Science Journal, Т. 11, 12 (s), pp. 336, 2014. [12] V. Prikhodko, A. Ivanov and S. Shadrin, “The development of additional services using vehicle to person (V2P) interface,” Life Science Journal, Т. 12, 12 (s), pp. 862, 2014. [13] R. Felix, J. Economou, K. Knowles, “Driverless Vehicles and LIDAR: Evaluation of Possible Security Threats on the Open Road,” SAE Technical Paper 2015-01-0219, 2015. doi: 10.4271/2015-01-0219. [14] Hybrid and Electric Vehicles, IEA, Germany, pp: 328, 2015. [15] S. Shadrin, A. Ivanov and K. Karpukhin, “Using data from multiplex networks on vehicles in road tests, in intelligent transportation systems, and in self-driving cars,” Russian engineering research, vol. 36, n. 10, pp. 811814, 2016. [16] H. Dakroub, A. Shaout and A. Awajan, “Connected Car Architecture and Virtualization,” SAE Int. J. Passeng. Cars – Electron. Electr. Syst., vol. 9, n. 1, pp. 153-159, 2016. doi: 10.4271/2016-01-0081. [17] S. Shadrin A. Ivanov, “Algorithm of autonomous vehicle steering system control law estimation while the desired trajectory driving,” ARPN Journal of Engineering and Applied Sciences, vol. 15, n. 11, pp. 9312-9316, 2016.
Development of Russian Driverless Electric Vehicle (Andrey Mikhailovich Saykin)
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 240~244 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp240-244
240
Design of an IOT based Online Monitoring Digital Stethoscope B. Revanth Reddy1, S. Roji Marjorie2 , P. Ramakrishna3 Department of Electronics and Communication Engineering, Saveetha School of Engineering, Saveetha University, India
Article Info
ABSTRACT
Article history:
Acoustic stethoscopes have low sound levels. Digital stethoscope overcomes this issue by amplifying body sounds electronically. As the sound signals are transmitted electronically, it can be wireless and can provide noise reduction. Acoustic stethoscope can be changed into a digital stethoscope by inserting an electric capacity microphone onto its head. Heart sounds received from the microphone are processed, sampled and sound signals are converted analog to digital and sent wirelessly using the Internet of Things(IOT) techniques, so that multiple doctors can do auscultation and monitor conditions of the patient.
Received Nov 27, 2017 Revised Mar 30, 2018 Accepted Apr 13, 2018 Keyword: Acoustic Stethoscope Auscultation Digital Stethoscope Internet of Things
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Roji Marjorie.S, Departement of Electronics and Communication Engineering, Saveetha School of Engineering, Saveetha University, Thandalam, Chennai 602105, Tamilnadu, India. Email: roji.marjorie@gmail.com
1.
INTRODUCTION A Stethoscope is a device that helps in listening to the sounds of heart and lungs in our body [7]. By using stethoscope, the doctor can check the problems of the heart and lung of patient. Acoustic stethoscope is cheaper than electronic stethoscope. The function of electronic stethoscope is same as acoustic stethoscope. Acoustic stethoscopes are common to most people, and operation of sound transmission from the chest piece, through hollow tubes, to listener's ears [7]. The chest piece generally consists of two sides that can be placed on the patient for hearing sound. If the diaphragm puts on the patient, body sounds vibrate the diaphragm and creates acoustic pressure waves [2]. Those created acoustic pressure waves travel through the stethoscope, resulting in hearing of body sounds. This stethoscope was invented by Rappaport and Sprague in the early 20th century [7]. Acoustic stethoscopes produce very low sound. An electronic stethoscope amplifies low level body sounds and require conversion of acoustic sound waves to electrical signals which can be amplified and processed for optimal listening [1]. The simplest and low cost method of sound detection can be achieved by placing a microphone in the chest piece. Electronic stethoscope module consists of different types of components that can be used to amplify and optimize the sound signals in different frequencies. Sound signals can be digitized, encoded and decoded to have noise reduction [3]. Processed data can be sent to the cloud using Internet of Things techniques. Internet of Things is a technology that uses internet to control or monitor the electronic devices [6]. Heart beat sounds of a digital stethoscope is monitored over internet using the IOT and then graphs can be drawn [5].
2.
DESIGN Preamplifier is used to amplify low level electrical signals for further processing. Filters allows selection of suitable frequencies, so that particular heart sound frequencies can be reproduced. ADC converts
Journal homepage: http://iaescore.com/online/index.php/IJAAS
IJAAS
ISSN: 2252-8814
241
analog signals to digital, so that the heart beat sounds can be processed and then sent over internet wirelessly. 2.1 Block Diagram
Figure 1. Block Diagram
2.2 Condenser Microphone A microphone or mic is an electric transducer that coverts the acoustic sounds to electrical signals. It is inserted into head of the mechanical stethoscope which converts heart sounds received by diaphragm to the electrical signals. 2.3 Preamplifiers A preamplifier is used to generate a small electrical signal for further amplification. Two op-amps of LM324 integrated circuit chip employed to amplify the signals from the piezoelectric crystal.2 This is AC coupled to the first amplifier stage, which provides a gain of about 20. The second stage, which is identical to the first stage in all respects, provides a subsequent gain of about 20 to give a total amplification of 400. The input signals to the pre amplifier from the piezoelectric transducer are in the order of 0.5mV. Continuous variation of gain is achieved through the 10kΩ potentiometer at the output of the second stage of the Preamplifier. The input signal to the pre-amplification stage is amplified twice using two identical noninverting amplifiers. Capacitor C1is the coupling capacitor for the first amplification stage and C4 for the second amplification stage respectively. Capacitors C2 and C5 are the feedback capacitances of stage one and two. These capacitances serve to improve the stability of the circuit and the low frequency response.
Figure 2. Pre-amplifier
2.4 Filters Filters permit selection of suitable frequency bands, so that particular heart sound frequencies can be reproduced. In general, the high frequency components of cardiovascular sound have a much smaller intensity than the low frequency components. In general, three to four sets of such filters with different cutoff frequencies are used for clinical recording of the PCG as the shape of a murmur is often characteristically Design of an IOT based Online Monitoring Digital Stethoscope (B Revanth Reddy)
242
ISSN: 2252-8814
displayed in one channel better than the other. Active low pass and high pass filters are used in this equipment because of the advantages like flexibility in gain and frequency adjustments, high input impedance and low cost. The requirement of the filter in this is to attenuate any frequency below 10Hz and above 1KHz. 2.5 Power Amplifier The LM386 is a power amplifier used in low voltage using applications. The inputs are grounded while output is automatically biased to one half of the input voltage. The quiescent power drain is only 24mW while operating from a 6v supply, makes the LM386 ideal for operation of battery.
Figure 3. Power amplifier
The output of filter is given to the LM386 audio amplifier. In power amplifier, Capacitor2 maintains DC bias levels in the gain adjustment circuit. Capacitor4 gives power supply decoupling, and Capacitor5 is output coupling capacitor. Capacitor6 & Resistor2 act as a zobel network providing a high frequency load to maintain stability where loudspeaker inductive reactance may become excessive. POT 1 provides adjustable input level attenuation. 2.6 Analog to Digital Converter A circuit which converts a signal from continuous to discrete or analog to digital. Analog signals can be directly measured. Digital signals have two binary states, 0 and 1. Microprocessors can perform only difficult processing on digitized signals. When signals are in digital they have less possibility of additional noise. Analog to digital converters are used, where an analog signal can be processed, stored, and transmitted in digital form. 2.7 Micro Controller The AT89S52 is a low-power 8-bit microcontroller with 8K bytes of in-system programmable Flash memory. The device is manufactured using Atmel’s high-density nonvolatile memory technology and is compatible with the Indus-try-standard 80C51 instruction set and pin out. The Idle Mode stops the CPU while allowing the RAM, timer/counters, serial port, and interrupt system to continue functioning. The Power-down mode saves the RAM con-tents but freezes the oscillator, disabling all other chip functions until the next interrupt or hardware reset. 2.8 UART The A UART is a part of an integrated circuit used for serial communications over a computer or peripheral device serial port. The Universal Asynchronous Receiver/Transmitter (UART) controller is a component of the serial communications to computer. The UART uses bytes of data and transfers the individual bits in a sequential pattern. Serial transmission of digital data through RS-232 cable.
IJAAS Vol. 7, No. 3, September 2018: 240 – 244
IJAAS
ISSN: 2252-8814
243
2.9 Working of Module A Condenser microphone is used in the chest-piece of the stethoscope which converts the heart sounds in to electrical signals. These electrical signals are preamplified and then given to filters. In each case of filters, heart sounds are processed in such a way that external noise is reduced. Obtained heart sounds are again amplified using power amplifier which helps in getting high voltage signals. ADC is used to convert the analog output at power amplifier to digital pulse. Pulse signals are counted using protocols in AT89S52 microcontroller and then heart beat rate per minute is displayed in LCD. 2.10 Transmission of Heart Beat Rates Through IOT Heart beat rates displayed in the LCD are transmitted to PC through UART, Asynchronous serial transmission and receiver using RS232 cable. Data received in the PC are processed in USB serial communication package which uploads the data to the cloud storage. An authentication step is used in retrieving the data, as this is related to health-related issues. Data can be retrieved and displayed in a website or android app, so that mutiple doctors can monitor the patient conditions from any where at any time.
3.
RESULTS AND ANALYSIS The pickup transducer is placed to get the heart sounds. The electric signals in each case are preamplifier and then processed by suitable filters. The filtered signal is provided to power amplifier and then converted to digital signals using analog to digital converters and digital signal processors. The Working of this IOT based Digital Stethoscope is, Heart beats are picked up using Condenser microphone and converts the audio sounds into electric signals. Converted Electrical signals are Amplified in the pre-Amplifier circuit. Amplified signals are then filtered with various types of low pass and high pass filters to reduce the noise. Reduced noise signals are again amplified with the power amplifier circuit. Input for this power amplifier is the output of high pass filter. Amplified signals with power amplifier are can be converted into digital using analog to digital converter and digital signal is processed. The processed signal is given to the microcontroller and then it sends to cloud or internet. The Heart sounds picked up by condenser microphone are amplified in this circuit. Once the Heart sounds are processed in the preamplifier and filters circuit, the signals can be converted from analog to digital using ADC.
Figure 4. Pre amplifier and filters
Figure 5. Wireless stethoscope using IOT
Figure 6. Heartrates in browser
Design of an IOT based Online Monitoring Digital Stethoscope (B Revanth Reddy)
244
ISSN: 2252-8814
The pre amplifier, active filters and power amplifiers. The pickup transducer is placed to get the heart sounds. The electric signals in each case are preamplifier and then processed by suitable filters. Implemented wireless stethoscope using FM transmitter can be modified by converting analog to digital signals using ADC and then can be sent to cloud using IOT techniques. Heartbeats counted by micro controller is displayed in the LCD as well as in a site using IOT..
5.
CONCLUSION A Digital Stethoscope has been implemented and signals can be transmitted through IOT. Heart beats are picked up using Condenser microphone and converts the audio sounds into electric signals. Converted Electrical signals are Amplified in the pre-Amplifier circuit. Amplified signals are then filtered with various types of low pass and high pass filters to reduce the noise. Reduced noise signals are again amplified with the power amplifier circuit. Input for this power amplifier is the output of high pass filter. Amplified signals with power amplifier are can be converted into digital using analog to digital converter and digital signal is processed. The processed signal is given to the microcontroller and then it sends to cloud or internet.
REFERENCES [1] Habin Wang, Jian Chen, Choi Samjin, “Heart Sound Measurement and Analysis System with Digital
Stethoscope”, International Conference on Biomedical Engineering and Informatics, 2009. [2] Yang Tang, Guitao Cao, Hao Li, “The design of electronic heart sound stethoscope based on Bluetooth”
4th International Conference On Bioinformatics and Biomedical Engineering (ICBBE), 2010. [3] Ashish Harsola, Sushil Thale, M.S. Panse, “Digital Stethoscope for Heart Sounds”, International
Conference and workshop on Emerging Trends in Technology (ICWET), 2011. [4] R. P. Singh, S.D. Sapre, “Communication Systems Analog & Digital” Second edition, Tata
McGrow–HillPublication. [5] “Sound sending over internet” by david middlecamp published in adafruit, music, particle. [6] ESP8266 Wi-Fi + Arduino upload to xively and Thingspeak byjanisalnis in Internet of Things. [7] Wikipedia, https://en.wikipedia.org/wiki/Stethoscope
IJAAS Vol. 7, No. 3, September 2018: 240 – 244
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 245~254 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp245-254
245
Analysis and Implementation of Unipolar PWM Strategies for Three Phase Cascade Multilevel Inverter Fed Induction Motor Drive Ravikumar Bhukya1, P. Satish Kumar2 Departement of Electrical Engineering, University College of Engineering, Osmania University, Hyderabad, Telangana, India
Article Info
ABSTRACT
Article history:
This paper presents unipolar pulse width modulation technique with sinusoidal sampling pulse width modulation are analyzed for three-phase five-level, seven-level, nine-level and eleven-level cascaded multi-level inverter. The unipolar PWM method offers a good opportunity for the realization of the Three-phase inverter control, it is better to use the unipolar PWM method with single carrier wave compared to two reference waves. In such case the motor harmonic losses will be considerably lower.The necessary calculations for generation of unipolar pulse width modulation strategies have presented in detail. The unipolar SPWM voltage switching scheme is selected in this paper because this method offers the advantages of effectively doubling the switching frequency of the inverter voltage. The cascaded multi level inverter fed induction motor is simulated and compared the total harmonic distroction for all level (five-level, seven-level, nine-level and elevel-level)of the inverter. Theoretical investigations were confirmed by the digital simulations using MATLAB/SIMULINK software.
Received Nov 13, 2017 Revised Feb 17, 2018 Accepted Mar 11, 2018 Keyword: CHB MLI Induction motor THD Unipolar PWM
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Ravi kumar Bhukya, Departement of Electrical Engineering, University college of Engineering, Osmania University, Hyderabad, Telangana, INDIA, 50007. Email: rkpurnanaik2014@gmail.com
1.
INTRODUCTION The pulse width modulation (PWM) strategies are the most effective to control multilevel inverters. The unipolar PWM are the most preferred pwm control techniques. Even though space vector modulation (SVM) is complicated, it is the preferred method to reduce power losses by decreasing the power electronics devices switching frequency, which can be limited by pulse width modulation. The operation theory will be discussed with the aspect of Unipolar PWM [1]. In the region of high voltage and power, a high quality inverter fed ac drive is more easily obtained by the use of multi-level and in the first turn of three-level inverters. The neutral point clamped three-level inverter topology is presented. Several PWM methods for this inverter have been elaborated previously [2], [3]. The most popular method is based on the intersection of a single sinusoidal reference with N-1 triangular carriers, originally proposed by carara et al [4], [5].three alternative carrier disposition schemes are suggested, namely PD,POD,APOD.Another important topology, named Cascade H-Bridge (CHB), has fewer components to achieve the same number of output voltage levels. It is one of the most widely used topologies due to its reliability and increased capacity to operate under fault conditions in the cells [6], [7]. In this paper the Unipolar PWM strategy of five-level, seven-level, nine-level and eleven-level inverters are compared for THD. The paper mainly deals with the computation and the comparison of the motor harmonic losses of unipolar PWM solutions and with the selection of the solutions providing the best Journal homepage: http://iaescore.com/online/index.php/IJAAS
246
ISSN: 2252-8814
results. Finally, the drive harmonic losses will be compared for (five-level, seven-level,nine-level and elevenlevel) inverters.
2.
THREE-PHASE N-LEVEL CASCADED H-BRIDGE MULTI-LEVEL INVERTER The three phase Series H-bridge inverter (SHI) or Cascaded Multi-Level Inverter (CMI) topology shown in Figure 1. by al will have broad strategy with an better harmonic performance. This strategy consists single carrier from each H-bridge cell (Where N is number of level). Single leg N-level inverter CHB inverter is shown in Figure 1. Each cell contains four active switches S1 to S4 with anti-parallel diodes D1 to D4.Each cell contains separated DC source. And each cell produced three different voltage level (+E,0,-E). The phase voltage consists of N-level inverter. the levels in phase voltage are 2N+1, here N is 1-ϕ inverter cells present in a phase and the number of levels in line voltage are 2M-1, where M is the number of level in phase voltage [8], [9].
R S1
S2
S3
S4
S5
S6
S7
S8
Vdc1
Vdc2
Sn1
Sn2
Vdcn Sn3
Sn4
N Figure 1. Single leg N-level cascaded H-bridge multilevel inverter
3.
GENERALIZED UNIPOLAR PWM TECHNIQUE FOR CASCADED H-BRIDGE MULTILEVEL INVERTER Figure 2. Illustrates the unipolar multi level modulation strategy.the modulation signals have the same frequency (fo) and amplitude (Am). The sinusoidal signals are sampled by a triangular carrier signal with frequency (fc) and amplitude (Ac) once every cycle, intersection between the sampled modulation signals and the carrier signal defines the switching instant of the PWM pulses. In order to ensure quarter wave symmetric properties of the PWM output waveform,the starting point of the modulation signals is phase shifted by one period of the carrier wave, in addition the frequency modulation ratio (mf) must also be even number [10], [11]. The unipolar PWM method offers a good opportunity for the realization of the Three-phase inverter control. In case of the five-level, seven-level, nine-level and eleven-level inverters it is better to use the unipolar PWM method with three carrier waves. In such case the motor harmonic losses will be considerably lower.The unipolar spwm voltage switching scheme is selected in this paper because this method offers the advantages of effectively doubling the switching frequency of the inverter voltage.A particular advantage of the unipolar PWM approach is that,this method reduces the harmonics in the three phase inverter. IJAAS Vol. 7, No. 3, September 2018: 245 – 254
IJAAS
ISSN: 2252-8814
247
That means,selecting unipolar PWM as switching scheme in the proposed inverter is appropriate as there is no filter at the inverter output compared to other spwm techniques. In this scheme, the triangular carrier waveform is compared with two reference signals which are positive and negative signal. The basic idea to produce SPWM with unipolar voltage switching is shown in Figure 2. The different between the Bipolar SPWM generators is that the generator uses another comparator to compare between the inverse reference waveform−Vr . The process of comparing these two signals to produce the unipolar voltage switching signal is graphically illustrated in Figure 2. In Unipolar voltage switching the output voltage switches between 0 and Vdc, or between 0 and −Vdc. This is in contrast to the Bipolar switching strategy in which the output swings between Vdc and−Vdc. Over modulation occurs when amplitude modulation index ma is greater than unity. It causes a reduction in number of pulses in the line to line voltage waveform leading to emergence of lower order harmonics. Moreover the notch and pulse widths near the center of positive and negative half cycle tend to vanish. To complete the switching operations of the device, minimum notch and pulse widths must be maintained. When minimum width notches and pulses are dropped, there will be some transient jump of load current. The generated signals are as shown in Figure 2 and the switching pulses generated shown in Figure 3.and the output currents at the load shown in Figure 4. The modulation index is the ratio of peak magnitudes of the modulating signal Vm and the carrier signal:
m=
(2)
Vm Vc
The modulation index in SPWM technique for cascaded multilevel inverter configuration is given by:
m=
(3)
Vm ( N − 1)Vc
where N is number of levels. For under modulation 0 < m < 1. For over modulation m>1. Generally, over modulation is not desired because of the presence of the lower frequency harmonics in the output voltage and subsequent distortion in the load current [12]. Frequency modulation is the ratio of frequency of the triangular carrier signal fc to the frequency of sinusoidal reference signal fs . It controls harmonics in the output voltage.
mf =
(4)
fc fs
Figure 2. Unipolar PWM technique pulse generation
Analysis and Implementation of Unipolar PWM Strategies for Three Phase Cascade… (Ravikumar Bhukya)
248
ISSN: 2252-8814
Figure 3. Switching Pulses generation using Unipolar PWM technique
Figure 4. Output current at the load
4.
SIMULATION RESULTS AND ANALYSIS The N-level cascaded multilevel inverter fed induction motor is considered for simulation. The simulation study is carried out using MATLAB/SIMULINK. The system parameters are shown in Table 1.
Table 1. The Parameter of the Induction Motor Parameters Input voltage Inverter voltage Rotor speed Fundamental frequency Switching frequency Reference speed
Specifications 400VRMS(PhasePhase) 100(Volts) 1440(RPM) 1200(Hz) 1200(Hz) 1500(RPM)
The simulation of the cascaded multi-level inverter fed induction motor and observe the stator current,torque,speed of the induction motor and THD of the line voltage (five-level, seven-level, nine-level and eleven-level). As per the principle of unipolar pwm two reference sine and single carrier waves are compared. The generated signals are as shown in Figure 2. Which contain single carrier comparead other spwm techniques (multi carrier techniques). 3.1. Unipolar PWM technique applied to five-level CHBMLI The output voltage of the inverter for line to line is about 410V when using Unipolar PWM. Here in Figure 5.one can observe the line voltages of five level inverter. Here in Figure 6. Indicates the three phase output stator currents of the load and we can observe the system is unstable from o to 0.15 sec.due to transient behavior of the system at the starting from 0.15 sec. Figure 7.indicates the torque, speed of the load of five level inverter from the figure it can be seen that the steady state operation of the system has achieved at 0.2 sec. The total harmonic distortion of the output voltage is about 23.13%, at ma=1.5, mf=31.84. When using Unipolar PWM technique on five-level cascaded multi-level inverter shown in Figure 8.
IJAAS Vol. 7, No. 3, September 2018: 245 – 254
IJAAS
ISSN: 2252-8814
249
Figure 5. Proposed five-level Unipolar PWM technique (Line voltage)
Figure 6. Output stator currents of the load (Five-level inverter)
Figure 7. Output torque and speed of the load (Five-level inverter)
Figure 8. THD For Unipolar technique ma=1.5,mf=31.84
Analysis and Implementation of Unipolar PWM Strategies for Three Phase Cascade… (Ravikumar Bhukya)
250
ISSN: 2252-8814
3.2. Unipolar PWM technique applied to seven-level CHBMLI The output voltage of the inverter for line to line is about 410V. Here in Figure 9. One can observe the line voltages of seven level inverter using unipolar PWM technique. Here in Figure 10. Indicates the three phase output stator currents of the load and we can observe the system is unstable from o to 0.2 sec. Due to transient behavior of the system at the starting from 0.2 sec. Figure 1I. Indicates the torque, speed of the load of seven level inverter from the figure it can be seen that the steady state operation of the system has achieved at 0.15 sec. Figure 12. Indicate the total harmonic distortion of the output voltage is about 15.33%, at ma=1.3, mf=31.84 when using Unipolar PWM technique on seven-level cascaded multi-level inverter.
Figure 9. Proposed seven-level Unipolar PWM technique (Line voltage)
Figure 10. Output stator currents of the load (Seven-level inverter)
Figure 11. Output torque and speed of the load (Sevenive -level inverter)
Figure 12. THD For Unipolar technique ma=1.3,mf=31.84 IJAAS Vol. 7, No. 3, September 2018: 245 – 254
IJAAS
ISSN: 2252-8814
251
3.3. Unipolar PWM technique applied to nine-level CHBMLI The output voltage of the inverter for line to line is about 410V. Here in Figure 13. One can observe the line voltages of nine level inverter using unipolar PWM technique. Here in Figure 14. Indicates the three phase output stator currents of the load and we can observe the system is unstable from o to 0.3 sec. Due to transient behavior of the system at the starting from 0.3 sec.and Figure 15. Indicates the torque, speed of the load of nine level inverter from the figure it can be seen that the steady state operation of the system has achieved at 0.35 sec. Figure 16. Indicate the total harmonic distortion of the output voltage is about 10.78%, at ma=1.2, mf=31.84 when using unipolar PWM technique on nine-level cascaded multi-level inverter.
Figure 13. Proposed nine-level Unipolar PWM technique (Line voltage)
Figure 14. Output stator of the load (Nine-level inverter)
Figure 15. Output torque and speed of the load(Nine-level inverter)
Analysis and Implementation of Unipolar PWM Strategies for Three Phase Cascade… (Ravikumar Bhukya)
252
ISSN: 2252-8814
Figure 16. THD For Unipolar technique ma=1.2,mf=31.84
3.3. Unipolar PWM technique applied to eleven-level CHBMLI The output voltage of the inverter for line to line is about 410V.Here in Figure 17. One can observe the line voltage of eleven level inverter using unipolar PWM technique. Here in Figure 18. Indicates the three phase output stator currents of the load and we can observe the system is unstable from o to 0.4 sec. Due to transient behavior of the system at the starting from 0.4 sec. Figure 19. Indicates the torque, speed of the load of eleven level inverter from the figure it can be seen that the steady state operation of the system has achieved at 0.45 sec. Figure 20. Indicate the total harmonic distortion of the output voltage is about 10.13%, at ma=1.0, mf=31.84 when using unipolar PWM technique on eleven-level cascaded multi-level inverter.
Figure 17. Proposed eleven-level Unipolar PWM technique (Line voltage)
Figure 18. Output stator current of the load (Eleven-level inverter)
IJAAS Vol. 7, No. 3, September 2018: 245 – 254
IJAAS
ISSN: 2252-8814
253
Figure 19. Output torque and speed of the load (Eleven-level inverter)
Figure 20. THD for Unipolar technique ma=1.0,mf=31.84
The total harmonic distraction waveforms obtained from the simulation results using MATLAB/SIMULINK and compared the all levels (five, seven, nine and eleven)of the cascaded multi level inverter. The comparison are made for the total harmonic distraction wave forms based on variations of parameters amplitude modulation (ma) and frequency modulation (mf).and we can observe the three phase stator currents, torque and speed of the induction motor for all levels of the inverter.the total harmonic distortion of the output voltage is about 10.13%,at ma=1.0,mf=31.84 on eleven-level cascaded multi-level inverter.The Comparison of THD analysis of a three-phase cascaded inverter fed to induction motor shown in Table 2.
Table 2. The Comparision of THD Of Five, Seven, Nine and Eleven-level Cascaded Inverter Output Voltage Level Five-level Seven-level Nine-level Eleven-level
Techniques Unipolar PWM Unipolar PWM Unipolar PWM Unipolar PWM
%THD(V) 23.13 15.33 10.78 10.13
5.
CONCLUSION In this paper the Unipolar PWM technique of most proffered control strategies applied to three phase cascaded inverter are presented.The waveforms clearly depicting that almost all the five-level, sevenlevel, nine-level and eleven-level control strategies are functioning well in controlling the speed of the induction motor, output line voltage, stator current, speed and torque. But THD of the Unipolar PWM strategies when compared, it is obvious that eleven-level cascaded inverter is the low THD of about 10.13%.
ACKNOWLEDGEMENTS We thank the University Grants Commission (UGC), Govt. of India, New Delhi for providing Major Research Project to carry out the Research work on Multi-Level Inverters. I also thank UGC for awarding me with RGNF FELLOWSHIP to carry out my research work (Ph.D). Analysis and Implementation of Unipolar PWM Strategies for Three Phase Cascade… (Ravikumar Bhukya)
254
ISSN: 2252-8814
REFERENCES [1]
S.Zainal,J.Aziz,S.S.Ahmed, “Single carrier PWM scheme for cascaded multilevel voltage source inverter,”Power Electronics and Drive systems,IEEE Conference (PEDC), 2003, Vol. 1, pp. 406-410.
[2]
Akira Nabae, Isao Takahashi and Hirofumi Akagi,” A New Neutral-Point-Clamped Pwm Inverter”, IEEE Trans On Industry Applications, Vol. Ia-17, No. 5. September/October 1981, Pp 518-523.
[3]
Irfan Ahmed, Vijay B. Borghate,” Simplified space vector modulation technique for seven-level cascaded H-bridge inverter”, IET Power Electron. 2014, Vol. 7, Iss. 3, pp. 604–613. Carrara G.et al., “A new multilevel PWM method;A theoretical analysis”, Power Electronics Specialists Conference ,1990,PESC’90 record, 21st annual IEEE, pp. 363-371. M. Calais, L.J. Borle, V.G. Agelidis, “Analysis of multicarrier PWM methods for a single-phase five level inverter,” IEEE Power Electronics Specialists Conference (PESC), 2001, Vol. 3, pp. 1351-1356.
[4] [5] [6]
[7] [8] [9]
Ravi kumar Bhukya, P. Satish kumar “Performance Analysis of Modified SVPWM Strategies for Three Phase Cascaded Multi-level Inverter fed Induction Motor Drive”, international journal of power electronics and drive systems, Vol. 8, No. 2, pp. 835-843, May.2017. J.Rodriguez, J.S.Lai, and F.Z.Peng, “Multilevel inverters: A survey of topologies, control and applications,” IEEE Trans.Ind.Electron., vol.49, no.4, pp.724-738, Aug.2002. M.Malinowski, K.Gopakumar, J.Rodriguez and A.Perez,” A Survey on Cascaded Multilevel Inverters,” IEEE Trans.Ind.Electron, vol.57, No.7, pp.2197-2205, July 2010. F. Salim andN.A. Azli, "Development ofan FPGABased Gate Signal Generator for a Multilevel Inverter." Proceedings of2003 International Conference on Power Electronics and Drive Systems PED 2003, Singapore. 17-20 November 2003.
[10] P. Palanivel, S.S. Dash “Analysis of THD and output voltage performance for cascaded multilevelinverter using carrier pulse width modulation technique” IET Power Electronics vol. 4, no. 8, pp. 951-958, 2010. [11] M. Satyanarayana, P. Satish Kumar, “Analysis and Design of Solar Photo Voltaic Grid Connected Inverter”, Indonesian Journal of Electrical Engineering and Informatics (IJEEI), IAES, December 2015, Vol. 3, pp 199-208. [12] Ravikumar Bhukya, P.Satish kumar, Modeling, Analysis and Comparative of Down Sampling based Clamping SV PWM for Cascaded and Diode Clamped Multilevel Inverter fed Induction Motor Drive. Indonesian Journal of Electrical Engineering and Computer Science Vol. 7, No. 3, September 2017, pp. 698 ~ 707 DOI: 10.11591/ijeecs.v7.i3.pp698-707.
IJAAS Vol. 7, No. 3, September 2018: 245 – 254
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 255~264 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp255-264
255
Energy and Load Aware Routing Protocol for Internet of Things S.Sankar1, P.Srinivasan2 1
Research Scholar, School of Computer Science and Engineering, VIT University, Vellore-632014, Tamilnadu, India; 2 Associate Professor, School of Information Technology and Engineering, VIT University, Vellore-632014, Tamilnadu, India
Article Info
ABSTRACT
Article history:
Maximizing the network lifetime is one of the major challenges in Low Power and Lossy Networks (LLN). Routing plays a vital role in it by minimizing the energy consumption across the networks through the efficient route selection for data transfer. IPv6 Routing Protocol for Low Power and Lossy Networks (RPL) is a IETF standardized IPv6 routing protocol for LLN. In this paper, we propose Energy and Load aware RPL (EL-RPL) protocol, which is an enhancement of RPL protocol. It uses a composite metric, calculated based on expected transmission count (ETX), Load and battery depletion index (BDI), for the route, selection. The COOJA simulator is used for performance evaluation. EL-RPL is compared with other similar protocols RER(BDI) RPL and fuzzy logic based RPL (OF-FL RPL). The simulation result shows that the EL-RPL improves the network lifetime by 812% and packet delivery ratio 2-4%.
Received Nov 28, 2017 Revised Feb 18, 2018 Accepted Mar 11, 2018 Keyword: Enery Efficiency Internet of Things IPv6 Routing Protocol for LowPower and Lossy Networks Load
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: P.Srinivasan Associate Professor, School of Informationa and Technology, VIT University, Vellore-632014, Tamilnadu, India; Email: srinivasan.suriya@vit.ac.in
1.
INTRODUCTION Internet of Things (IoT) is an emerging research area of last few years and it is a future internet technology. IoT enables the machine-to-machine communication that able to exchange the data between them and take the decision accordingly without involvement of human [1-2]. Nowadays, IoT system is involved in various field and the application areas such as smart home, wearable, smart grid, smart city, connected car, smart retail, smart farming, smart supply and etc.,. IoT is a new paradigm that introduced a new way of communication and it is denoted as Low Power and Lossy Networks (LLN) [3]. In LLN, the nodes are highly resource constrained i.e, low power, processing and memory capacity. The existing routing protocols namely on-demand distance vector (AODV), optimized link state routing(OLSR), open shortest path first (OSPF) and intermediate system to intermediate system (IS-IS) for ad-hoc network that unable to fulfill the requirements of LLN. So, IETF working group is standardized an IPv6 Routing protocol for Low power and Lossy networks (RPL) for LLN [4-5]. RPL is a distance vector and source routing protocol and it follows the Destination oriented Directed Acyclic graph (DODAG). The top of the node is represented as root node. The edges directed towards the root called upward routing and edges are far away from root called downward routing. RPL protocol generates more than one RPL instances in a network and each instance contains more than one DODAG. The entire RPL instance maintains in separate RPL instance ID or DODAG ID. Each node has own IPv6 address. In LLN, the top of the node acts as border router and remaining entire node acts as host node of DODAG. The host node gets the address from root ID prefixes. The border router or root node maintains the entire Journal homepage: http://iaescore.com/online/index.php/IJAAS
256
ISSN: 2252-8814
network information in a routing table. RPL has two types of nodes i.e, storing and non- storing mode. The storing mode can collect and forward the data to other node. The non-storing mode simply forwards the data to other node. In RPL, DODAG construction process is used for three control messages namely, DODAG Information Solicitation (DIS), DODAG Advertisement Object (DAO), DODAG Advertisement ObjectAcknowledgement (DAO-ACK). The participant node sends the willingness to DODAG or DODAG sends the DIO control messages to all the participants. Then the participant node sends the DAO control messages with in trickle time. Finally, the DODAG sends the DAO-ACK confirmation messages to all participants [6-7]. The proposed work main objective is to maximize the network lifetime by minimizing the node energy consumption. The contribution of this work is to introduce a combination of ETX, Load and BDI based composite metric in RPL. This composite metric follows the minimizable property. DODAG sends DIO control messages to all participant nodes. The participant node selects the best parent from DODAG rank. The rank calculates from minimum value of the composite metric in the DODAG. Finally, sender or participant node sends the data to DODAG root towards the best parent in the DODAG. Thus, it improves the packet delivery ratio, reduce the traffic load and improve network lifetime. 1.1. Problem Statement RPL is designed and implemented with variaous features such as delay, self healing, loop-free topology and load balance. Nevertheless, they are not considered at a time, these metrics such as Load, Residual Energy and Link quality and this imbalance is a significant limitation in RPL. This problem creates more impact to drain the node energy in near to the sink node and as well as intermediate node. The proposed routing protocol (EL-RPL) balances the energy and load among the network nodes and it avoids the bottle neck in near to the sink node and also intermediate node. 1.2. RPL Protocol Based Related Works In this section briefly discuss energy efficient RPL protocol issues in LLN. The composite metric indicates the composition of more than one routing metric in DODAG. The composite metric provides the requirement of convergence, optimality, and loop-freeness in LLN. Zahariadis et’al [8] proposed a design guideline for routing metrics composition in LLN and this document is standardized by IETF. It is clearly stated the properties, rules and requirement of composite metrics in LLN. The composition of the primary metric can be combined additive and lexical manner. Additive property is suitable for composition of primary metric consists of either minimizable or maximization property. Fuzzy logic is suitable for composition of single metric consists of both minimizable and maximization. Lexicographic property is suitable for composition of single metric and it is inspected first metric and only if possible path have equal value then it considers the next metric from the composition. Oana Iova et’ al [9] proposed expected lifetime (ELT) as routing metric and it estimates the bottleneck of the network node. It mainly focused on network holes near the sink node. It is considered the factors such as traffic and link reliability. ELT metric avoids the early node battery depletion during the data transmission and network holes near sink node. Ali Hassan et’al [4] proposed an improved routing metrics for RPL protocol in LLN. This composite metric based on the combination of residual energy (RES), expected transmission count (ETX) and battery depletion index (BDI). It improves the network lifetime and reduces the battery depletion. However, this work is not given the preferences to link quality (ETX). Patrick-Olivier Kamgueu et’ al [10] proposed a fuzzy inference mechanism based composite metric. This composite metric based on the combination of delay, ETX and residual energy. It mainly concentrated on quality of service in LLN. It minimizes the energy consumption and improves the quality of service. Panagiotis Karkazis et’al [11] evaluated the routing metric composition to improve the Quality of Service (QoS) in LLN. It focused on three things. a). it introduced a composite metric from the single metric such as Remaining energy (RE), Expected transmission count (ETX), Packet forwarding indication (PFI). b). It provided better QoS and it proved the efficiency using routing algebra. c). it achieved better performance and it is provided the optimal loop- free paths.
2.
SYSTEM MODEL Figure 1 indicates the IoT low power and lossy networks system model. Thousands of sensor nodes are connected to the Border Router (BR) via mesh network. The BR or Low power Border router (LBR) connects the LLN edge devices in to both public or private internet and various application servers are IJAAS Vol. 7, No. 3, September 2018: 255 – 264
IJAAS
ISSN: 2252-8814
257
connected in the internet. In LLN, the edge devices are generating the data and send to the border router and various applications access those sensor data through internet. In LLN, we can perform the upward and downward routing. If one device wants to communicate with other device, that node sends the data to border router and it forwards the data to corresponding destination through downward routing [12, 13].
Figure 1. Low Power and Lossy Networks System Model
3.
THE PROPOSED EL RPL PROTOCOL In this paper, we propose energy and load aware composite routing metric (EL-RPL). EL-RPL is a composite routing metric and it is based on the combination of load, battery depletion index (BDI) and expected transmission count (ETX). In EL-RPL, the participant node selects the best parent among the preferred parent based on DODAG rank. The rank calculates from objective function and rank increase. Each node and link associates with Load, BDI and ETX metrics in LLN. This composite metric improves the network lifetime. 3.1. Metrics of Interest 3.1.1. Expected Transmission Count (ETX) Expected Transmission Count is a link metric and it is used to predict the link quality based on transmission and including retransmission. The ETX metric formula calculates from Equation (1) and (2). Link ETX Link ETX represents the forward and reverse data delivery of link. The forward data delivery (df) represents the probability that a data packet, successfully arrives at the recipient. Reverse data delivery (dr) represents the probability that the ACK packet received successfully from the recipient [14]. The link ETX calculates from Equation (1).
ETX ( Ni ) =
1 df * dr
(1)
Route ETX Route ETX uses to find link quality of particular path Px. The Route ETX calculates from Equation
ETX ( Px) = ∑i =1 ETX ( Ni ) n
(2)
Energy and Load Aware Routing Protocol for Internet of Things (S.Sankar)
258
ISSN: 2252-8814
3.1.2. Load Network data traffic is an amount of data transfer across the network at given amount of time. Load balance is a technique and it is used to balance the traffic across network. It is mainly concentrated on number of child present in each parent node [15]. The participant node selects the parent node based on less number of child accumulated parent node in the DODAG.The traffic load calculates from equation (3) and (4). If the number of children increases in a parent node in the DODAG, EL-RPL reconstructs the DODAG. a. To calculate the Load In EL-RPL, load of Path(x) calculation is based on the cumulative of node traffic or child set.
Load ( Path( x)) =
n
∑ Node _ TrafficLoad (M )
(3)
M =1
b. To calculate the Node Traffic In EL-RPL, node traffic calculates from children count of the respective parent node n
Node _ TrafficLoad ( M ) = ∑ children _ count
(4)
i =1
3.1.2. Battery Depletion Index (BDI) Battery depletion Index (BDI) indicates that how much percentage of energy depleted from battery present in the node. The residual energy calculates from initial energy and remaining energy of the node [4]. The residual energy calculates from equation (5).
RER( Mi ) =
Eremaining Einitial
(5)
The residual energy is a remaining energy in the node Mi and it is represented in terms of 0 to 1. The BDI calculation is calculated from Equation (6).
BDI ( Mi) = (1 − RER( Mi))
(6)
The BDI follows the puductive rule and BDI of Path Px calculates from Equation (7). n
BDI ( Px) = ∏ BDI ( Mi)
(7)
i =1
3.2. Objective Function (OFEL) In DODAG, the parent selection is based on DODAG rank. The DODAG rank calculates from Min hop rank increase and objective function. The objective function takes into account the following issues such as loop freeness, data load in upward and downward routing and bottle neck near sink node. The proposed EL-RPL is evaluated the performance and fine tuned the weight values. Moreover, it provides the better efficiency, where the weight values are w1,w2 and w3=1/3. min OF(LB, BDI, ETX) = w1 × Load (Pi) + w2 × BDI (Pi) + w3 × ETX (Pi)
(8)
Step = w1 × Load (Pi) + w2 × BDI (Pi) + w3 × ETX (Pi)
(9)
3.2. Rank Calculation In EL-RPL, DODAG rank calculates from parent rank and rank increase value. The rank increase calculates from step value and MinHopRankIncrease. The MinHopRankIncrease default value is 256 [7]. The step value calculates from objective function and it is denoted in Equation (8). The rank calculates from Equation (9-12).
IJAAS Vol. 7, No. 3, September 2018: 255 – 264
IJAAS
ISSN: 2252-8814
259
Fixed_Point or Rank_increase = Step + MinHopRankIncrease
(10)
Rank(N) = floor(Rank (N)/MinHopRankIncrease)
(8)
Rank (N) = Rank (Parent_Node) + Rank_ increase Finally, the Rank calculation based on the below equation (12)
(8)
For Instance, we have considered 6 nodes in the DODAG and each node contains Load, BDI and ETX of value. We considered that root node as “A” and its rank as 1. Followed by, calculate the rank of Node “B and C” based on the fixed point value and Step value. Node “B and C” are bottle neck node, because it is near to sink node. EL-RPL balances the load and energy in the networks. So it avoids the early battery depletion than other similar RPL protocol. The node “D and E” calculates the rank. The participant node “F” wants to join in the existing DODAG and it has two possibilities either E or C. According to ELRPL, the DODAG computes the rank and it is decided to select node “C” as parent node.
Figure 2. Rank caculation represention
The proposed EL-RPL rank calculation process is discussed in below the Table 1.
Table 1. EL=RPL Rank Calculation Process Node Id A B C D E
Fixed Point Value 256 546.32 537.41 842.25 842.25
Step Value 68.64 34.32 25.41 39.93 39.93
Rank increase 324 290.32 281.41 295.93 295.93
Rank R=floor(324/256)=1 R=floor(546.32/256)=2 R=floor(537.41/256)=2 R=floor(842.25/256)=3 R=floor(842.25/256)=3
3.2. Parent Selection Process In DODAG, request process is carried out in two ways. i). the participant node sends the DODAG Information Solicilisation (DIS) message to DODAG. ii). DODAG sends the DODAG Information Object (DIO) message to all participant node in periodic manner. If the participant node accepts DIO request and it wants to join in the existing DODAG, it sends the DAO control message to the parent node in DODAG. Finally, the DODAG sends the DODAG Advertisement Object-Acknowledgement (DAO-ACK) to the respective participant node. Then, the request and response based on the trickle timer interval. If the trickle time expires, the DODAG resends the DIO control messages to all participant nodes and the process is carried out the above in the same. The parent selection algorithm is given below.
Energy and Load Aware Routing Protocol for Internet of Things (S.Sankar)
260
ISSN: 2252-8814
Parent Selection Algorithm 1: Input 2: Node N, ParentNodeID, SenderNode_ParentID, BestParent_Rank=∞; 3: Output 4: Preferred_Parent (N) 5: for Preferred_Parent ∈ Parent _List do 6: Rank (Node) ←Rank (Parent_Node) + Rank _Increase; 7: Rank_ Increase ← Step+MinHopRankIncrease; 8: Step= w1× Load (Pi) + w2×BDI (Pi) +w3× ETX (Pi); 9: If BestParent_Rank>=Prefered_ParentRank (P) then 10: BestParent_Rank←Prefered_ParentRank (P); 11: end if 12: end 13: while Prefered_ParentRank (P) = BestParent_Rank do 14: SenderNode_ParentID←Preferred_ParentNodeID; 15: end
4.
PERFORMANCE EVALUATION We conducted the simulation of EL-RPL and it is an advancement of the standard version of RPL protocol. EL-RPL follows the additive property and it is evaluated the performance in terms of network lifetime, number of parent changes, remaining battery power, packet delivery ratio, number of parent changes and end-to-end delay. The simulation is conducted using COOJA network simulator. It is an open source operating system and it specifically designed for IoTdevices [19]. The simulation parameters represented in table-2. In MAC layer, RPL uses the default radio duty cycling mechanism and EL-RPL avoids the multi point -to-point data traffic in upward routing [20]. Most of the cases, the composite metric provides the better performance against single metric in RPL. Hence, the performance of EL-RPL compared with RER(BDI) RPL, OF-FL RPL. We conducted the simulation and it fine-tunes the weight values of the EL-RPL protocol.
Table 2. Simulation Configuration for Experiments Node Id Operating System Node Type Minimum DIO interval DIO interval doubling Routing Protocol MAC/Adaptation Layer Radio Environment Number of Nodes Simulation Duration Full Battery Transmission Range Data Packet Timer RPL Parameter
Fixed Point Value Contiki 2.7 Tmote sky 12 10 RPL ContikiMAC/6LowPAN Unit Disk Graph Medium (UDGM) 30 120 Hrs 3000 mJ 400*400 m2 60 sec MinHopRankIncrease=256
4.1. Performance Evaluation Metrics EL-RPL Performance is analyzed and evaluated the following metrics. a. Number of parent changes: It indicates parent changes in a DODAG at given amount of time. b. Network lifetime: Network lifetime as the duration of time until the first node fails in a network due to the battery exhaustion. c. Packet delivery ratio: Packet delivery ratio denotes that number of packets received successfully at receiver and number of packets send by sender. d. Remaining energy: It indicates the average remaining energy present in all nodes of network. e. End-to-end delay: It indicates the duration between start to transmit the data packet and it received by the DODAG root.
IJAAS Vol. 7, No. 3, September 2018: 255 – 264
IJAAS
ISSN: 2252-8814
261
f. Number of hopcount: It represents that the number of jump or hopcount from parent node to DODAG root. 4.2. Evaluation Results We conducted the simulation and evaluate the performance of EL-RPL with RER(BDI) and composition routing metric using lexicographic and additive property. 4.2.1. Average Number Of Parent Changes We measured an average number of parent changes in EL-RPL and it is compared with other similar RPL protocol such as RER(BDI) RPL and OF-FL RPL. Fig.3 shows that the number of parent changes with different composite metric based RPL protocol. X-axis indicates the different composite metric based RPL protocol and Y-axis indicates the average number of parent changes. The average number of parent changes in RER (BDI) RPL, OF-FL RPL and EL-RPL are 0.29, 0.27 and 0.25 respectively. The number of parent changes value indicates the network stability and it is considered the lower value for stability network. However, EL-RPL provides the better network stability than RER (BDI) RPL and OF-FL RPL.
Figure 3. Various RPL protocol vs. Average number of parent changes
4.2.2. Network Life Time In this scenario, we considered a DODAG consists of 1 DODAG root and 30 RPL router or host node. Through our simulation, we observed the node remaining energy of the entire node in a network. The EL-RPL network lifetime is compared with RER(BDI) RPL and OF-FL RPL. The entire simulation is conducted up to 120 hours. Fig.4 shows that the average remaining energy of the network nodes. In this simulation, we have set the throughput as 1 packet per minute. After the simulation, the total number of alive node count as RER(BDI) RPL is 24, OF-FL RPL is 26 and EL-RPL is 27.In EL-RPL, the remaining energy distribution ratiocontains1 node have a remaining energy between 10-15%, 7 nodes have remaining or current energy between 15-20%, 8 nodes have remaining or current energy between 20-25%, 7 nodes have remaining or current energy between 25-30% and 4 nodes have remaining or current energy between 3035%. In EL-RPL, Load is one of the routing metric and it avoids the data traffic over the network. So ELRPL improves the network lifetime against RER(BDI) RPL and OF-FL RPL.
Figure 4. Comparison of remaining energy vs Number of node
Energy and Load Aware Routing Protocol for Internet of Things (S.Sankar)
262
ISSN: 2252-8814
4.2.3. Average Packet Delivery Ratio Packet delivery ratio indicates the relaiblity of the node present in the network. X-Axis indicates the RX ratio and Y-Axis indicates the packet delivery ratio (PDR).In this simulation, we take into account the Transmit ratio (TX) value is 100 and Receive ratio (RX) is dynamically changed. Fig. 5 shows that packet delivery ratio with respect to different RX value. The proposed EL-RPL protocol PDR value compared between RER(BDI) RPLand OF-FL RPL. RER(BDI) indicates that packet delivery ratio values are 60, 63, 72, 83, 87,89 and 91 for the RX values 40, 50,60,70,80,90 and 100 respectively. OF-FL RPL indicates that packet delivery ratio values are 64,65,74,83, 89,90 and 92 for the RX values 40, 50,60,70,80,90 and 100 respectively. Likewise, EL-RPL indicates that packet delivery ratio values are 68,68,80,84, 91,93 and 95 for the RX values 40, 50,60,70,80,90 and 100 respectively. Finally, EL-RPL provides the better performance nearly 2-5% against RER(BDR) RPL and OF-FL RPL.
Figure 5 Packet delivery ratio vs. RX ratio
4.2.4. Average End-to-End Delay The end-to-end delay is calculated from number of hop count from parent node to the DODAG root node. The delay is measured the time between sender sends the data and receiving the acknowledgement from receiver.
Figure 6 Number of hops vs. Average end-to-end delay (ms)
IJAAS Vol. 7, No. 3, September 2018: 255 – 264
IJAAS
ISSN: 2252-8814
263
Figure 6 shows that the average end-to-end delay of RER(BDI) RPL, OF-FL RPL and EL-RPL. This work is considered the average end-to-end delays undertake 1 sec. Neverthless, load is a prominent metric in EL-RPL and it reduces the average end-to-end delay over the network. Moreover, the figure shows that EL-RPL has less latency and it compared with RER(BDI) RPL and OF-FL RPL. EL-RPL keeps the maximum delay undergo a data packet is 0.5 sec vs. 1sec for RER(BDI) RPL . 4.2.5. Average End-to-End Delay Figure 7 shows that the comparison of average hop count in RER(BDI) RPL, OF-FL RPL and ELRPL. X-axis indicates network size and Y-axis indicates average number hop count. In EL-RPL, we used load is one of the metric and it is calculated from cumulative children count from parent node to DODAG root. However, EL-RPL allows lower hop count than RER(BDI) RPL and OF-FL RPL. In our simulation, the maximum hop count is 5 among the 30 nodes present in the network. In EL-RPL, the network size is 30 and number of hop count is 3 vs. 5 hop count for RER(BDI) RPL. Finally, EL-RPL reduces the number of hop count within the DODAG
Figure 7 Network size vs. Average number of hop count
5.
CONCLUSION In this paper, we proposed a new energy and load aware routing protocol (EL-RPL). It deploys an objective function, which considers the metrics load, BDI and ETX, to calculate the DODAG rank, where other existing work does not consider the load metric in combination with ETX and BDI.The route with minimum value for the objective function will be the shorter route with less traffic, which is selected to send the data to DODAG root.Using the COOJA simulator, we compared the performance of EL-RPL with RERBDI RPL and OF-FL RPL. The simulation result shows that EL-RPL provides the better performance in terms of network lifetime, packet delivery ratio and end-to-end delay compared to RERBDI RPL and OF-FL RPL. As part of future work, it is planned to apply mobility to the nodes in Low Power and Lossy Networks (LLN) and deploy it in real time environment.
REFERENCES [1] [2] [3] [4]
[5]
A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, M. Ayyash, Internet of things: A survey on enabling technologies, protocols, and applications, IEEE Communications Surveys & Tutorials 17(2015) , 2347-76. Sankar, S., and P. Srinivasan, Internet Of Things (Iot): A Survey On Empowering Tsechnologies, Research Opportunities And Applications, International Journal of Pharmacy & Technology 8(2016), 26117-26141. Gaddour, Olfa, Anis Koubâa, and Mohamed Abid. "Quality-of-service aware routing for static and mobile ipv6based low-power and lossy sensor networks using RPL." Ad Hoc Networks 33 (2015): 233-256. Hassan, Ali, Saleh Alshomrani, Abdulrahman Altalhi, and Syed Ahsan, Improved routing metrics for energy constrained interconnected devices in low-power and lossy networks, Journal of Communications and Networks 18 (2016), 327-332. Bhattacharjee, Sanghita, and Subhansu Bandyopadhyay. "Interference aware energy efficient multipath routing in multihop wireless networks." Journal of High Speed Networks 20.4 (2014): 263-276.
Energy and Load Aware Routing Protocol for Internet of Things (S.Sankar)
264 [6]
[7] [8] [9] [10]
[11]
[12] [13] [14] [15]
[16] [17]
ISSN: 2252-8814
Tripathi, Joydeep, Jaudelice Cavalcante de Oliveira, and Jean-Philippe Vasseur, A performance evaluation study of rpl: Routing protocol for low power and lossy networks, Information Sciences and Systems (CISS), 44th Annual Conference on. IEEE, (2010). T. Winter, P. Thubert, A. Brandt, J. Hui, R. Kelsey, P. Levis, K. Pister, R. Struik, JP. Vasseur and R. Alexander, RPL: IPv6 Routing Protocol for Low-Power and Lossy Networks, RFC6550, 2012. Zahariadis, Theodore. "Design guidelines for routing metrics composition in LLN." (2011). Iova, Oana, Fabrice Theoleyre, and Thomas Noel. "Improving the network lifetime with energy-balancing routing: Application to RPL." Wireless and Mobile Networking Conference (WMNC), 7th IFIP. IEEE, 2014. Kamgueu, Patrick Olivier, et al. Energy-based routing metric for RPL. Diss. INRIA, 2013.P. Karkazis, T. Panagiotis, C.L. Helen, S. Lambros, P. Ioannis and Z. Theodore, Design of primary and composite routing metrics for RPL-compliant Wireless Sensor Networks, Telecommunications and Multimedia (TEMU), International Conference on. IEEE, 2012. P. Karkazis, T. Panagiotis, C.L. Helen, S. Lambros, P. Ioannis and Z. Theodore, Design of primary and composite routing metrics for RPL-compliant Wireless Sensor Networks, Telecommunications and Multimedia (TEMU), International Conference on. IEEE, 2012. Kim, Hyung-Sin, et al. "Load balancing under heavy traffic in RPL routing protocol for low power and lossy networks." IEEE Transactions on Mobile Computing 16.4 (2017): 964-979. Zhu, Yanmin, et al. "On deploying relays for connected indoor sensor networks." Journal of Communications and Networks 16.3 (2014): 335-343. D.S. De Couto , D. Aguayo , J. Bicket, R. Morris, A high-throughput path metric for multi-hop wireless routing, Wireless Networks 11 (2005), 419-434. M. Qasem, A.Al-Dubai, I.Romdhani, B.Ghaleb and W.Gharibi., A new efficient objective function for routing in Internet of Things paradigm, Standards for Communications and Networking (CSCN), IEEE Conference on. IEEE, 2016. Ali, Hazrat., A performance evaluation of rpl in Contiki, (2012). F. Osterlind, A. Dunkels, J. Eriksson, N. Finne and T. Voigt, Cross-level sensor network simulation with cooja, Local computer networks, proceedings 31st IEEE conference on. IEEE, (2006).
IJAAS Vol. 7, No. 3, September 2018: 255 – 264
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 265~272 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp265-272
265
Angular Symmetric Axis Constellation Model for Off-line Odia Handwritten Characters Recognition Pyari Mohan Jena1, Soumya Ranjan Nayak2 1
Departement of Computer Science and Engineering, College of Engineering and Technology, Bhubaneswar, India 2 Departement of Computer Science and Engineering, K L University, Andhra Pradesh, India
Article Info Article history: Received Oct 12, 2017 Revised Feb 20, 2018 Accepted Mar 26, 2018 Keyword: Feature Vector Optical Character Recognition Pattern Recognition Random Forest SVM
ABSTRACT Optical character recognition is one of the emerging research topics in the field of image processing, and it has extensive area of application in pattern recognition. Odia handwritten script is the most research concern area because it has eldest and most likable language in the state of odisha, India. Odia character is a usually handwritten, which was generally occupied by scanner into machine readable form. In this regard several recognition technique have been evolved for variance kind of languages but writing pattern of odia character is just like as curve appearance; Hence it is more difficult for recognition. In this article we have presented the novel approach for Odia character recognition based on the different angle based symmetric axis feature extraction technique which gives high accuracy of recognition pattern. This empirical model generates a unique angle based boundary points on every skeletonised character images. These points are interconnected with each other in order to extract row and column symmetry axis. We extracted feature matrix having mean distance of row, mean angle of row, mean distance of column and mean angle of column from centre of the image to midpoint of the symmetric axis respectively. The system uses a 10 fold validation to the random forest (RF) classifier and SVM for feature matrix. We have considered the standard database on 200 images having each of 47 Odia character and 10 Odia numeric for simulation. As we have noted outcome of simulation of SVM and RF yields 96.3% and 98.2% accuracy rate on NIT Rourkela Odia character database and 88.9% and 93.6% from ISI Kolkata Odia numerical database. Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Soumya Ranjan Nayak, Departement of Computer Science and Engineering, K L University, Greenfields, Vaddeswaram, Guntur, Andhra Pradesh, 522502-India. Email: nayak.soumya@kluniversity.in
1.
INTRODUCTION In the era of digital image processing, the character recognition is one of the significant and useful emerging research topics is the area of pattern recognition. The main intend of character recognition is to translate human readable character to machine readable code so that machine can efficiently recognize the character. There are mainly two broad category of character recognition system are found such as offline and online recognition process. In case of online character recognition process, it represents the two dimensional co-ordinates of successive points of the handwriting as a function of time are stored in particular order described by [1].where as in case of the offline handwriting, only the completed writing is available as an image describe by [2]. In this paper, our research intend confined with offline handwritten character recognition. Our recognition stage comprises of three broad stages including acquisition, feature extraction and classification step. Beside that a recognition system mostly depends upon a well-defined feature extraction procedure along with a good classifier, in order to achieve high success rate [3]. In order to achieve a good recognition system for handwritten format is quite still challenging because of variation in Journal homepage: http://iaescore.com/online/index.php/IJAAS
266
ISSN: 2252-8814
writing skills, shapes and orientation. Various approaches are followed up by different researcher to various scripts like Arabic, Chinese, and English etc are reported [4]. Basically Odia Script language is one the language which is derived from Devangiri scripts. It is one of the regional languages of India, mostly spoken at eastern part (Odisha) and some south, north part of India. To achieve a good accuracy of recognition for handwritten characters of Odia character is quite impressive. Though a good number of works has done for Indian regional languages but a less in number related to Odia script. In these past recent years different authors make an attempt for analysis with respect to Odia scripts are reported in [5]. The feature extraction technique for recognition of handwritten character is a challenging task in the research field of pattern recognition. In this regard a large number of feature extraction technique and classification algorithm have been presented in recent year described by [6]. Several character recognition technique of different language is found in many literatures [7-9]. In line to character recognition the extensive survey has been reported based on different kinds of feature extraction technique [10]. In this survey paper, author reported different feature extraction technique applied on Template matching, Projection histograms, Deformable templates, Contour profiles, Unitary image transforms, Zoning, Graph description, Zernike moments, Spline curve approximation and Fourier descriptors has been applied on gray level character, Binary character, Character Contour , character skeleton and character graph image representation form in the pre processing steps. As the Indian language is concerned, the optical character recognition plays a vital role now days. In this paper we have made an attempt to design a novel approach that efficiently recognize the odia character by implementing angular measurement and Euclidian distance by taking the midpoint from the axis, which was generated by taking the midpoint of two boundary edge of row symmetric axis as well as column symmetric axis to the centre of the images. Odisha state, so far has been able to uphold the pride of having the largest number of palm leaf manuscripts (over 20,000 manuscripts) in the world. [11]. Million books would have been printed from starting where “New Testament” that got printed in 1809 was first published.[12]. Odia got classical status except 5 other Indian languages on the basis of its literary heritage following approval of the Union cabinet.
2.
RELATED BACKGROUND WORK Odia script has been extracted using Bhrami scripts and one of the most ancient languages among Indian regional language most spoken eastern part of India basically in state Odisha, West-Bengal, Gujarat etc. The most important scenario of this language that it has no lower and upper case format. Here in the script is has no upper case lower structure. A certain well-defined approaches are adopted by different researchers to achieve high recognition rate. Recognition is the process of accepting the unknown samples of handwritten character image or words and then proceeds into a pattern recognition problem for testing. Recognition process can be achieved either in three important way, which is described as template matching, statistical technique and neural network techniques. These character recognition approaches uses either top down approaches or analytical strategies for recognition. Template matching is the simplest form of training and recognition. Here is the idea is to match the stored predefined prototype with the unknown handwritten characters. In this matching technique only selected pixel are compared with data samples and ruled based decision tree analysis. Rule based decision technique were used by chaudhuri et. at in 2002 [13]. Statistical technique considered as more effective while recognition of Odia characters. In this regard obaidullah et al [14] in 2014 uses the linear logistic regression model by using higher order statistical decision model to provide better performance rather than the linear model in performance. In 2007 pal et al [15] used quadratic function for classification is based on Bayesian estimation. In 2009 and 2005 a similar techniques of pseudo Bayesian estimation technique was adopted by waxabyashi et al [15] , and roy at al[16] for odia handwritten numerical recognition. They used conventional quadratic discriminant function. In 2006 Hidden Markov Model (HMM) was purposed by Bhowmik et al [17]. This is used non homogeneous quadratic method for training and recognition of handwritten numerical. In 2014 Dash et al [18]-[19] have adopted a discriminative learning based quadratic discriminant classifier (DLQDF) and Non-redundant Stockwell transform based feature extraction for handwritten digit recognition. Neural network is the parallel processing method having interconnection of neurons inside this technique. It perform computation at higher speed in comparison with statistical and template matching. Neural network can be performed either in two ways like feed forward network (FFNA) and back propagation network (BPNN). In 2013 mishra et al [20] perform the classification with BPNN and got a high accuracy of 90.44 percentage. In 2011 Majhi et al [21] authors have proposed a nonlinear neural network classifier it is an analogy of functional link artificial neural network (FLANN) classifier. In 2012 Chanda et al [22] propose a method for writer identification from Odia handwritings which uses the SVM for classification. In 2015 kalyan et al[23] purposed BESAC symmetric axis constellation model using classifier SVM, Nearest neighbour and Random forest having accuracy 98.90, 99.48, 96.76 percentage respectfully. The details comparison of recognition accuracies are described in below Teble. 1. I JAAS Vol. 7, No. 3, September 2018: 265 – 272
IJAAS
ISSN: 2252-8814
267
Table 1. List of All Features in Recognition of Odia Characters with Accuracy Method Pal et al. (2007a)
Features Gradient + curvature
Classifier MQDF
Database Odia basic characters on IITBBS
Accuracy (%) 94.60
Weighted gradient
MQDF
Odia basic characters on IITBBS
95.14
Directional
Quadratic
94.12
Bhowmik et al. (2006)
Scalar
HMM
Bhowmik et al. (2006)
Scalar
HMM
Mishra et al. (2013)
DCT
BP NN
Mishra et al. (2013)
DCT
BP NN
Dash et al. (2014b)
Hybrid topology
DLQDF
Dash et al. (2014a)
Stockwell transform
k-Nearest neighbor
Dash et al. (2014b)
Hybrid topology
DLQDF
Dash et al. (2014a)
Stockwell transform
k-NN
BESAC
Random forest SVM Nearest neighbor
Odia numerals on IITBBS Odia numerals on ISI Kolkata Odia numerals on IITBBS Odia numerals on ISI Kolkata Odia numerals on IITBBS Odia numerals on ISI Kolkata Odia numerals on ISI Kolkata Odia numerals on IITBBS Odia numerals on IITBBS Odia numerals on ISI Kolkata Odia numerals on IITBBS
97.30 98.56 98.90
Wakabayashi et al. (2009)
Roy et al. (2005)
Kalyan S Dash et al. (2016)
Kalyan S Dash et al. (2016)
BESAC
Random forest SVM Nearest neighbor
90.50 91.26 92.00 90.44 98.50 98.80 98.28 99.00 98.44 99.02 99.35
3.
PROPOSED HANDWRITTEN CHARACTER RECOGNITION SYSTEM In this section, we have made a novel technique that efficiently recognizes the Odia character. The complete proposed method is described graphically in Figure 1. These proposed systems are carried out by including the certain steps like Image like Image Acquisition, Pre-processing, Feature Extraction, and Classification. The details discussion can be made in several sub-chapters in subsequent section.
Figure 1. Schemitic Model of Recognition Model
3.1. Image Acquisition As per our proposed methodology described above we have consider the standard database of odiya character named as Nit Rourkela Odia database, which was developed at NIT, Rourkela by Mishra et al. [20]. In this database they had composed of various 15040 numbers of images of both character and numerals. In this research analysis, we have considered 47 characters having 200 numbers of samples for our experimental study. The modern Odia script consists of 12 vowels, 3 vowel modifiers, 37 simple consonants, 10 numerical digits and about 159 composite characters (juktas). Odia script is a curved appearance of writing patterns on Angular Symmetric Axis Constellation Model for Off-line Odia Handwritten... (Pyari Mohan Jena)
268
ISSN: 2252-8814
Palm leaves which have been secure from tear if writer uses too many straight lines. Table. 2 describes about Odia characters with their Phonetics.
Table 2. List of All Features in Recognition of Odia Characters with Accuracy Phonetics a
Letter
Phonetics Ka
letter
Phonetics Da
Letter
Phonetics Ma
letter
Phonetics Ek
Aa
Kha
Dha
Ja
Due
E
Ga
Na
Ra
Tini
Ee
Gha
Ta
La
Chari
U
Una
Tha
Laa
Paanch
Uu
Cha
Da
Ya
Chha
Ru
Chha
Dha
Sha
Saath
Ruu
Ja
Na
Sha
Aath
A
Jha
Pa
Sa
Na
Ae
Nya
Pha
Ha
Shun
O
Ta
Ba
Khya
Ou
Tha
Bha
letter
3.1. Image Pre Processing Pre-processing is an important step during the image acquisition process in order to get higher accuracy result by means producing noise free images as well as free of skewnes. In this analysis step, our pre-processing steps are done by using different phases like noise reduction, Normalization, skew or slant adjustment and segmentation. The details description of these pre-processing steps are summarised in the following sub sections. 3.2. Noise Reduction Noise is the unwanted output comes with the pixel intensity value in the scanned document whereas reduction of noise is the process of eliminating spurious points due to the poor sampling rate of the scanner. 3.3. Normalization Normalization is the process of separating what data we get and what data we required. We adopt binarization as the intensity normalization in the pre-processing step. Then we adjust the size of each sample as 81*81 dimensions for size normalization. 3.4. Skew or Slant Adjustment Skewness in the image undergoes some rotation of scanned image. This is very important to eliminate rotation in the pre-processing step. Rotation can be eliminated by implementing the elimination of degree of tilt angle and rotation of opposite direction. 3.5. Segmentation Segmentation is the process of separation of text and non text area in the scanned handwritten document. It is the challenging part for pre-processing there are 2 types of segmentations can have in the preprocessing steps, External segmentation perform separation of paragraph, words or sentence from scanned documents whereas internal segmentation is the process of separation character from each word. 3.6. Feature Extraction Feature extraction techniques are used to evaluate the uniqueness of each character image by which they differs from the rest character images. In this section we have implemented a unique algorithm for evaluation of feature vector by considering the mean distance of row, mean angle of row, mean distance of column and mean angle of column from centre of the image to midpoint of the symmetric axis respectively. All the operations were performed over skeletonized image of handwritten characters. Our feature extraction
I JAAS Vol. 7, No. 3, September 2018: 265 – 272
IJAAS
ISSN: 2252-8814
269
implementation is mainly focuses on five unique steps, and this is considered as the key feature values of our proposed system. The details of the five steps are described as follows: 1. This unique technique extracted the feature of image according to binary image point passing through the angle of 45°, 135°, 180°, 225°, 270°, 315°, 360° respectively from the centre of the image represented in Figure 3. 2. Extract the point positions of the sample image having passes through these angles represented in Figure 4. 3. Plotting all the points create a unique polygon shaped image for each sample represented in Figure 5. 4. Find row symmetry axis and column symmetry axis based on Figure 6 & 9 respectively. 5. Estimate mean angle and distance of both the symmetry axis based on Figure 8 & 11 respectively. For the above all empirical calculation of our implementation methodology, we have developed two algorithms which were depicted in Algorithm I and Algorithm II respectively.
The proposed character recognition method are divided the images into two parts of operation and the first part operation included a chord that is drawn from each boundary pixel to straight of boundary pixel in row wise and the second part consisting a chord that is drawn from each boundary to straight of its boundary pixels in column wise and the complete description of these two steps are discussed in Algorithm 1 and Algorithm 2 respectively. For N no of boundary pixel and K no of boundaries; the number of available cord is (N/2)*k in row wise and column wise. However, we discard those boundary chords which having less than 3 pixels in that cords. The remaining cord is called row chords and column chords because these chords are present in the same row and same column. These chords are parallel present in row chords, which is Angular Symmetric Axis Constellation Model for Off-line Odia Handwritten... (Pyari Mohan Jena)
270
ISSN: 2252-8814
presented in Figure 6 and the cord are vertically present in column chords presented in Figure 9 respectively. In our subsequent step we have group the row chords and column chords, in order to find symmetry axis from parallel row chords and vertically column chords. The midpoint of the parallel row chords and vertical column chords could generate a number of row symmetry axes as well as column symmetry axes which are presented in Figure 7 and Figure 10 respectively. In order to find the accurate symmetry axes to represent the perceptual parts, we propose midpoint criteria of the respective chords to be verified in the following method. Yr−Xr
Midpoint of row chords= 2 Where r=(1,2,3,....n) No of boundaries Yr =Y point of r ’th row boundary point Xr =X point of r ’th row boundary point Yc−Xc Midpoint of column chords= 2 Where c=(1,2,3,....n) No of boundaries Yc =Y point of c ’th column boundary point Xc =X point of c ’th column boundary point After successfully analysed of above implementation model, we have obtained the final set of row and column symmetry axis. Then we have developed the constellation model according to their relative symmetric axis pixel position and midpoint pixel angle from centre of the image. This constellation model generates two set of parameter for each row symmetry axis and column symmetry axis. Where one parameter show mean value of relative distance of every symmetry axis pixel position to centre of the image and other parameter shows the angle between the midpoints of the symmetry axis to the centre of the image. Thereafter we found four parameter of each image having two parameter each for row symmetry axis and column symmetry axis presented in Figure 8 and Figure 11 respectively.
Figure 2. Sample image
Figure 3. Skeletonized images
Figure 4. Angle Pixel point extraction
Figure 5. Plotting pixel point
Figure 6. Row symmetry axis
Figure 7. Midpoint of row symmetry axis
I JAAS Vol. 7, No. 3, September 2018: 265 – 272
IJAAS
ISSN: 2252-8814
Figure 8. Angle and distance from centre
Figure 9. Column symmetry axis
271
Figure 10. Midpoint of column to midpoint in row symmetry
Figure 11. Angle and distance from centre to midpoint in column symmetry axis
3.7. Classification Classification is one of the important phases of any recognition model. According to our implementation model we have adopted a two way strategy for recognition. In this regard we have chosen two well liked classifier namely support vector machine (SVM) [24] and random forest tree (RFT) [25] for recognition of handwritten characters. After evaluating the desired key feature values we process these vector to classifier separately and noted down the overall recognition accuracy. We have first evaluated the SVM [16] classifiers which are multi-class classifier and supervised one. Secondly random forest tree [25] which is work based on the idea of bagging and random selection of features. All the performance was listed depending upon the value of the mean square error. And tells about which classifier is the best one.
4.
RESULT AND DISCUSSION All the implementation of our proposed method were carried out with the system having specification with windows 8, 64 bit operating system, and Intel (R) i7 – 4770 CPU @ 3.40 GHz, and all the simulation is done through matlab14 (a) over a standard database. As per standard Database containing 200 samples from each of the 47 categories named as NIT Rourkela Odia database and considering numeric database from ISI Kolkata having 16 samples from each of the 10 categorised. After getting the four key feature vector values from each database as mean distance of row, mean angle of row, mean distance of column and mean angle of column from centre of the image to midpoint of the symmetric axis from each image. Hence total size of input for Odia character becomes 4*9400 and numeric character becomes 4*9400 and makes these as input to well defined classifier such as SVM and random forest and also performed the validation by implementing 10 fold-cross validations to the system. Consequently all the observation was counted to certain as 75, 25 ratio as training and testing. At first SVM classifier is implemented followed up by random forest classifier. We have also made a comparison analyses among these two classifiers, and listed 93.6% as the recognition rate for SVM and 98.2% for the random forest for NIT Odia character, similarly for ISI numeric character the recognition rate for both SVM and random forest as 88.91% and 96.3% respectively.
5.
CONCLUSION In this paper, we have presented an angular symmetric constellation technique for offline Odia characters recognition. This system uses row and column symmetric axis for generating four key feature vector values from each database as mean distance of row, mean angle of row, mean distance of column and mean angle of column from centre of the image to midpoint of the symmetric axis from each image. For classification purpose, SVM and RF model is used. An experimental result from this research gives Angular Symmetric Axis Constellation Model for Off-line Odia Handwritten... (Pyari Mohan Jena)
272
ISSN: 2252-8814
satisfactory recognition result over the standard dataset, but still the development is in its infancy. Further, other techniques are to be explored for better recognition accuracy.
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]
[21] [22] [23] [24] [25] [26]
H. Swethalakshmi, et al., “Online handwritten character recognition of devanagari and telugu characters using support vector machines,” 2006 International Workshop on Frontiers in Handwriting Recognition. 2006. U. Pal, et al., “A system for off-line Oriya handwritten character recognition using curvature feature,” 2007 international conference on information technology, India, pp. 227-229, 2007. V. K. Govindan, and A. P. Shivaprasad, “Character recognition a review,” Pattern Recognition, vol. 23, pp. 671683, 1990. J. Mantas, “An overview of character recognition methodologies,” Pattern Recognition, vol. 19, pp. 425-430, 1986. U. Pal and B. B. Chaudhuri, “Indian script character recognition: a survey,” Pattern Recognition, vol. 37, pp. 18871899, 2004. R. Plamondon, S. N. Srihari, “On-Line and off-line handwritten recognition: A comprehensive survey,” IEEE Trans on PAMI, vol.22, pp.62-84, 2000. N. Arica and F. Vural, ''An Overview of Character Recognition Focused on Off-Line Handwriting,'' IEEE Transactions on System Man Cybernetics-Part C: Applications and Reviews, vol.31, pp. 216-233, 2001. L. M. Lorigo and V. Govindaraju, ''Offline Arabic Handwriting Recognition: A Survey,'' IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, pp. 712-724, 2006. G. Nagy, ''Chinese Character Recognition, A twenty five years retrospective,'' 1998 IEEE international conference on pattern recognition, Italy, pp.109-114, 1988. A. K. Jain and T. Taxt, ''Feature Extraction Methods for Character Recognition-A Survey,” Pattern Recognition, vol. 29, pp. 641-662, 1996. http://www.odishamuseum.nic.in/?q=dept/manuscripts P. Pattnaik, “Presentation on digitization of Odia books” in Utkal University (21 February 2014). B. B. Chaudhuri, et al., “Automatic recognition of printed Oriya script,” Sadhana, vol. 27, pp. 23-34, 2002. M. Obaidullah, et al., “Structural feature based approach or script identification from printed indian document,” 2014 IEEE international conference on signal processing and integrated networks, India, pp. 120-124, 2014. T. Wakabayashi, et al., “F-ratio based weighted feature extraction for similar shape character recognition,” 2009 IEEE international conference on document analysis and recognition, Spain, pp. 196-200, 2009. K. Roy, et al., “Oriya handwritten numeral recognition system,” 2005 IEEE 8th international conference on document analysis and recognition, South Korea, pp. 770-774, 2005. T. K. Bhowmik, et al., “An HMM based recognition scheme for handwritten Oriya numerals,” 2006 IEEE international conference on information technology, India, pp. 105-110, 2006. K. S. Dash, et al., “A hybrid feature and discriminant classifier for high accuracy Odia handwritten numeral recognition,” 2014 IEEE technical symposium, Malaysia, pp. 531-535, 2014. K. S. Dash, et al.,, “Non-redundant Stockwell transform based feature extraction for handwritten digit recognition,” 2014 IEEE signal processing and communication, India, pp. 1-4, 2014. T. K. Mishra, et al., “A comparative analysis of image transformations for handwritten Odia numeral recognition,” 2013 IEEE international conference on advances in computing, communications and informatics, India, pp. 790– 793, 2013. B. Majhi, et al., “Efficient recognition of Odiya numerals using low complexity neural classifier,” 2011 IEEE international conference on energy, automation, and signal, India, pp. 1-4, 2011 S. Chanda, et al.,, “Text independent writer identification for Oriya script,” 2012 IEEE international workshop on document analysis systems, Australia, pp. 369-373, 2012. K. S. Dash, et al., “BESAC: Binary External Symmetry Axis Constellation for unconstrained handwritten character recognition,” Pattern Recognition Letters, vol. 38, pp. 413-422, 2016. B. Xu , et al., “An improved random forest classifier for image classification,” 2012 IEEE International Conference on Information and Automation, China, pp. 795-800, 2012. C. Mitra and A. K. Pujari, “Mining Intelligence and Knowledge Exploration.Editor.Mining Intelligence and Knowledge Exploration,” 2009 Springer Lecture Notes in Computer Science, India, pp. 82-84, 2013. L. Breiman, “Random Forests,” Machine Learning, vol. 45, pp. 5-32, 2001.
I JAAS Vol. 7, No. 3, September 2018: 265 – 272
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 273~285 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp265-285
273
An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid Message Encryption Over Wsns Y .M. Wazery1, Mona A. S. Ali2 1 2
Faculty of computers & information, Minia University, Egypt Faculty of computers & information,Benha University, Egypt
Article Info
ABSTRACT
Article history:
WSN is a way of handling dangerous and hostile environments safely. It replaces human existence with nodes and units that could sustain its existence under extreme circumstances. The significance of WSN arises from the importance of the data gathered through its nodes. Due to the fact of WSN that it is open air environment, security issues must be considered, for example authentication of new units and the encryption of data transmitted between these units. This research provides a new model covering two important aspects in WSN. The first aspect is the creation of the key that will be used for the current session between a pair of nodes. In this step the research introduces the intuitionistic fuzzy sets to handle the WSN criteria simultaneously and efficiently, in order to decide the exact key length required depending on the status of the network parameters. The second aspect is the distribution of the key between the units desiring communications. This phase starts by authenticating each entity to each other and to the cluster head, then one unit suggests a key and the other one confirms. It then starts communication using that key. This phase shows the hybrid cryptography applied in which the algorithm uses asymmetric encryption for authentication then uses symmetric encryption to secure the connection between the two units. Experimental results in this research could categorized also into two classes. The first class is key size model in which the proposed model compared to ordinary KNN and fuzzy model related to the determination of the key size. The proposed model shows an overall efficient way relating to decide the key size. The second class of experiments is to distribute the intermediate key efficiently; at this point the proposed model shows resilience and efficiency compared to distributing the key directly through cluster head.
Received Dec 21, 2017 Revised Mar 24, 2018 Accepted Apr 26, 2018 Keyword: Encryption Hybrid cryptography Intuitionistic fuzzy sets Key KNN WSN (Wireless sensor network)
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Mona A. S. Ali, Faculty of computers & information, Benha University, Egypt. Email: Mona.abdelbaset@fci.bu.edu.eg
1.
INTRODUCTION The study introduces WSN (wireless sensor network) with wide rang about its meaning and applications , it is not regarded as any ordinary network systems but it considers one of the most essential ways to introduce perfect and secure network service [1]. To assure that we need to provide some circumstances and follow conditions which help users to access information in fields of interest of WSN easily without any obstructions or problems. In this study also Encryption or encoding is recognized and implemented to provide a secure means of transmission and communication by knowing its origin, meaning and it`s way of working in addition to its main purpose which protects data storage.
Journal homepage: http://iaescore.com/online/index.php/IJAAS
274
ISSN: 2252-8814
In this research a new model for securing the WSN is proposed. The proposed model used to secure the creation and transmission of the secret key which is used for temporarily communication between a couples of entity. The creation of the temporarily depends on some parameters those are passed to an intuitionistic fuzzy model which decides the exact number of bits will be used under the current circumstances. After the number of bits is clearly decided; the model starts another phase in which the key is passed to a pair of units issued communication. The model starts with a communication request from a node to the Cluster Head. The CH authenticates each unit to the other then the session key is created and passed between the two units for a certain amount of time decided by the CH according to the intuitionistic fuzzy model [2]. WSN is the type of networks that is based on ad-hoc technology but provides more adequate and stable infrastructure. WSN provides so many applications both military and civilian environments. Security in WSN is emerging many researchers since a lot of attacks occurred frequently. These attacks requires continuous development of defense systems to face these attacks. The intuitionistic fuzzy sets provides a very elastic and strong methods for decision making within a very changing environment like WSN. The rest of this section provides insights on WSN, Key distribution, cryptography and intuitionistic fuzzy sets. Those topics are the orbits for this research. 1.1. WSNs A wireless sensor network (WSN) means the wireless network that consists of spatially distributed over a range of autonomous devices by using sensors to provide the ability for controlling the environmental conditions. A WSN system incorporates a gateway that provides wireless connectivity back to the wired world and distributed nodes. The wireless protocol of choice works by depending on the application needs and requirements [3]. Applications of WSN: Engineers create WSN applications for areas involving health care, utilities, surveillance and remote monitoring. In health care field, wireless devices create less invasive monitoring for patients [4]. For services utilities such as the electricity power grid, streetlights, and outdoor water municipals, wireless sensors give a lower-cost way for collecting system health data to decrease energy usage and better manage resources. Remote controlling and control covers a wide range of applications including ways where wireless systems can sequel wired systems by decreasing costs of wiring structures and allowing many new types of measurement applications. Remote monitoring and surveillance applications include: a. Industrial large machine monitoring, saving human life from danger b. Structural monitoring for large buildings and bridges c. Environmental monitoring and assessment of air, soil , and water [5] d. Process monitoring for watching over the steps involved in the automated processes without human intervention. e. Important objects tracking Wireless technology gives many advantages that help users to make wired and wireless systems and allow users to take advantage of the greatest technology for their applications. 1.2. Components of a WSN Node A WSN node has several technical ingredients involving the radio, battery, microcontroller, sensor interface and analog circuit. When WSN technology is used, trade-offs among those composts must be kept in mind. In systems those are mainly battery-powered, the use of more frequent radio besides higher radio data rates implies more power consumption. Usually two to three years battery life is required, so most of the WSN systems today are built on ZigBee because of the low-power consumed in Zigbee, due to battery life and power management technology are evolving and due to the availability of IEEE 802.11 bandwidth, WiFi will be an interesting technology [6].
Figure 1. General form of WSN IJAAS Vol. 7, No. 3, September 2018: 273 – 285
IJAAS
ISSN: 2252-8814
275
The other technology requirement in WSN architecture is the battery itself. In addition to long life required, the size and weight of batteries must be considered as well as internationally existing standards for the shipping of batteries and the availability of battery. The low cost and wide availability of carbon zinc and alkaline batteries make them a common choice. To enlarge the battery life, a WSN node continuously wakes up and transmits data by powering on the radio and then powering it back off to keep energy. WSN radio technology must transmit a signal efficiently and allow the system to go back to sleep with lower power use. This means the processor included must also be able to power up, wake and return to sleep mode in an efficient way. WSNs Microprocessor direction involve reducing consumption of power while reserving or increasing processor speed. Much like any radio choice, the processing speed and power consumption trade-off is a key issue and concern when selecting WSNs processors [7]. This makes processors of the family x86 architecture a very hard option for any battery-powered units. The rest of this paper is categorized as follows; section one provides literature review and preliminaries for the technologies and methods used in this research; section two provides the introduced model for securing WSNs; the experimental results performed in this research is illustrated in section three; finally the conclusions drawn from this research is figured out in section four.
2. RESEARCH METHOD 2.1. Intuitionistic Fuzzy Sets Intuitionistic fuzzy set (IFS), introduced by Atanassov is considered a powerful tool to handle deal with vagueness. A prominent obvious characteristic of IFS is that it assigns to each element first a membership degree and secondly a non-membership degree, and thus, the IFS constitutes an advancement and extension of Zadeh’s fuzzy set, which assigns only a membership degree to each element [8]. Many authors have paid attention for the applications of the IFS theory. Those applications and theories has been successfully used and applied in different fields such as; logic programming [9], medical diagnosis, decision making problems etc. Recently various applications of IFS clustering and classification of artificial intelligence have appeared; for example (IFNN) intuitionistic fuzzy neural networks, (IFES) intuitionistic fuzzy expert systems, (IFML) intuitionistic fuzzy machine learning [9]. (IFDM) intuitionistic fuzzy decision making, (IFSR) intuitionistic fuzzy semantic representations etc. Intuitionistic fuzzy sets: Let a set E be fixed. An IFS A* in Е is an object having the form: A* = {〈x, µA(x), νA(x)〉 | x ∈ E},
Figure 2. Space Allocation of IFS where the functions µA(x): E → [0, 1] and νA(x): E → [0, 1] define the degree of membership and the degree of non-membership of the element x ∈ E to the set A, which is a subset of E (for simplicity below we shall write A instead of A*), respectively, and for every x ∈ E: 0 ≤ µA(x) + νA(x) ≤ 1.
2.1.1. Intuitionistic Fuzzy Sets (IFS) Model For the purpose of providing the accurate size of the intermediate encryption key, the algorithm must keep track of the rapidly changing parameters in the WSN. Hence the expected level of security depends mainly on the difficulty of breaking the secret key between each pair of communicating devices. The crucial step in the proposed model is the design of the IFS function that produces the actual size of the session key (intermediate encryption key) by processing the parameters given. This process handles five variables (Nodes Count (NC), Node Log (NL), Trusted Neighbors Count (TNC), Frequency of key Changes (FKC) and Length of Temporarily Key (LTK)) each of those variables will have a degree of membership µA(x) and a degree of non-membership νA(x) as illustrated below: 1. Nodes count(NC): a counter Intuitionistic fuzzy variable that holds the number of nodes currently An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid... (Mona A. S. Ali)
276
2.
ISSN: 2252-8814
registered to the WSN taking two fuzzy values a. Small with µA(sml) and νA(sml) b. Many with µA(ma) and νA(ma) Node log(NL): an Intuitionistic fuzzy variable that monitors the history of the node’s authentication attempts, this variable takes three values: a. Good: the node had been registered many times and causes no susceptibility with µA(Go) and νA(Go) b. Moderate: In states that the node had been registered many times but causes small number of susceptibilities with µA(Mod) and νA(Mod) c. Bad: declares that the node is a potential risk either by its self or through a BOTnet attack with µA(Bad) and νA(Bad) The node log variable will be expressed graphically as follows in Figure 3 and 4
Figure 3. NL IFR
Figure 4. NL IFR
3.
Trusted Neighbors count(TNC): counter Intuitionistic fuzzy variable that keeps track of the number of neighbors with a certain threshold of distance from the node A. that variable takes three values: a. Little: indicates that the number of neighbors is small hence the amount of attacks is relatively small, this variable is associated with two states µA(Lit) and νA(Lit) b. Medium: indicates that the number of neighbors is small hence the amount of attacks is relatively medium, this variable is associated with two states µA(Med) and νA(Med) c. Many: indicates that the number of neighbors is small hence the amount of attacks is relatively large, this variable is associated with two states µA(Man) and νA(Man)
Figure 5. TNC IFR
Figure 6. TNC IFR
IJAAS Vol. 7, No. 3, September 2018: 273 – 285
IJAAS
ISSN: 2252-8814
277
4.
Frequency of key Changes (FKC): the frequency of changing the session key which is an intuitionistic fuzzy variable, the more frequent changing key is for sure safer and provides more resilience to the security overall but still creates more processing and more resources consumption . This variable handles two values: a. S: ordinary traffic flow and small number of changes to the session key, this variable is associated with two states µA(S) and νA(S) b. F: high traffic and fast changing the session key, this variable is associated with two states µA(F) and νA(F) 5. Length of Temporarily Key (LTK): the length of the temporarily session key, which is also an intuitionistic fuzzy variable for outputting the key length. This variable handles three values: a. S: the fewest number of bits for a session key usually 16:64 bits depends on the inputs. b. M: moderate number of bits for a session key usually 64:184 bits depends on the inputs. c. L: large number of bits for a session key usually 184:512 bits depends on the inputs. The reason to handle and process those variable is to obtain the desired security by find the exact value of another variable Session Key Scale (SKS) variable with the values ranging (very low, low, normal, high, very high) the table below illustrates the intuitionistic rules applied in each case
Table 1. IFS Inputs and Outputs NC µA(sml) ~ νA(sml) µA(ma) ~ νA(ma)
Intuitionistic fuzzy inputs NL TNC µA(Go) ~ νA(Go) µA(Lit) ~ νA(Lit) µA(Mod) ~ νA(Mod) µA(Med) ~ νA(Med)
FKC µA(S) ~ νA(S) µA(F) ~ νA(F)
µA(sml) ~ νA(sml) µA(sml) ~ νA(sml) µA(ma) ~ νA(ma) µA(ma) ~ νA(ma) µA(sml) ~ νA(sml) µA(sml) ~ νA(sml) µA(ma) ~ νA(ma) µA(sml) ~ νA(sml) µA(ma) ~ νA(ma) µA(ma) ~ νA(ma) µA(sml) ~ νA(sml) µA(ma) ~ νA(ma) µA(ma) ~ νA(ma)
µA(Bad) ~ νA(Bad) µA(Mod) ~ νA(Mod) µA(Go) ~ νA(Go) µA(Go) ~ νA(Go) µA(Mod) ~ νA(Mod) µA(Go) ~ νA(Go) µA(Bad) ~ νA(Bad) µA(Bad) ~ νA(Bad) µA(Go) ~ νA(Go) µA(Mod) ~ νA(Mod) µA(Bad) ~ νA(Bad) µA(Mod) ~ νA(Mod) µA(Go) ~ νA(Go)
µA(S) ~ µA(F) ~ µA(S) ~ µA(F) ~ µA(S) ~ µA(F) ~ µA(S) ~ µA(F) ~ µA(S) ~ µA(F) ~ µA(S) ~ µA(F) ~ µA(S) ~
µA(Man) ~ νA(Man) µA(Man) ~ νA(Man) µA(Lit) ~ νA(Lit) µA(Med) ~ νA(Med) µA(Lit) ~ νA(Lit) µA(Man) ~ νA(Man) µA(Med) ~ νA(Med) µA(Man) ~ νA(Man) µA(Lit) ~ νA(Lit) µA(Med) ~ νA(Med) µA(Lit) ~ νA(Lit) µA(Man) ~ νA(Man) µA(Med) ~ νA(Med)
νA(S) νA(F) νA(S) νA(F) νA(S) νA(F) νA(S) νA(F) νA(S) νA(F) νA(S) νA(F) νA(S)
LTK S M
Output SKS Very low Low
Output Number of bits 12 – 16 24 - 32
L M S S L L M M S L M L S
Normal Normal Low High Normal High Very high Normal Very high High Low High Normal
48 - 64 48 - 64 24 - 32 128 - 160 48 - 64 128 - 160 256 - 300 48 - 64 256 - 300 128 - 160 24 - 32 128 - 160 48 - 64
Table 1 provides the basis for the fuzzificaction process implementing the IF-Then rules and all of the other steps in the process. The value of the variable SKS determines the fuzzy value for the session key length. The Defuzzification process shown in the table 1 which provides the actual length of the session key. 2.2. Key Distribution Key distribution might be defined as the process of distributing (cryptographic) keys to different parties. Usually this distribution includes techniques regarded "out-of-band", i.e. techniques that don't use the channel again of later connections to transmit keys. Alternative method for key distribution can be achieved through the relying of the distributing new keys onto the safe distribution of old keys, that's what a KDC is doing [10]. The standard meaning for distributing keys propose administration over the entire lifetime of the key. Key management and distribution is a piece of key administration, however it additionally includes key creation, key escrow (for reinforcement purposes), key erasure, key repudiation, key utilization and key trust in administration. Cryptography is likewise used to help the procedures for validating entities between sets of nodes. Authentication rules and protocols are about dissemination and administration of secret keys [11]. Key management and distribution in an appropriated environment is a usage of dispersed verification protocols. Based on this thought many key dissemination and verification conventions have been proposed. Generally, all protocols and mechanisms expect that some secret data is held at first by every management unit. Authentication and verification is accomplished by one central node exhibiting the other that manages that key. All frameworks accept that strategy condition is exceptionally unstable and is open for assault. So any message arrived from a central unit must have its authentication, integrity and freshness An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid... (Mona A. S. Ali)
278
ISSN: 2252-8814
confirmed. To accomplish these objectives, most frameworks need to depend on a confirmation server and this server ought to have the accompanying highlights [12]. Ability: An Authentication server conveys great quality session keys and disperses them to the asking for principals safely. Trustability: Authentication server keeps up a table containing a name and a private key for every unit. The secret key is utilized just to confirm unit's actions to the verification server and to transmit messages safely between customer forms and the confirmation server [9]. Key distribution and validation Protocols are isolated into two classifications to assure the confirmation of a message. To begin with class utilizes nonce and test/reaction handshake to check freshness, illustration is Needham-Schroeder Protocol. Second classification utilizes timestamps and expect that all machines in appropriated framework are clocksynchronized; case is Kerberos Protocol [13]. 2.2.1. Encryption Encryption is the transformations of electronic data from a form to another, aka cipher text, which difficult to be understood or decrypt by anyone except authorized units. The main reason and goal of encryption is to preserve the secrecy and confidentiality of digitally stored or transmitted data files through the Internet or other computer networks used. Modern ciphering algorithms play a truly crucial part in the security and assurance of electronic communication systems as they can preserve both confidentiality and the following vital key security elements [14]. Authenticity: The source of a message could be assured and confirmed. Integrity: a proof of the case that message contents have been modified or not since its transmission. Non-repudiation: the originating unit of a message cannot claim that the message does not belong to him [3]. Symmetric ciphers (single key encryption) are the type of encryption in which all entities share one secret key, for both ciphering and deciphering a file. AES is considered to be one of the most widely used single-key encryption [9]. Symmetric encryption provides a faster processing than asymmetric-key encryption, but with a drawback that the sender must somehow transforms the secret key used to encrypt the data to the other unit(s) before start using that key. This is a basic condition to securely manage and distribute huge amounts of symmetric keys implies that most cryptography models use a symmetric encryption algorithm to cipher data efficiently, on the other hand they use asymmetric encryption algorithm for the purposes of secret key transmission [15]. Asymmetric (double-key) cryptography, or public-key cryptography, this type of ciphering uses a couple of mathematically related but different keys. The first key is public (available to some or all other units) and the other key is to be private (secret). One of a commonly used asymmetric encryption is RSA algorithm, basically because both keys (public and private) can be used to encipher a transmission; in the same time only the other key from the one used to encipher a message can be used to decipher it. This condition allows a trust way of assuring not only integrity, but also the confidentiality, non-reputability and authenticity of an electronic connection [11]. 2.2.2. Key Distribution Model After deciding the suitable length of the intermediate encryption key, this length must sent in a secret way to both nodes that asking for communication. In order to do so a way of asymmetric encryption must be used. For the purpose of asymmetric encryption RSA was preferred for the implementation to obtain the public key and private key for each node within the WSN. These keys are distributed within the shake hand protocol when the node is first registering to the WSN, the hand shake protocol is to be done with the Cluster Head (CH) which is one node responsible for keeping track of each node within the WSN in its range, if this CH tries to leave the WSN or goes down for any reason it transmits all of its control information and data bases to a node with the highest Node Log in the WSN. Now another problem arises which is the way of distributing the intermediate encryption key. To distribute the session keys safely means of hierarchal communication must be used. One of the most famous and practical solution for the hierarchal communication is the PKI. A customized version of the PKI is used in this algorithm taking three stages as shown in Figure 7.
Figure 7. Key Distribution Model IJAAS Vol. 7, No. 3, September 2018: 273 – 285
IJAAS
ISSN: 2252-8814
279
a) Stage 1: node X tries to gain access to the public key of node Y, this can be done over four steps 1) Step 1 ( X requests Y’s public key from CH): node X sends a request message to the CH, requiring the public key of node Y this message is in the form Cto which means cipher to (CH) Cto(CH)=EKU(CH) (EKR(X)(id , Y))
(1)
This message contains the id of the message to prevent replay attacks and the name of the node Y. The message is encrypted two times first with the private key of node X (to assures that it come from node X since it cannot be decrypted by any key except by X’s public key which is available to the CH) then the message is encrypted with CH’s public key, which is available to every node in the WSN. The reason behind the last encryption to assure that only CH can open the message by its private key. 2) Step 2(CH decrypt X’s message): CH receives the message from X and analyze it in the form of reversing the order of the message Mfrm(X) which message from X. CH first decrypts by its own private key then with Y’s public key. Mfrm(X) = DKR(CH) (EKU(X)(id , Y)) (id,Y)
(2)
3) Step 3 (CH sends Y’s public key to X): CH encrypts a message in the form Cto(X) = (EKU(X) (EKR(CH)(id , KU(Y)))
(3)
Double encryption in this step serves as encryption of the message and authentication of (CH). 4) Step 4 (X retrieve Y’s public key from CH’s message): X decrypts the message as the following equation: Mfrm(CH) = DKR(CH) (EKU(X)(id , KU(Y))) id, KU(Y)
(4)
b) Stage 2: node X asks node Y for communication 1) Step 1 (X sends a communication request to Y): X sends a non-encrypted message to Y including X’s public key and an id to declare the timing of the message Mto(Y) = (id , KU(X))
(5)
2) Step 2 (Y sends a query to CH ): node X sends a request message to the CH, requiring the public key of node Y this message is in the form Cto which means cipher to (CH) Cto(CH)=EKU(CH) (EKR(Y)(id , KU(X)))
(6)
The message includes the id of the message to prevent replay attacks and the public key of the node X. This message is encrypted two times first with node Y's private key to authenticate Y to CH. Then the message is encrypted with CH’s public key that prevents any other node from reading the content of the message. 3) Step 3 (CH decrypt Y’s message): CH receives the message sent from Y and analyzes in the form of reversing the order of the message Mfrm(Y) which message from Y Mfrm(X) = DKR(CH) (EKU(Y)(id , KU(X))) (id, KU(X))
(7)
4) Step 4 (CH sends X’s public key to Y): CH encrypts a message containing the public key of X as a confirmation to Y in the form Cto(Y) = (EKU(Y) (EKR(CH)(id , KU(Y)))
(8)
5) Step 5 (Y confirms X’s public key from CH’s message): Y decrypts the message as the following equation: Mfrm(CH) = DKR(CH) (EKU(Y)(id , KU(X))) id , KU(X)
(9)
An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid... (Mona A. S. Ali)
280
ISSN: 2252-8814
Up to this step X and Y are both confirmed to each other and to CH that leads the flow to stage 3 in which they are communicating through a common shared key with the size given from the intuitionistic fuzzy model for key distribution that is mentioned earlier. c) Stage 3: Establish a connection between X and Y 1) Step 1 (Y sends a confirmation to X): in order to inform X that Y got the confirmation about X from CH, Y sends a message in the form Cto(X) = (EKU(X) (EKR(Y)(id , KU(X)))
(10)
This message contains the public key of X to confirm the acknowledgment about X this message will be encrypted two times first encrypted with the private key of the Y node to authenticate Y to X, then the message is encrypted with X’s public key that prevents any other node from reading the content of the message. 2) Step 2 (X decrypt Y’s message): X receives the message sent from Y and analyze in the form of reversing the order of the message Mfrm(Y) which message from Y Mfrm(Y) = DKR(X) (EKU(Y)(id , KU(X))) (id, KU(X))
(11)
Now both X and Y are confirmed and authenticated to each other so they can communicate with a common shared key that will be used for any symmetric encryption technique. 3) Step 3 (X sends the intermediate encryption key to Y): in order to start a private session between X and Y X sends an intermediate session key (IKE) to Y. Cto(Y) = (EKU(Y) (EKR(X)(id , IEK))
(12)
This message contains the IEK this message will be encrypted two times first encrypted with the private key of the X node to authenticate X to Y, then the message is encrypted is encrypted with Y’s public key that prevents any other node from reading the content of the message. 4) Step 4 (Y extracts the IKE of the message from X): Y receives the message sent from X and analyze in the form of reversing the order of the message Mfrm(Y) which message from Y Mfrm(X) = DKR(Y) (EKU(X)(id , IEK)) (id, IEK)
(13)
5) Step 5 (Y sends a hello message encrypted with the intermediate encryption key to X ): the final step before both nodes can start their communication using IEK, is that Y sends a message encrypted using the symmetric encryption algorithm using the IEK. Cto(X) = EIEK (id,“HEllo”)
(14)
By the time this message arrives to X. It is confirmed that Y got the IEK so they can start their session using that IEK.
IJAAS Vol. 7, No. 3, September 2018: 273 – 285
IJAAS
ISSN: 2252-8814
281
Figure 8. Sequence diagram for communication
3.
RESULTS AND ANALYSIS During the implementation of this research a set of experiments performed and a set of experimental results recorded. The first type of experiments was to discriminate among different tools and methods of deciding the size of the Intermediate session key as will be discussed in section 3.1. In section 3.2 the other set of experiments was in the key distribution methods was clarified. This section will compare the results of applying the proposed model to some algorithms and models used to provide the same functionality. To compare different type of algorithms two criteria are used in each case. 1. Processing time: the time required to process the nodes under consideration by the algorithm, in other words the time required to evaluate the mathematical and logical operations required by the algorithm. 2. Gained security level: the level of security provided by each algorithm this criteria will be measured through monitoring two indicators: a. False rejection: number of non-attacker nodes which the model rejected or eliminated. An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid... (Mona A. S. Ali)
282
ISSN: 2252-8814
b. False acceptance: number of attacker nodes which the algorithm did not discovered. 3.1. Intuitionistic Fuzzy Model (Key Size) The key size determination function is the first step in the proposed model. The results of the intuitionistic implementation were compared to two will know algorithms in the field, KNN and Fuzzy implementation. Results were monitored and recorded in each case regarding to the two applied criteria. 3.1.1. Processing Time The it takes the model to generate the key is referred to as the processing time, it differs from environment to another for example some applications are delay tolerant like mail transmission and FTP where some other environments does not provide any tolerance for delay such as real time environment. The concern of this research is on the first type delay tolerant applications. In the same time the processing time must be bounded and reasonably accepted in order to keep the processing capabilities intact and healthy. These results are in milliseconds and shown in Figure 9.
Figure 9. ASL of Intuitionistic Fuzzy vs. Non-intuitionistic Fuzzy Classification
Figure 9 shows the processing time for the intuitionistic fuzzy function is larger for the small number of units which is logically due to the number of parameters handled and the computation time required but as the system grows up and the number of units increases the proposed model outperforms the two other algorithms with respect to the processing time. 3.1.2. Gained security level: The gained security levels for three types are measured in terms of false acceptance and false rejections and shown in Tables 2, 3.
Table 2. False Acceptance of Intuitionistic Fuzzy vs. Fuzzy Function vs. KNN False acceptance
No. nodes Non fuzzy KNN fuzzy function Intuitionistic fuzzy function
10 2 2 1
15 5 5 3
25 8 9 4
50 12 11 7
75 18 17 8
100 25 22 10
175 45 39 17
200 55 48 22
225 60 52 25
250 70 67 90
Table 3. False Reject of Intuitionistic Fuzzy vs. Fuzzy Function vs. KNN False reject
No. nodes Non fuzzy KNN fuzzy function Intuitionistic fuzzy function
10 1 2 1
15 2 3 2
25 3 4 2
50 6 7 3
75 9 8 4
100 10 9 4
175 15 11 6
200 20 14 7
225 23 16 9
250 28 19 9
Tables 2 and 3 show how intuitionistic fuzzy decision making for the size of the key outperforms the other two models with a very significant ratio in both false acceptance and false rejections these superiorities condones the time delay obtained for the small number of units.
IJAAS Vol. 7, No. 3, September 2018: 273 – 285
IJAAS
ISSN: 2252-8814
283
3.2. Key Distribution Model The most important step in the proposed algorithm which is the decision of the key size is done with its experiments and proven that the intuitionistic fuzzy model outperforms the other two alternatives. Now it is the time to examine the next stage in the model which is the distribution of the key. One way is to provide each of the units with IEK directly through the CH which will be referred as DCH and the other alternative is to use the infrastructure provided by the algorithm shown in the proposed model which will be referred as IPM. In this section both paradigms will be compared in terms of the two common criteria. 3.2.1. Processing Time The processing time for both mechanisms is measured in milliseconds and shown in Figure 10.
Figure 10. Security of DCH vs, IPM
Figure 10 shows one weakness of the proposed model that it takes more time to distribute the key, in fact this amount of time as it is really obvious it is still negectable regarding to the security provided by the proposed model. 3.2.2. Gained Security Level One other important criterion that must be considered while developing any solution for a security model that is the gained security of the implementation, since the algorithm is capable of deciding whether to accept or reject the node and alter its log variable, then the gained security has to be in consideration. The next tables show the number of false acceptance and reject shown by implementing key distribution scenarios.
Table 4. False Acceptance of DCH vs. IPM False acceptance
No. nodes DCH IPM
10 1 1
15 3 3
25 8 4
50 15 6
75 21 11
100 30 14
175 48 18
200 57 19
225 62 21
250 65 23
Table 5. False Acceptance of DCH vs. IPM False reject
No. nodes DCH IPM
10 1 1
15 2 2
25 4 3
50 8 4
75 10 6
100 12 7
175 18 7
200 22 8
225 25 8
250 30 9
Tables 4 and 5 show that the IPM provides a far more enhanced security levels than DCH in both cases; either the decision to accept a node or to reject it. Specifically when it comes to rejection IPM provides semi stable performance as the number of nodes continues to grow other than the decaying performance of DCH which provides more false reject and acceptance than IPM.
4.
CONCLUSION The importance of the WSN comes from the fact that it could replace human existence in dangerous and hostile environment. WSN provides a scalable, easy to implement and very resilient infrastructure. An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid... (Mona A. S. Ali)
284
ISSN: 2252-8814
This research provides a novel model for securing the WSN, via making sure that the intermediate encryption key is used for short periods and distributed safely. The model illustrated could be viewed as a two phase’s paradigm, the first phase decides the number of bits required for the IEK according to the current network conditions, this number of bits is to be passed to the second phase which is the key distribution model. In the later phase each unit tries to make sure that it will communicate with the right unit, this done through securely communicate to a trusted third unit which is called CH. Experimental results performed in this research could be also categorized into two categories the first handles the key size determination in which the research shows that the use of intuitionistic fuzzy model will increase the security levels significantly, where as it provides a little bit of increment in processing time but this amount of time is relatively small compared to the advancement in security levels. The other category of the experimental results handles the IEK distribution model, in this step IPM shows advancement in the security levels that overcomes the delay it provides regarding to the processing time. For future research we are looking forward to apply some colony classification such as bee or whale to decide the key size. Additionally we will implement elliptic curve strategy for key distribution
REFERENCES [1] [2] [3] [4] [5]
[6] [7] [8]
[9] [10]
[11] [12] [13] [14] [15]
Stallings, W., & Tahiliani, M. P.. Cryptography and network security: principles and practice (Vol. 6). London: Pearson. 2014 P. Udaya Sri, S. Sravani, G. Priyanka, Security protocols for Wireless Sensor Networks, International Journal of Engineering And Computer Science, Vol. 6 , PP. 20289-20292, Feb. 2017. Y.M. Wazery and Mohamed Abd-ELfattah. PKI Session Key Distribution in WSN Using Fuzzy Rule Based System. International Journal of Signal System Control and Engineering Application, 10: 48-60. 2017 Begum, S. S., & Srinivasan, R. A Study on Properties of Intuitionistic Fuzzy Sets of Third Type. rn, 55, 7. Hanafy, I. M., Salama, A. A., Abdelfattah, M., & Wazery, Y. M. (2013). AIS Model For Botnet Detection In MANET Using Fuzzy Function. International Journal of Computer Networking, Wireless and Mobile Communications (IJCNWMC), 3(1), 95-102. 2016 Waleed Al Shehri, A Survey On Security In Wireless Sensor Networks, International Journal of Network Security & Its Applications (IJNSA) Vol.9, No.1. 2017 Garg, H. Generalized interaction aggregation operators in intuitionistic fuzzy multiplicative preference environment and their application to multicriteria decision-making. Applied Intelligence, 1-17. 2017. Ford, V., Siraj, A., & Rahman, M. A.. Secure and efficient protection of consumer privacy in Advanced Metering Infrastructure supporting fine-grained data analysis. Journal of Computer and System Sciences, 83(1), 84-100. 2017 Hanafy, I. M., Salama, A. A., Abdelfattah, M., & Wazery, Y. Security in Mant based on Pki using fuzzy function. IOSR Journal of Computer Engineering, 6(3), 54-60. 2012. Abdelmged, A. A., Saad, A. H. S., & Hussien, N.. A Combined Approach of Steganography and Cryptography Technique based on Parity Checker and Huffman Encoding. International Journal of Computer Applications, 148(2) 2016. Ejegwa, P. A., Akowe, S. O., Otene, P. M., & Ikyule, J. M. An overview on intuitionistic fuzzy sets. International journal of scientific & technology research, 3(3), 142-145 2014. Bhattacharya, J. (2016). A few more on intuitionistic fuzzy set. Journal of Fuzzy Set Valued Analysis, 2016(3), 214-222. Houssein, E. H., & Ismaeel, A. A. Ant Colony Optimization based Hybrid Routing Protocol for MANETs. Journal of Emerging Trends in Computing and Information Sciences, 6(11) 2015. Praveena, A., & Smys, S. Efficient cryptographic approach for data security in wireless sensor networks using MES VU. In Intelligent Systems and Control (ISCO), 2016 10th International Conference on (pp. 1-6). IEEE. ahmed Elsawy Walid Osamy “Genetic based algorithm for base-station position determination in wireless sensor network”, International Journal of Advancements in Computing Technology, 8(5),16-27 2016.
IJAAS Vol. 7, No. 3, September 2018: 273 – 285
IJAAS
ISSN: 2252-8814
285
APPENDIX
Algorithm 1: Key Size generation by using Intuitionistic fuzzy model Input: Node Count (NC), Node Log (NL), Trusted Neighbors Count (TNC), Frequency of key Changes (FKC) and Length of Temporarily Key (LTK) Output: length of the intermediate encryption key (in bits) (IEK) For each input variable; Calculate µA(X) ~ νA(X) Apply Intuitionistic fuzzy system conductive IF-Then rules End Algorithm 2: Key Distribution Input: intermediate encryption key (in bits) (IEK) Output: Key distributed // Stage 1 1. X sends a message to CH Cto(CH)=EKU(CH) (EKR(X)(id , Y)) 2. CH receives Message and analyze Mfrm(X) = DKR(CH) (EKU(X)(id , Y)) (id, Y) 3. If decryption error or id is obsolete: Increase NL(X)//consider attack Discard the message Else search entire DB for KU(Y) 4. If KU(Y) does not exist Send an error message to X Else Send a message to X contains Y’s KU Cto(X) = (EKU(X) (EKR(CH)(id , KU(Y))) 5. CH receives Message and analyze Mfrm(CH) = DKR(CH) (EKU(X)(id , KU(Y))) id, KU(Y) 6. If decryption error or id is obsolete: Discard the message Else Store Y’s public key and pass it to stage 2 // Stage 2 7. X sends a message to Y Mto(Y) = (id , KU(X))) 8. Y receives Message and analyze id , KU(X) 9. If id is obsolete: Discard the message Else Y sends a query to CH Cto(CH)=EKU(CH) (EKR(Y)(id , KU(X))) 10. If decryption error or id is obsolete: Discard the message Increase NL(Y)//consider attack Else search entire DB for KU(X) 11. If KU(X) does not exist Send an error message to Y Else send a message to Y contains X’s KU Cto(Y) = (EKU(Y) (EKR(CH)(id , KU(X))) 12. Y confirms X’s public key from CH’s message Mfrm(CH) = DKR(CH) (EKU(Y)(id , KU(X))) 13. If decryption error or id is obsolete: Discard the message // Stage 3 14. Y sends a confirmation to X Cto(X) = (EKU(X) (EKR(Y)(id , KU(X))) 15. X decrypt Y’s message id , KU(X) 16. If id is obsolete: Discard the message Else X sends the intermediate encryption key to Y Cto(Y) = (EKU(Y) (EKR(X)(id , IEK)) 17. Y extracts the IKE of the message from X Mfrm(X) = DKR(Y) (EKU(X)(id , IEK)) (id, IEK) 18. If decryption error or id is obsolete: Discard the message Else store IEK 19. Y sends a hello message encrypted with IEK Cto(X) = EIEK (id,“HEllo”) 20. X decrypts the message obtaining the id and “HEllo" confirming the IEK
An Intuitionistic Fuzzy Sets Implementation for Key Distribution in Hybrid... (Mona A. S. Ali)
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 286~297 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp286-297
286
Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM Communications Renuka Bhandari, Sangeeta Jadhav Department Of (E&TC), Dr. D.Y. Patil Institute of Engineering & Technology, Pimpri, Pune-411018 Army Institute of Technology Pune-411015, India
Article Info
ABSTRACT
Article history:
With emerge of increasing research in the domain of future wireless communications, massive MIMO (multiple inputs multiple outputs) attracted most of researchers interests. Massive MIMO is high-speed wireless communication standards. A channel estimation technology plays the essential role in the MIMO systems. Efficient channel estimation leads to spectral efficient wireless communications. The critics of Inter-Symbol Interference (ISI) are the challenging tasks while designing the channel estimation methods. To mitigate the challenges of ISI, we proposed the novel blind channel estimation method which based on Independent component analysis (ICA) in this paper. Proposed channel estimation it works for both blind interference cancellation and ISI cancellation. The proposed Hybrid ICA (HICA) method depends on pulse shape filtering and ambiguity removal to improve the spectral efficiency and reliability for MIMO communications. The Kurtosis operation is used to measure the complex data at first to estimate the common signals. Then we exploited the advantages of 3rd and 4th order Higher Order Statistics (HOS) to priorities the common signals during the channel estimation. In this paper, we present the detailed design and evaluation of HICA blind channel estimation method. We showed the simulation results of HICA against the state-of-art techniques for channel estimation using BER, MSE, and PAPR.
Received Jun 9, 2017 Revised Feb 20, 2018 Accepted Mar 11, 2018 Keyword: Blind channel estimation Error rates Independent component analysis Interference MIMO-OFDM
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Renuka Bhandari, Department Of (E&TC), Dr. D.Y. Patil Institute of Engineering & Technology, Pimpri, Pune-411018, Army Institute of Technology Pune-411015, India. Email: bhandarirenuka33@gmail.com
1.
INTRODUCTION Since from last decade, continues growth in the requirements for higher data rates on constrained resources and available bandwidth. This demand for higher data rate is resulting in the significant attention of researchers to initiate the working towards future wireless communications [1]. In future wireless communications, the resource utilization using MIMO (Multiple Input Multiple Output) is better than other methods [2] [3]. Hence massive MIMO is the crucial technology for future communication systems like 5G. The MIMO combined with transmission method like OFDM (orthogonal frequency division multiplexing) also the CDMA (code division multiple access). The critical part for MIMO systems is the design of efficient channel estimation to improve the overall communication performance and minimize the error rates. Therefore, several channel estimation methods designed in literature. The methods reported in [4-7] depend on ICA and wavelet-based channel estimation for MIMO wireless systems. The channel estimation techniques of MIMO-OFDM divided into the three types such as semi-blind, blind channel, and training based evaluation methods [4]. In training based techniques, plan known training Journal homepage: http://iaescore.com/online/index.php/IJAAS
IJAAS
ISSN: 2252-8814
287
samples used to perform the extensive channel estimations. The least square (LS) and MMSE are the cases of training based channel assessments techniques. In blind channel estimation techniques, SOS (second-order stationary statistics) or HOS techniques used for transporting maximum spectral efficiency. In the semi-blind channel estimation, the main order statistics utilized. The semi-blind channel estimation technique consolidated attributes of training based and blind based channel estimation are interconnected with MIMOOFDM respectively. The semi-blind strategy sets aside longer time for the channel estimation and subsequently increases the overall communication costs. The blind channel estimation technique is superior to other two types of channel estimation methods. Thus in this paper we focused on designing the novel blind channel estimation method. In OFDM systems, block-based symbol transmission method used; hence for 4G communications, the technique of block-based and iterative channel estimation introduced. For future wireless communication systems, same methods will be applicable. Therefore, it motivates to use blind channel estimation to obtain the initial symbol estimation and furthermore it is utilized the initial symbol evaluation to increases the more considerable dedication channel estimations. This channel estimation process repeated iteratively with soft information exchange to enhance the performance of channel estimation and data symbol estimations [8]. Further, consequently prompts the fast increment in low mobility based application and hence influence to design the blind channel estimation which requires a large number of samples based on quasi-static channel conditions. The current investigations exhibited the SOS and HOS based blind channel estimation techniques. However, none of such process equipped for interference signals cancelation for MIMO-OFDM [11]. An interference signal generates the blind channel or scalable client in MIMO-OFDM wireless systems. The HOS relies upon ICA strategy for interference cancelation in MIMO systems, yet not expressly outlined and tended to for blind channel estimation. Moreover, ISI (inter-symbol interference) likewise have the critical effect on the performance of blind channel estimation and spectral efficiency. In this paper, we proposed distinctive spectral efficient channel estimation based on modified ICA approach called HICA. The proposed HICA method designed to mitigate the challenges of blind interference, ISI cancellations, and error minimizing efficiently using the pulse shape filtering, Kurtosis, HOS functions, and ambiguity removal. We present the extensive simulation analysis for proposed blind channel estimation method using BER, MSE, and PAPR performance metrics. The PAPR evaluation for MIMO systems recently introduced in [13]-[18], hence we conducted the PAPR performance investigation in this paper. In part II, the brief literature of past works exhibited, In part III, the proposed method for blind channel estimation discussed. In part IV, the simulation outcomes illustrated. The conclusion based on the results presented in section V.
2.
RELATED WORK The channel estimation is the crucial phase of wireless communications; designing channel estimation for MIMO gains the significant attention since from last decade. We discussed some of the recent methods presented for channel estimation as well as PAPR analysis. The blind channel estimations by using the repeated index approach reported in [8]. They proposed subspace blind channel assessments design relies upon the rehashed list with the same yield from separates to ultimately strategies with a few quantities of symbols. Our work is not the same as this work as we in light of altered ICA way to deal with blind interference cancellations and error minimization. In [9], another blind channel estimation technique in the MIMO-OFDM system with the equilateral spacetime piece code investigated. This approach composed the new weighted covariance matrix of data recognize in the direction to obtain the errors in the code. All the non-scalar ambiguities are resolved using method reported in [9], but failed to address the challenges of common signal estimation and prioritizing the estimated common signals. Further, the SOS based blind channel estimation method reported in [10] based on the novel blind recursive approach in MIMO-OFDM systems. The subspace based tracking designed for the fast varying wireless channels. To get SOSs update fast, data used from time and frequency domain due to frequency correlation of this wireless channels. In [11], the blind channel estimation technique using the SOS and subspace models for the MIMOOFDM system reported. Authors efficiently utilized the null space produced by OSTBC. The method designed with the goal of minimum received blocks. The experimental results demonstrated on the single input and single output only. Our approach depends on multiple inputs and multiple output antennas for the channel estimation. Another recent approach for channel estimation reported in [12]. The blind channel estimation techniques utilized for the LTE (Long-Term Evaluation) networks. The channel estimation performed using the wavelet transform with ICA. The wavelet transforms used preceding ICA channel estimation to diminish Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM… (Renuka Bhandari)
288
ISSN: 2252-8814
the errors in approaching signals. The denoising approach was intended to deal with the blind interference cancelation. It is exploiting the ICA. The ISI cancelation techniques are not utilized in this technique. In this paper, we proposed the novel channel estimation strategy which differs from previous works of channel estimation discussed above. The proposed blind channel estimation method based on ICA aims to ISI and blind interference cancelations. At first, to mitigate the noise interference and ISI effect, we designed the pulse shape filtering to apply on modulated signals. At channel estimation phase, the source of signals estimated using the kurtosis values, then common signals are detected and then prioritize common estimation signals using 3rd and 4th order HOS. The lightweight method used for the ambiguity removal. Additionally, we presented the PAPR evaluation of proposed blind channel estimation technique for MIMO-OFDM systems in this paper. The PAPR analysis of space-time block-coded (STBC) MIMO-OFDM method for the 4G wireless networks evaluated in [13]. The numbers of techniques introduced to decreases PAPR in the (STBC) MIMO-OFDM system such as SLM, clipping and filtering and partial transmit sequence. Their simulation outcomes claim that clipping and filtering delivered effective PAPR reduction than the others methods and only SLM technique conserves the PAPR reduction in reception part of the signal. We studied some recent works reported for PAPR performance evaluation in [14]-[18].
3.
SYSTEM MODEL The system model for proposed HICA channel estimation is showing in Figure 1 for MIMO-OFDM systems.
Figure 1. Proposed MIMO-OFDM System Model
In proposed system model, at transmitter data X from users is randomly generated and then forwarded to modulation process on each symbol. IFFT and CP operations performed at each transmitter. To mitigate the effect of ISI cancellation, we applied the pulse shape filtering on IFFT data. Once symbols are ready to transmit over the Rayleigh channel through transmitter’s antennas towards receiver’s antennas, the measurement of PAPR performs on each symbol. At the receiver side, after the pulse reshaping, CP removal and FFT operations, proposed HICA method for efficient channel estimation and signal detection performed. Figure 2 showing the complete functionality of HICA block. After estimation, the reverse operations applied at the receiver side to get the original data at the receiver side. The modulation and demodulation are performed by using QPSK technique. Let’s consider S is the sample of signal, M is a transmitted signal which include the many numbers of samples. An is an attenuation for nth channel. The Factor of An is complex number. The authors investigates the pulse shaping for the deleting the ISI impacts to reduces the error rates and optimize the spectral efficiency. The HICA method for blind channel estimation consists of steps like source estimation, common signal estimation; prioritize the signals, ambiguity removal, and final signal detection at the receiver side. The functionality of complete architecture presented in the form of algorithms below.
IJAAS Vol. 7, No. 3, September 2018: 286 – 297
IJAAS
ISSN: 2252-8814
289
Figure 2. HICA channel estimation phases
Algorithm 1: MIMO Transmitter Input: X; random data n; number of transmitter and receivers Output: T1…Tn; transmitted data 1. 2. 3. 4. 5. 6. 7. 8.
Generate random data X Representation of X in different symbols as per the number of transmitter and receiver antennas X1…Xn Apply the modulation on X1…Xn. IFFT on X1…Xn Apply the Pulse shaping on input symbols using Eq. (1) X1…Xn = Pf * (X1…Xn) (where Pf is pulse shaping filter) Cyclic Prefix on X1…Xn Measure PAPR X1…Xn using Eq. (11). Transmit data X1…Xn through Rayleigh channel
Algorithm 2 is showing the process at receiver side. Algorithm 2: MIMO Receiver Input: Yr = [T1…Tn]; set of transmitted signals Output: X` = [X`1…X`n];set of received signals 1. 2. 3. 4. 5. 6.
Pulse Reshaping T1…Tn =(1/Pf * (T1…Tn)) (where Pf is pulse shaping filter) FFT on estimated signals T1…Tn CP Removal T1…Tn Channel estimation using algorithm 3 T`1…T`n = HICA (T1…Tn) X`= Demodulation T`1…T`n Return X`
The algorithm 3 shows the process for proposed channel estimation method called HICA. The proposed channel estimation approach is the iterative process to optimize the channel estimation performance and minimize the errors of estimation.
Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM… (Renuka Bhandari)
290
ISSN: 2252-8814
Algorithm 3: Channel Estimation Input: Yr = [T1…Tn]; set of received signals Output: ŜiF = [T`1…T`n]; set of estimated signals 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
Initialization HICA iterations in IT and set it = 0; Random W initialization Defining the objective function Jold ← J (W). Gradient computation of objective function using Eq. (4). W updating according to negative gradient direction, W ←W − µJw. W normalization according to unitary constraint, W ←W / W . If (Jold − J (W) < ε ) then go back to step 2 end if (where, Yr is the Set of signals estimation ŝ [n] = WH x Yr set of received signal) Form every estimated signal ŝ [n] as vector which represented by Vit[n] If it <1 goes to step 2 else, continue. Find the set of the common vectors for all runs of algorithm up to itth run. If (no common vectors) Goto step 2 Else break end if Apply Eq. (5) to prioritize the common estimated signals with J (ŝit). Selection of desired signals (m) with largest J (ŝit) in order performs the blind interference cancellation. Ambiguity Elimination on estimation common signals using Eq. (9). ŜiF = [T`1…T`n]; Return estimated signals ŜiF
Algorithm 3 HICA In the proposed method, we first enable the pulse shaping techniques in each user symbol in the order to minimize the internal symbol interference at the transmitter side. Pulse shape is simple ISI cancellationtechniques. We utilized the square root increased cosine filter in sampling region to perceive the pulse shaping filter. Table 1 shows the symbols used by proposed channel estimation method.
Table 1. List of Symbols Symbol gt (t) gr (t) T R L iR iT W v K
Meaning Transmitter side pulse shape filter for each symbol Receiver side pulse shape filter for each symbol Number of transmitters Number of Receivers Number of taps ith receiver ith transmitter Separation Matrix non-gaussianity for random variable Kurtosis
IJAAS Vol. 7, No. 3, September 2018: 286 – 297
IJAAS
ISSN: 2252-8814
291
The composite channel is represented as T * R with matrix H (t). The (iR, iT) channel using pulse phase filtering is represented by: ℎ𝑖 𝑅′ 𝑖 𝑇 (t) = ℎ𝑖 𝑅′ 𝑖 𝑇 ′ 𝑐 (𝑡) ∗ 𝑔𝑡 (𝑡) ∗ 𝑔𝑟 (𝑡)
(1)
𝐾 [𝑠] = 𝐸[|𝑣|4 ] − 2(𝐸[|𝑣|2 ])2 − 𝐸[𝑣. 𝑣]𝐸[𝑣 ∗ 𝑣]
(2)
𝑊 = {min𝑊 𝐽(𝑊) = ∑𝑀 𝑗=1 𝐾[ ŝ[𝑛]]
(3)
Where hi R' iT'c (t)is the (iR, iT) element of H (t). Here the channel can be represented as the L tap FIR filters array for blind channel estimation [19]. After the designing of pulse shaping filter which we applied to time domain (i.e., after IFFT operation) and perform the subsequent tasks of transmitter side. The data is transmitted from the transmitter’s antennas towards the receiving antennas via Rayleigh channel. At receiver end, the channel estimation is performed after the pulse re-shaping, FFT and cyclic prefix removal operations (Algorithm 2). For the channel estimation, we proposed the algorithm 3 called HICA. As observed in algorithm 3, initially signals sources from the receiver analysis mixture are determine by using the partition of matrix W. The proposed channel estimation algorithm is based on the iterative process for excellent channel estimation, thus at each iteration, the order of initial signal estimation is varying due to the random initialization of separation matrix. But, if the important information discovered in estimated signals, then it will be utilized for the iteration. The identified important information called as common estimated signals. The HOSs 3rd and 4th order derivatives applied to the common estimated signals to prioritize them. The outcome of HOS is the selected desired signals (m) by canceling the interference components from the signals. The objective function based on W introduced based on the principle of non-gaussianity maximizing observation signals. The outcome of non-gaussianity for random data (v) containing the complex data which is measured by the operator Kurtosis K[s]:
Where (.)* is represents the complex conjugate. Before prioritising the common estimation signals, the Kurtosis function is applied to discover the common estimation signals. The signal estimation using W based on the minimization of objective function J (W) within the unitary constraint The estimation of W is performed by the minimization of J (W) objective function within the WW H = IR unitary constraint. The unitary constraint used to alleviate the negative outcomes of kurtosis value for various modulation methods. Thus, the objective function using the estimation signals ŝ[𝑛] kurtosis values represented by:
The minimization of objective function J (W) satisfied using the gradient computation represented by: Jw =
∂J(W) ∂W
= K(W H ŝ[n])[E{ŝ[n])3 }]
(4)
Where (w) is one of the vectors from W separation matrix. According to the constraint for the objective function (Jw), the object function gradient complemented projecting W over the interval after each step performed by dividing the W by its norm. The iterative execution using different random initialization matrix W performed to evaluates the general signals at the receiver side. The two estimated signals of various computations are examine as different if the spectral angle mapper (SAM) of their vectors vi[n] and vj[n] is beyond the threshold value ε. The general signal estimation results used to prioritize the signals based on the 3rd (Eq. 5) and 4th (Equation 5) higher order statics. 1
1
𝐽�ŝ𝑞 � = � � [𝑄𝑞3 ]2 + ( )[𝑄𝑞4 − 3]2 12
48
(5)
1
Where, 𝑄𝑞3 = 𝐸�ŝ3𝑞 � = � � ∑𝑇𝑛=1(ŝ𝑞 [𝑛])3 is the 3rd order of statistics and 1
𝑇
𝑄𝑞4 = 𝐸�ŝ4𝑞 � = � � ∑𝑇𝑛=1(ŝ𝑞 [𝑛])4 is the 4th order of statistics of estimated signals, and q is the 𝑇 execution index. Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM… (Renuka Bhandari)
292
ISSN: 2252-8814
Once the signal estimated common signals prioritise, the next step of ambiguity removal applied to suppress the unwanted components from the estimated signals. Scientifically the estimated ŝi [n] is not matching to the original transmitted s[n] due to presence of ambiguity matrix A in ŝi [n]. This can be represented as: ši [n] =A Ⅹ ŝi [n]
(6)
Two indeterminacies forming the A as: A = P×D
(7)
Where P is the permutation ambiguity matrix and (D) is phase rotation ambiguity matrix. The removal of such ambiguities performs to reduce the error rates in proposed channel estimation techniques. The ambiguities suppressed by multiplying the estimated signals and remove from the original signals in first step represented by𝐿𝐹 : 𝐿𝐹 = arg min𝐿ℰ𝐺 ∥ 𝑠[𝑛] − ŝ[𝑛] 𝑥 A ∥2
(8)
ŝiF [n] = LF x ŝ[n]
(9)
After the 𝐿𝐹 computation, the multiplying it with the estimated signals as:
For the performance evaluation along with BER and MSE, we computed the PAPR for MIMO systems. The computation of PAPR is done as per given below: The PAPR of the signal, x (t), is then provided the peak prompt power to the average power, following are shown the formula, PAPR =𝑙𝑙𝑙10 ∗
𝑚𝑚𝑚0≤𝑡≤𝑇|𝑥(𝑡)|2
(10)
𝐸[|𝑥(𝑡)2 |]
Where E [.] is the predicated operator. From the central limit formulas, for most values of N, the X (t) is a real and imaginary values. The PAPR computation showing in Eq. (10) is applicable for Single Input Single Output (SISO) systems, therefore in MIMO systems; same formula is utilized to compute the PAPR for each transmitter given by: 𝑃𝑃𝑃𝑃𝑀𝐼𝐼𝐼−𝑂𝑂𝑂𝑂 =𝑚𝑚𝑚𝑖≤𝑖≤𝑛𝑡 𝑃𝑃𝑃𝑃𝑖
(11)
Where PAPRi denotes the PAPR of transmit ith antenna.
4.
SIMULATION RESULTS AND DISCUSSION The performance analysis of the innovated channel estimation technique for MIMO systems has been shown in this section against the state-of-art techniques. The simulation performed for this study under different network conditions such as varying number of sender antennas and accepted antennas. Table 2 shows the simulation parameters used to evaluate the proposed method. For comparative study, we used the conventional ICA technique [20] and recent wavelet denoising of ICA (WD-ICA) [12] method. Table 2. MIMO-OFDM Simulation Parameters FFT Size Transmit Antennas Receive Antennas Block size Sub band size SNR level Number of iterations Blind Estimation Method Modulation Technique Number of subcarriers Filter type Oversampling factor Channel Type
IJAAS Vol. 7, No. 3, September 2018: 286 – 297
64, 128, 256 2, 4 2, 4 8 20 0:5:30 10 ICA [20], WD-ICA [12] and Proposed HICA QPSK 256 Pulse shaping filter 4 Rayleigh
IJAAS
ISSN: 2252-8814
293
To check the reliability of proposed channel estimation method, we performed the simulations based on (1) changing the number of transmitter and receiver antennas such as 2x2 and 4x4 and (2) varying FFT sizes such as 64, 128, and 256. The performance measured in three parameters such as BER, MSE, and PAPR to check the trade-off between spectral and PAPR efficiency for each investigated method. 4.1. 2x2 Transmit/Receive Antennas This section presents the simulation results for 2x2 (2-transmit and 2-receiving antennas) for MSE, BER and PAPR performance metrics for each varying FFT size. The performance of MSE and BER is significantly improved by proposed HICA channel estimation method due to iterative error minimization approach using the Kurtosis and HOS functions for the common channel estimations. However, the performance PAPR is not promising for all the methods. In an average of 64 FFT size, WDICA method showed minimum PAPR as compared to ICA and proposed HICA method. The Figures 6 to 8 shows the results using the FFT size 128. The PAPR ratio increased as the FFT size increases. The Figures 9 to 11 shows the BER, MSE and PAPR results using the FFT size 256. The proposed channel estimation method was demonstrating the superior performance against the state-of-art ICA based methods for the channel estimation using varying FFT size. However, the PAPR results show the proposed channel estimation method more PAPR rate as compared to conventional ICA and WD-ICA methods. Figure 11 shows the FFT size having the impact of PAPR performance, as the FFT size increases the PAPR is also growing. The result of BER and MER proves the proposed channel estimation techniques are the dependable techniques MIMO- OFDM systems. The impacts of FFT size on BER and MSE is showing in Tables 1 and 2 respectively. Increasing FFT size leads to minimizing the BER and MSE values for all the methods as the number of subcarriers increases to transmit each symbol. However, the PAPR performance has the reverse impact as shown in table 3. Increased FFT size is showing the worst PAPR performance for all the methods. For each FFT size, the average PAPR result of WD_ICA having minimum PAPR rate. In this sub-section, we evaluated the impact of varying FFT size on BER, MSE, and PAPR performances using 2x2 transmit/receive antennas.
Figure 3. 2x2 BER analysis using FFT size 64
Figure 4. 2x2 MSE analysis using FFT size 64
Figure 5. 2x2 PAPR analysis using FFT size 64
Figure 6. 2x2 BER analysis using FFT size 128
Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM… (Renuka Bhandari)
294
ISSN: 2252-8814
Figure 7. 2x2 MSE analysis using FFT size 128
Figure 8. 2x2 PAPR analysis using FFT size 128
Figure 9. 2x2 BER analysis using FFT size 256
Figure 10. 2x2 MSE analysis using FFT size 256
Figure 11. 2x2 PAPR analysis using FFT size 256
Table 3. Average BER Performance Analysis (2x2) FFT Size 64 128 256
ICA 0.0445 0.0408 0.0368
WDICA 0.0068 0.0062 0.0065
HICA 0.0032 0.003 0.0027
Table 4. Average MSE Performance Analysis (2x2) FFT Size 64 128 256
ICA 0.0779 0.0385 0.0198
WDICA 0.0529 0.0261 0.013
Table 5. Average PAPR Performance Analysis (2x2) FFT Size 64 128 256
ICA 13.4883 16.2325 19.0180
IJAAS Vol. 7, No. 3, September 2018: 286 – 297
WDICA 13.4443 16.1961 19.010
HICA 13.4823 16.2552 19.0375
HICA 0.0064 0.0032 0.0016
IJAAS
ISSN: 2252-8814
295
4.2. 4x4 Transmit/Receive Antennas After the investigation of FFT size on channel estimation methods, to claim the reliability and scalability of proposed channel estimation, we used 4x4 transmit/receive antennas for MIMO-OFDM communications. This section presents the simulation results for 4x4 (4-transmit and 4-receiving antennas) results for MSE, BER and PAPR performance metrics for each varying FFT size. The varying transmitting and receiving antennas not having the significant impact on HICA performance. However, the HICA performance is significantly improved as compared to state-of-art methods using 64 FFT size (Figure 12 and Figure 13), using 128 FFT size (Figure 15 and Figure 16) and using 256 FFT size (Figure 18 and Figure 19). The channel estimation method should be lightweight and spectral efficiency. The results of MSE and BER show that HICA the spectral efficient way for MIMO-OFDM systems. The PAPR ratio is increased as the FFT size increases using 4x4 antennas scenario as well. The effect of increasing the antennas on MSE, BER and PAPR performances are review in Tables 6, 7 and 8 accordingly. The increasing antennas from 2x2 to 4x4 does not show the significant impact of performances, except little increase in MSE and BER rates as the compared 2x2 scenario. The purpose of evaluating the channel estimation methods using different antenna sizes is to check the reliability of the proposed method. The results show that for all the cases of varying FFT size and antenna scenarios, suggested HICA method delivered efficient channel estimation performance as compared to state-of-art methods. There is little decrease in PAPR rate using 4x4 situations as compared to 2x2 scenarios. From all the results, the performance of MSE and BER are the contrast to PAPR performance for all the channel estimation methods. Achieving the trade-off between PAPR and error rates is one of the leading research problems for MIMOOFDM systems. Finally, Table 9 shows the average processing time performance for each channel estimation method. The results of processing time demonstrate the proposed method has very less computation overhead as differentiates to both state-of-art channel estimation techniques. The simulation conducted on the I5 processor with 4GB RAM.
Figure 12. 4x4 BER analysis using FFT size 64
Figure 13. 4x4 MSE analysis using FFT size 64
Figure 14. 4x4 PAPR analysis using FFT size 64
Figure 15. 4x4 BER analysis using FFT size 128
Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM… (Renuka Bhandari)
296
ISSN: 2252-8814
Figure 16. 4x4 MSE analysis using FFT size 128
Figure 17. 4x4 PAPR analysis using FFT size 128
Figure 18. 4x4 BER analysis using FFT size 256
Figure 19. 4x4 MSE analysis using FFT size 256
Figure 20. 4x4 PAPR analysis using FFT size 256
Table 6. Average BER performance analysis (4x4) FFT Size 64 128 256
ICA 0.0525 0.045 0.04
WDICA 0.0081 0.0071 0.0062
HICA 0.0037 0.0033 0.0028
Table 8. Average PAPR performance analysis (4x4) FFT Size 64 128 256
ICA 13.2 16.17 18.99
WDICA 13.18 16.16 18.98
HICA 13.21 16.17 18.97
IJAAS Vol. 7, No. 3, September 2018: 286 – 297
Table 7. Average MSE performance analysis (4x4) FFT Size 64 128 256
ICA 0.0788 0.0394 0.0195
WDICA 0.0528 0.0261 0.013
HICA 0.0066 0.0033 0.0016
Table 9. Average processing time performance Method Time (Seconds)
ICA 49.55
WD-ICA 47.2
HICA 36.23
IJAAS
ISSN: 2252-8814
297
5.
CONCLUSION AND FUTURE WORK In this research paper, we investigate the novel hybrid blind channel estimation technique based on modified ICA technique using pulse shape filtering approach. The HICA effectively utilized the Kurtosis and HOS features to estimate and prioritize the common signals respectively. The idea of ambiguity removal helps to reduce the error rates further. The proposed HICA method designed and evaluated with conventional ICA and recent WD-ICA regarding BER, MSE and PAPR results. HICA method is showing the BER, and MSE performances significantly improved as compared state-of-art methods with the different antenna and FFT sizes using the QPSK modulation methods. However, the PAPR performance is not efficient using the HICA technique. HICA method failed to reduce the PAPR rate against the state-of-art methods. For future work, it will be interesting to try and check PAPR reduction methods with HICA to optimize the PAPR performance as well.
REFERENCES [1] [2] [3] [4] [5]
[6] [7] [8] [9]
[10] [11]
[12]
[13]
[14] [15]
[16] [17]
[18]
[19] [20]
C. Shin, Jr., R. W. Heath, and E. J. Powers, “Blind channel estimation for MIMO-OFDM systems,” IEEE Trans. Veh. Technol., vol. 56, no. 2, pp. 670–685, Mar. 2007 Yufei.jiang, xuzhu, enggeelim, yihuang. Orthogonal Sequences Based Multi-CFO Estimation and Semi-Blind ICA Based Equalization for Multiuser Comp Systems. Computer Science &Information Systems. 2012; 9(4): 1385-1406. J Gao, X Zhu, AK Nandi. Independent Component Analysis for Multiple-Input Multiple-Output Wireless Communication Systems. Signal Processing. 2011; 91(4): 607-623. Chiu Shun Wong, Dragan Obradovic. Independent Component Analysis (ICA) for Blind Equalization of Frequency Selective Channels. 14thIEEE workshop Neural Networks for Signal Processing. 2003; 419-428. Obradovic D, Madhu N, Szabo A, Wong CS. Independent Component Analysis for Semi-Blind Signal Separation in MIMO Mobile Frequency Selective Communication Channels. IEEE International Joint Conference on Neural Networks (IJCNN). 2004; 1: 53–58. M Sifuzzaman, MR Islam, MZ Ali. Application of Wavelet Transform and its Advantages Compared to Fourier Transform. Journal of Physical Sciences. 2009; 13: 121-134. Karanpreet Kaur. Disceret Wavelet Transform based OFDM System using Convolutional Encoding. Master thesis. Patiala: Thapar university; 2014 Xiangli Shao, Jian Chen and Yonghong Kuo, "Blind Channel Estimation for MIMO-OFDM Systems Based on Repetition Index", 2011 International Conference on Internet Computing and Information Ser-vices, IEEE, 2011 Nima Sarmadi, Marius Pesavento, "Closed-Form Blind Channel Estimation In Orthogonally Coded Mimo-Ofdm Systems: A Simple Strategy To Resolve Non-Scalar Ambiguities", 2011 IEEE 12th International Workshop on Signal Processing Advances in Wireless Communications. Chao-Cheng Tu and Benoît Champagne, "Blind Recursive Sub-space-Based Identification of Time-Varying Wideband MIMO Channels", IEEE Transactions On Vehicular Technology, VOL. 61, NO. 2, FEBRUARY 2012 Jian-Da Jiang, Tzu-Chiao Lin, and See-May Phoong, "New Sub-space-Based Blind Channel Estimation for Orthogonally Coded MIMO-OFDM Systems", 2014 IEEE International Conference on Acoustic, Speech and Signal Processing (ICASSP). Gamal Mabrouk Abdel-Hamid, Reham S. Saad, "Blind Channel Estimation Using Wavelet Denoising of Independent Component Analysis for LTE", Indonesian Journal of Electrical Engineering and Computer Science Vol. 1, No. 1, January 2016. Leila Sahraoui, Djmail Messadeg, Nouredinne Doghmane, "Analyses And Performance Of Techniques Papr Reduction For Stbc Mimo-Ofdm System In (4g) Wireless Communication", International Journal of Wireless & Mobile Networks (IJWMN) Vol. 5, No. 5, October 2013 P.Sunil Kumar, M.G.Sumithra, E.Praveen Kumar, "Performance Analysis of P APR reduction in STBC MIMOOFDM System", 2013 Fifth International Conference on Advanced Computing (ICoAC) Ying-Che Hung and Shang-Ho (Lawrence) Tsai, Senior Member, "PAPR Analysis and Mitigation Algorithms for Beam forming MIMO OFDM Systems", IEEE Transactions On Wireless Communications, Accepted For Publication, 2014 S. Sujitha, R. Ramachandran, "Performance Analysis of PAPR Reduction in MIMO OFDM System Using Modified Constant Modulus Algorithm", IJAREEIE, Vol. 3, Special Issue 2, April 2014 Sadhana Singha, Arvind Kumarb, "Performance Analysis of Adaptive Clipping Technique for Reduction of PAPR in Alamouti Coded MIMO-OFDM Systems", 6th International Conference on Advances in Computing & Communications, ICACC 2016, 6-8 September 2016, Cochin, India Gourav Misra, Arun Agarwal, "A Technological Analysis and Survey on Peak-to-Average Power Reduction (PAPR) in MIMO-OFDM Wireless System", International Conference on Electrical, Electronics, and Optimization Techniques (ICEEOT) – 2016 Feng Wan, Wei-Ping Zhu and M.N.S. Swamy, “Semi-Blind Channel Estimation of MIMO-OFDM Systems with Pulse Shaping”, 2008 IEEE International Symposium on Circuits and Systems LIU Haiyuan, SUN Jiancheng, “Blind MIMO-OFDM Channel Estimation Based on ICA and KRLS Algorithm, “2009 5th International Conference on Wireless Communications, Networking and Mobile Computing.
Spectral Efficient Blind Channel Estimation Technique for MIMO-OFDM… (Renuka Bhandari)
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 298~302 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp298-302
298
Mobile Learning Technologies Khalil Alsaadat Department of Educational Policies, College of Education, King Saud University, Riyadh, Saudi Arabia
Article Info
ABSTRACT
Article history:
Technological development have altered the way we communicate, learn, think, share, and spread information. Mobile technologies are those that make use of wireless technologies to gain some sort of data. As mobile connectedness continues to spread across the world, the value of employing mobile technologies in the arena of learning and teaching seems to be both self-evident and unavoidable The fast deployment of mobile devices and wireless networks in university campuses makes higher education a good environment to integrate learners-centered m-learning . this paper discusses mobile learning technologies that are being used for educational purposes and the effect they have on teaching and learning methods.
Received Aug 9, 2017 Revised Feb 20, 2018 Accepted Apr 21, 2018 Keyword: Mobile learning Technology
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Khalil Alsaadat, Department of Educational Policies, College of Education, King Saud University, Riyadh, Saudi Arabia Email: alsaadat@gmail.com
1.
INTRODUCTION Because m-Learning is such a new field the research is still in a stage where different categories of m-Learning pedagogy are being developed, identified, and researched. With this developmental stage in mind, the very existence of m-Learning, not to mention its growing application, is directly tied to the growth of mobile technology. This fact is why it is so important for researchers and practitioners to be familiar with mobile technology applicable to m-Learning. It simply is not possible for someone to log onto a learning management system (LMS) wirelessly from a personal digital assistant (PDA) if wireless networks don’t exist or if PDAs do not support wireless connectivity [1].
2.
THE MEANING OF MOBILE LEARNING The term mobile learning (m-learning) refers to the use of mobile and handheld IT devices, such as Personal Digital Assistants (PDAs), mobile telephones, laptops and tablet PC technologies, in teaching and learning. As computers and the internet become essential educational tools, the technologies become more portable, affordable, effective and easy to use. This provides many opportunities for widening participation and access to ICT, and in particular the internet. Mobile devices such as phones and PDAs are much more reasonably priced than desktop computers, and therefore represent a less expensive method of accessing the internet (though the cost of connection can be higher). The introduction of tablet PCs now allows mobile internet access with equal, if not more, functionality than desktop computers [2]. M-Learning, or "mobile learning", has different meanings for different communities. Although related to e-learning and distance education, it is distinct in its focus on learning across contexts and learning with mobile devices. One definition of mobile learning is: Learning that happens across locations, or that
Journal homepage: http://iaescore.com/online/index.php/IJAAS
IJAAS
ISSN: 2252-8814
299
takes advantage of learning opportunities offered by portable technologies. In other words, mobile learning decreases limitation of learning location with the mobility of general portable devices. The term covers: learning with portable technologies, where the focus is on the technology (which could be in a fixed location, such as a classroom); learning across contexts, where the focus is on the mobility of the learner, interacting with portable or fixed technology; and learning in a mobile society, with a focus on how society and its institutions can accommodate and support the learning of an increasingly mobile population that is not satisfied with existing learning methodologies. M-learning is convenient, in the sense that it is accessible virtually from anywhere(class, taxi, laundry room, bathroom) which provides access to all the different learning materials available. Moreover, it is collaborative; that is sharing is almost instantly among everyone using the same content, which will in turn also lead to receiving instant feedback and tips. M-Learning also brings strong portability by replacing books and notes with small RAMs, filled with tailored learning contents. In addition, this kind of learning is engaging and fun. With this kind of learning, it is much easier to combine gaming and learning for a more effective and entertaining experience [3]. Access : How widely available is the wireless network that will distribute the mobile content? Richness : Do pages load quickly? Do animations play in a smooth and seamless manner? Does the streaming media (media that is consumed-read, heard, viewed-while it is being delivered) flow at a sufficiently rapid rate? Efficiency : How large is the client that will be required to make use of a particular media player? How fast will the application load and play? Flexibility : Will the application be viewable on a variety of devices? Can content designed for use with one kind of device or operating system be played on other devices with some expectation of comparable quality? Security : Is the interactive mobile device protected from worms and viruses? Is the shared content protected from being intercepted by unintended recipients? Reliability : Will content be displayed in a consistent manner, regardless of the browser, device, and screen size? Interactivity : Does the application allow users to interact freely with the display and the content[4]. Through the use of mobile learning, user can access learning content without any borders so that it can be accessed at any time so that it can be accessed at any time with interesting illustration[5].
3.
MOBILE HARDWARE Various types of derivative information have been increasing exponentially, based on mobile devices and social networking sites, and the information technologies utilizing them have also been developing rapidly. Technologies to classify and analyze such information are as important as data generation [6]. Caudill said the first device that comes to mind when mobile hardware is discussed is the PDA. These devices offer many of the features of a full-size laptop computer but in a package that fits in a pocket. As discussed, mobility is a primary component of m-Learning hardware, and few devices offer the combination of mobility and features that the PDA does. From the start, the PDA experience lends itself to being ideal for the m-Learning environment. Whether a Palm or PocketPC operating system, a PDA will start up almost instantly, as opposed to the boot process that is required for a larger computer. This advantage by itself is a significant one; if a learner wants to check e-mail or reply to a message board while in between appointments, the time spent booting up and shutting down a traditional computer platform is a very real deterrent, the PDA interface eliminates that wasted time. Once the PDA is active, it provides a wide variety of applications that in the past were not available on mobile devices. Andronico and colleagues 2003, investigate three areas of applications using PDAs in the m-Learning environment: The use of PDA as an enhanced organizer, by uploading/ downloading data with the central system in order to align periodically or on demand the agenda of the user (teacher, student, or other actors of the system) with all the academic appointments. This will imply the integration of the data schema of the agenda software of the portable device with the data coming from the central system. The browsing of newsgroups managed by the central Learning management system (LMS) on the PDA screen, in case the user has no keyboard attached to the portable device, or the full interaction with the newsgroup in the other case.
Mobile Learning Technologies (Khalil Alsaadat)
300
ISSN: 2252-8814
The browsing of the LMS web pages where it is possible to download the educational material and consult it with specific viewers (at the moment those related with the Office suite and with Acrobat PDF format [1]. a. First, producers of hardware and operating systems often minimise costs and maximise effect through product development partnerships b. Second, demand from consumers and businesses is the influence on the type of product being developed, with the common requests being: easier to use, smaller, faster, smarter, and greater security. This has resulted in advancements such as multiple security layers, Bluetooth, car and business kits, hands free and infrared/ data cables. When asked about the future applications or capabilities that are planned for existing products, the response was: ‘smaller, faster, better, cheaper, and more wireless technology to send bigger files faster.’ What percentage of mobile technologies are purchased for business purposes and what percentage is purchased for consumer use? The following quote tells the story: There is such a cross-over between personal and business use we’re unable to tell them apart. Higher end products (i.e., Bluetooth or wireless, products with extra security, or products containing enterprise solutions) are marketed only to business clients, but the simplest phone can be used for business as well as personal purposes Much is made of the potential of m-Learning, but what is actually happening from a developer perspective? The two following quotes illustrate the developers' insights: a. A flash-based mobile interface is now being produced for m-Learning, so that animated material can be used on mobile phones; the technology is now moving quickly to respond to the increasing speed of mLearning uptake b. As an add-on to other modes of deliver, m-Learning will increase. But it won’t replace other forms of eeducation because screens are too small and hard to read, and if you make them bigger, the device isn’t as mobile. m-Learning is most useful when it’s in a mobile, field environment[7]. Traxler turned our attention to some categories of mobile learning that are emerging,he listed the following: a. Technology-driven mobile learning – Some specific technological innovation is deployed in an academic setting to demonstrate technical feasibility and pedagogic possibility b. Miniature but portable e-Learning – Mobile, wireless, and handheld technologies are used to re-enact approaches and solutions already used in 'conventional' e-Learning, perhaps porting some e-Learning technology such as a Virtual Learning Environment (VLE) to these technologies or perhaps merely using mobile technologies as flexible replacements for static desktop technologies c. Connected classroom learning – The same technologies are used in classroom settings to support collaborative learning, perhaps connected to other classroom technologies such as interactive whiteboards d. Informal, personalised, situated mobile learning – The same technologies are enhanced with additional functionality, for example location-awareness or video-capture, and deployed to deliver educational experiences that would otherwise be difficult or impossible e. Mobile training/ performance support – The technologies are used to improve the productivity and efficiency of mobile workers by delivering information and support just-in-time and in context for their immediate priorities (for an early account, see Gayeski, 2002) f. Remote/ rural/ development mobile learning – The technologies are used to address environmental and infrastructural challenges to delivering and supporting education where 'conventional' e-Learning technologies would fail, often troubling accepted developmental or evolutionary paradigms Mobile distance learning could fall into any of these categories (with the exception of the 'connected classroom learning'); how it develops will depend in part on the affordances of any given situation. These affordances might include: a. Infrastructure, meaning power supply, postal services, Internet connectivity, etc. b. Sparsity, giving rise to infrequent face-to-face contact, lack of technical support, etc. c. The wider policy agenda including lifelong learning, inclusion (of rural areas for example), assistivity, participation, and access d. Mobile distance learning within a framework of blended distance learning and the affordances of other delivery and support mechanism [8].
4.
MOBILE NETWORKING Caudill noticed that the most prevalent and most widely recognized mobile networking technology is the IEEE 802.11 specification, commonly called Wi-Fi. Wi-Fi works by using a series of access points, which are transmitter/ receiver stations that wireless devices can connect to via their own Wi-Fi networking IJAAS Vol. 7, No. 3, September 2018: 298 – 302
IJAAS
ISSN: 2252-8814
301
card. Initially seen as external cards that were used in a Personal Computer Memory Card International Association (PCMCIA) slot on laptops, Wi-Fi networking devices are now being integrated into standardsize PDAs and even smaller platforms such as mobile gaming devices. If a mobile device does not have a built-in Wi-Fi card there are a wide variety of add-on cards available, some small enough to fit into the SD slots on handheld devices. Thus, many devices not originally configured to access wireless networks can be converted to do so.he aso said: at 11 Megabits per second (Mbps), and 802.11g (g), which transmits at 54 Mbps. Both of these common standards are interoperable, There are two widely used standards, 802.11b (b), which transmits meaning that a g device can operate at a slower speed on a b network, and a b device can access a g network. Most devices today that include built-in Wi-Fi connectivity are using either the g or b standard, as are most publicly accessible Wi-Fi access points, which are referred to as hotspots[1].
5.
DISCUSSION The distance learning is an instruction system which has various and independent instruction environment, customized lecture design, method and techniques. Distance learning has been improving from letter learning to mobile learning [9]. Because of its high mobility and spontaneity, m-learning offers learners access to learning objects and resources that are distributed around us. In the ‘anytime, anywhere, learning area. what really matters is whether learners can access the right resources at the right time in the right place. Learners should be able to interact with the learning objects both in the real world and in the virtual world. Apart from that, the ultimate goal of m-learning is to enable mobile devices to offer individualized guidance and support during the learning process, and replace the one-size-fits-all receptive style of learning. It enables students to walk out of the classroom and actively explore their learning environment to gain more experience in collaboration and problem solving [10]. Traxler argued that mobile education, however innovative, technically feasible, and pedagogically sound, may have no chance of sustained, wide-scale institutional deployment in higher education in the foreseeable future, at a distance or on site. This is because of the strategic factors at work within educational institutions and providers. These strategic factors are different from those of technology and pedagogy. They are the context and the environment for the technical and the pedagogic aspects. They include resources (that is, finance and money but also human resources, physical estates, institutional reputation, intellectual property, and expertise) and culture (that is, institutions as social organisations, their practices, values and procedures, but also the expectations and standards of their staff, students and their wider communities, including employers and professional bodies). Implementing wireless and mobile education within higher education must address these social, cultural, and organisational factors. They can be formal and explicit, or informal and tacit, and can vary enormously across and within institutions. Within institutions, different disciplines have their own specific cultures and concerns, often strongly influenced by professional practice in the 'outside world' – especially in the case of part-time provision and distance learning. Because most work in mobile learning is still in the pilot and/ or trial phase, any explorations of wider institutional issues are still tentative [11]. Peters in alsaadat asked Is the promise of mobile technologies as a trigger to generate learning cultures realistic? And is m-Learning any more likely to increase interest in learning than any other form of delivery? Articles about the link between mobile technologies and learning organisations appear to fall into three categories: 1. A database focus that captures organisational knowledge 2. A human systems focus that allows synchronous communication and information sharing at the worksite 3. A learning development focus that suggests that learning about new technologies generates a more general drive for learning The database focus has, to a large degree, become the accepted wisdom in organisations that use structured processes to collect, codify, and manage knowledge. Mobile technologies have the potential to collect a greater range and percentage of data, through recording of activity on the device (and subsequent analysis of the patterns of access to specific information or information sources) and through the reduction of paper-based records as electronic systems replace paper in the field. Peters thinks that the capacity of mobile technology to deliver synchronous communication and knowledge-sharing can provide benefits to human (or soft) systems. Evidence of these benefits has been reported by Ragus 2004, who found that m-Learning encouraged simultaneous personal development, such as networking and socialisation, outside of normal working groups – an unexpected, and positive result of the m-Learning trials [12].
Mobile Learning Technologies (Khalil Alsaadat)
302
ISSN: 2252-8814
6.
CONCLUSION Mobile technology, both hardware and networking applications, is a necessary component for the existence of m-Learning. As instructors and designers, practitioners of m-Learning need to be fluent in the use of these technologies and cognizant of what technologies their learner population has access to. Application of specific pedagogical theories is directly connected to the technologies in use in a m-Learning system and as such, design of m-Learning environments demands a systems approach, where development accounts for all aspects of the environment. As technology continues to improve and innovate the options open to m-Learning will expand; the key is to focus on the fact that the goal of m-Learning is to facilitate learning, no matter what form the delivery may take [1].
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
[11] [12]
Caudil,Jason g,2007."The growth of m learning and the growth of mobile computing":paralle development” Exellence gateway.org.uk Wikipedia (reference:http://www.grayharriman.com/mlearning.htm) Wagner ellen D.2005,"enabling mobile learning" ,Educause review,V40,N3,pp40-53. Setiawan, alexsander. hanreas handojo.hadi,renadra , "indonisian culture learning application based on android",2017, international journal of electrical and computer engineering, vol 7,No 1, 2017 Jung se hoon.chan kim,jong.bo sim chun,"prediction data processing scheme using on artificial network and data clustering for big data", international journal of electrical and computer engineering, Vol 6,No 1,2016. Peters,2007."M learning:positioning educators for a mobile,connected future", the international review of research in open and distance learning,VOL8,NO2. Traxler,2007,"Defining,Discussing and Evaluating mobile learning : the moving finger writes" ,the international review of research in open and distance learning.VOL8,NO2. Soykan,emran,"the review of puplished articles on mobile learning area in Ebesco data base",procedia social and behavioral science,Vol 182,2015 Shin,ju ling . chun chu ,hui,hawang ,gwo gen.kinshuk,"an investigation of attitudes of students and teachers about participating in a context aware ubiquitous learning activity",British journal of educational technology ,Vol 42 ,No 3,2011. Alsaadat,khalil, "m learning and college education",European journal of education studies, Vol 3, Issue 5 ,2017. Alsaadat,khalil,"the importance of m learning in the educational arena",European journal of open education and e learning studies,Vol 2, Issue 1 ,2017.
IJAAS Vol. 7, No. 3, September 2018: 298 – 302
International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 3, September 2018, pp. 303~308 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i3.pp303-308
303
Requirement Elicitation Model (REM) in the Context of Global Software Development Muhammad Yaseen1, Umar Farooq2 1
University of Engineering & Technology, Peshawar, Malaysia 2 Universiti Tun Hussein Onn Malaysia, Malaysia
Article Info
ABSTRACT
Article history:
Contxext:Requirement elicitation is difficult and critical phase of requirement engineering and the case is worst in global software development (GSD). The study is about requirement elicitation in the context of GSD. Objective: Development of requirement elicitation model (REM) which can address the factors that have positive impact and the factors that have negative impact during elicitation in GSD. The propose model will give solutions and practices to the challenges during elicitation. Method: Systematic literature review (SLR) and empirical research study will be used for achieving the goals and objectives. Expected outcomes: The expected results of this study will be REM that will help vendor organizations for better elicitation during GSD.
Received Jan 19, 2018 Revised Mar 12, 2018 Accepted Apr 21, 2018 Keyword: Systematic literature review Requirement engineering Global software development Requirement elicitation model
Copyright © 2018 Institute of Advanced Engineering and Science. All rights reserved.
Corresponding Author: Khalil Alsaadat, Department of Educational Policies, College of Education, King Saud University, Riyadh, Saudi Arabia Email: alsaadat@gmail.com
1.
INTRODUCTION Requirement engineering (RE) is the systematic and discipline way of collecting user requirements for a software system and to manage it [1, 2]. The purpose of RE process is satisfaction of user needs and what customer wants from a software product [2-5]. Requirement elicitation is the first phase of RE during software development life cycle [6]. During elicitation phase we do direct communication with users or customers and gather requirements by applying various elicitation techniques. The quality of software system is more depended on the quality of how better the requirements are gathered [1, 7]. RE is very difficult when we do it locally but the case very difficult when the development is done globally where clients and vendors are separated by distance and face challenges like culture issues, time zone difference and difference in languages and terminologies. Due to extra challenges in GSD proper elicitation process is affected [8, 9].
2.
MOTIVATION AND RELATED WORK Miguel Romero [10] discusses that culture and time differences are big challenges in GSD. For reducing the effect of challenges it is necessary for people to share knowledge about requirements and have a proper knowledge management system. The author describes relevant skills for elicitation in GSD are; English language skills, understanding of cultures of others, computer mediated communication skills, use of proper communication protocols, ability to resolve conflicts and teamwork skills are needed for effective elicitation. Nosheen Sabahat [6] after doing survey and interviews explains the effectiveness of elicitation techniques. She concluded that in GSD the most effective elicitation techniques are prototyping, scenarios Journal homepage: http://iaescore.com/online/index.php/IJAAS
304
ISSN: 2252-8814
and interview. The prototype is considered to be the best technique in GSD because prototypes represents product earlier and customers are more satisfied. Scenarios are best in case where prototypes are difficult to make. Questionnaires and other traditional techniques are not suitable in GSD. The author propose iterative model where elicitation and analysis works iteratively at same time. Bin Wen [11] discuss that traditional elicitation techniques are not enough to apply in GSD during collecting of requirements, collaborative techniques like social intelligent for networked software and semantic web technology should be selected during elicitation. Gabriela N. Aranda [12] in his paper suggests the strategies to overcome the challenges like lack of face to face meetings and culture issues in GSD. Culture differences cannot be ignored but stakeholders can learn about these differences and training is the best solution for that. The author has explained in detail about the trainings and its strategies. Use of ontologies is suggested as best way to reduce language differences. Using ontologies can clarify the structure of knowledge and as well as reduce conceptual and terminological confusion. Technology selection must be discussed in team while doing elicitation because technology selection process is carried out by studying and confronting the personal preferences of people who need to work together. Fabio Calefato [13] discuss that as face to face communication during elicitation in GSD is difficult but still it is possible to have systems and technologies that can be alternative to face to face communication. After empirical studies the author design computer mediated tool for synchronous text solution in case where face to face communication is difficult. The tool contains audio and video conferencing facility. The tool was further evaluated through case study using students. Neetu Kumari [3] proposes the model which will address the issues with elicitation in GSD but the levels of this model are not defined. The model is limited to find the challenges only not the solutions. The characteristics of model are not explained. Further the methodology of collecting data from literature is not systematic so we need advance model which can address the challenges, critical success factors and the solutions for problems and challenges.
3.
OBJECTIVES The objective of this research study is to develop REM in order to support vendor organizations in better elicitation of requirements in the context of GSD. This model will address challenges and success factors during requirement elicitation. The propose model will address the solutions and practices needed to better implement success factors during elicitation and to reduce the effect of the challenges. For achieving the objectives, we will do the SLR and empirical study to find factors that are important during elicitation of requirements in GSD. For the implementation of factors, practices and solutions will be extracted through SLR and then questionnaire survey will be conducted to validate success factors, challenges and practices. Survey will be conducted in software industry to mention some new practices not mentioned in literature before. After finding the practices and solutions the propose model will be developed. The aim is to reduce the gap between researchers and software development vendors. Other researchers also adopted the same methodology in other fields to suggest such models [14].
4.
RESEARCH QUESTIONS The work reported in this paper is based on the five research questions which have posted in the following way: RQ1: What are the factors as identified from literature that have positive impact during requirement elicitation phase of RE in the context of GSD? RQ2: What are the factors in real practice that have positive impact during requirement elicitation phase of RE in the context of GSD? RQ3: What are the factors and challenges as identified from literature that have negative impact during requirement elicitation phase of RE in the context of GSD? RQ4: What are the factors in real practice that have negative impact during requirement elicitation phase of RE in the context of GSD? RQ5: Are there differences between the factors identified through the literature and the real-world practice? RQ6: Is the REM practically successful in terms of finding and alleviate implementation factors and challenges faced by vendor organizations during elicitation in GSD? Through SLR, RQ1 and RQ3 will be address i.e. studying what factors (CSFs and Cs) have already been reported in the literature. In future in order to facilitate vendors in implementing factors important for success of requirement elicitation, we will program/code the REM in the form of software. Moreover for
IJAAS Vol. 7, No. 3, September 2018: 303 – 308
IJAAS
ISSN: 2252-8814
305
overcoming challenges faced by vendor organizations in GSD the REM will suggest solutions. This tool will produce different assessment reports and do different activities for the software vendor organizations.
5.
DESIGN OF REM AND RESEARCH METHODOLOGY The methodology of the REM consists of the following three phases. Phase#1: SLR will be conducted for data collection. Phase#2: Empirical study will be conducted to validate the result of SLR and to find the practices for the mentioned factors. Phase#3: For evaluation and validation of REM, Case study will be conducted. To explain the aforesaid three phases the following subsections are added. 5.1. Collection of Data and its Analysis CSFs (critical success factors): Factors that have a positive impact during requirement elicitation in GSD. CRs (critical Risks): Factors that have a negative impact during requirement elicitation in GSD. Practices: For implementing CSFs, practices will be extracted and used. SLR will be used to identify factors (CSFs and Cs). Through SLR we will extract, analyse and will explore data relevant to our research questions. SLR is different from ordinary literature reviews being formally planned and more systematic. According to Kitchenham [15] SLR is divided into 3 phases. First phase is planning the review, second is conducting the review and implementation is the last phase. Before conducting SLR a protocol will be designed which include all the steps needed for SLR. From research questions a search string will be constructed for different libraries accordingly. Search procedure and plan will be defined and then protocol will be executed. After execution, inclusion and exclusion criteria will be define to tell which paper to include in final list.
Figure 1. REM development cycle activities
To successfully implement factors (CSFs and Cs), a questionnaire survey will be conducted with experts working in the software industry. The purpose of this survey is: a. Validation of the results of SLR. b. To find new factors (CSFs and Cs) which are not previously identified. c. To identify practices for the success implementation of (CSFs and Cs). Requirement Elicitation Model (REM) in the Context of Global Software Development (Mohammad Yaseen)
306
ISSN: 2252-8814
After REM is successfully design then for evaluation in real world environment case studies will be conducted in software industry. 5.2. Development of REM For the design of REM we have used the five stages as shown in Figure 2. A similar approach has also been used by another researcher[16, 17]. The development of REM is the first stage; it is used to set criteria for REM success. The below mentioned two criteria will be used for the development and assessment of REM. a. User satisfaction: This criteria focus on the satisfaction of end user from the result of REM. He / She should be able to use the REM without any confusion or ambiguity to promote objectives according to their requirements and assumptions. b. Usability: This criteria emphasis on the structure easiness of the REM. It states that the structure of the REM should be flexible and easy to understand because organizations do not accept complex models and standards which require resources, training and effort. Data collection and analysis is the stage 2. Rationalization and structuring of results will be performed in stage 3. Development of REM based on the results of empirical studies in stage4. Evaluation and validation of the REM via case studies will be performed in final stage i.e. stage 5. Planned structure of the REM is shown in Figure 3. Relationship between REM levels is also shown, factors/risks and practices used to address risks and success implementation of RE process. 5.3. REM Structure We will build the structure of the REM on the bases of following three extensions. a. REM levels b. Factors (CSFs, Cs) in each level c. Practices and solutions for the implementation of factors Classification of CSFs and Cs in different categories will be the base for defining the levels of REM. Each level will be consisting of different factors (CSFs and Cs). For each factor in the particular level, practices will be given for its proper implementation. Like CMMI and other models, for organizations to achieve certain level they must address and should follow the practices for each CSFs and Cs under that particular level.
Figure 2. REM development stages
IJAAS Vol. 7, No. 3, September 2018: 303 – 308
IJAAS
ISSN: 2252-8814
307
Figure 3. REM structure
5.4. REM Evaluation REM will be validated through industrial case studies. For case studies maximum of five organizations will be enough. Case study is the best tool for evaluating any model in real environment. A focus group session will be arranged to get feedback from the participant about REM. The criteria will be ease of use and user satisfaction as discussed in section 5.2. In focus group evaluation two or more people interact and generate ideas without getting help from researcher. Focus group is more open as compared to individual interview.
6. a. b. c. d. e. f.
RESEARCH CARRIED UPTO DATE We have done the following research work so far: Identification of problem and objectives Research questions specification Selection of research methodology Defining structure of REM Evaluation method selection Conduct of SLR
7.
CONCLUSION AND FUTURE WORK In our paper we have presented the structure of REM with different levels and phases. We have discussed how this model will help vendors in better implementation of success factors during elicitation in GSD. Detail methodology for the development of REM was introduced. This model will be used as tool for software developers and will produce different assessment reports in different situations.
ACKNOWLEDGMENT We are thankful to all members of the Software Engineering Research Group (SERG-UOM) at University of Malakand for their constructive reviews and also thankful to the anonymous reviewers of the conference.
REFERENCES [1]
[2]
D. Pandey, U. Suman, and A. Ramani, "An effective requirement engineering process model for software development and requirements management," presented at Advances in Recent Technologies in Communication and Computing (ARTCom), 2010 International Conference on, 2010. W. J. Lloyd, M. B. Rosson, and J. D. Arthur, "Effectiveness of elicitation techniques in distributed requirements engineering," presented at Requirements Engineering, 2002. Proceedings. IEEE Joint International Conference on, 2002.
Requirement Elicitation Model (REM) in the Context of Global Software Development (Mohammad Yaseen)
308 [3]
[4]
[5] [6]
[7]
[8]
[9] [10] [11] [12] [13] [14] [15] [16] [17]
ISSN: 2252-8814
S. Neetu Kumari and A. S. Pillai, "A survey on global requirements elicitation issues and proposed research framework," presented at Software Engineering and Service Science (ICSESS), 2013 4th IEEE International Conference on, 2013. M. Kauppinen, M. Vartiainen, J. Kontio, S. Kujala, and R. Sulonen, "Implementing requirements engineering processes throughout organizations: success factors and challenges," Information and Software Technology, vol. 46, pp. 937-953, 2004. B. Arthi, "Distributed Requirements Negotiations Using Mixed Media," Int’l Journal of Eng. and Technol, vol. 1, 2009. N. Sabahat, F. Iqbal, F. Azam, and M. Y. Javed, "An iterative approach for global requirements elicitation: A case study analysis," presented at Electronics and Information Engineering (ICEIE), 2010 International Conference On, 2010. A. Ahmad, A. Shahzad, V. K. Padmanabhuni, A. Mansoor, S. Joseph, and Z. Arshad, "Requirements prioritization with respect to Geographically Distributed Stakeholders," presented at Computer Science and Automation Engineering (CSAE), 2011 IEEE International Conference on, 2011. M. Yaseen, S. Baseer, and S. Sherin, "Critical challenges for requirement implementation in context of global software development: A systematic literature review," presented at 2015 International Conference on Open Source Systems & Technologies (ICOSST), 2015. T. Illes-Seifert, A. Herrmann, M. Geisser, and T. Hildenbrand, "The Challenges of Distributed Software Engineering and Requirements Engineering: Results of an Online Survey," presented at WORKSHOP P, 2007. M. Romero, A. VizcaÃno, and M. Piattini, "Teaching requirements elicitation within the context of global software development," presented at Computer Science (ENC), 2009 Mexican International Conference on, 2009. B. Wen, Z. Luo, and P. Liang, "Distributed and Collaborative Requirements Elicitation based on Social Intelligence," presented at Web Information Systems and Applications Conference (WISA), 2012 Ninth, 2012. G. N. Aranda, A. VizcaÃno, A. Cechich, and M. Piattini, "Strategies to minimize problems in global requirements elicitation," CLEI electronic journal, vol. 11, 2008. F. Calefato, D. Damian, and F. Lanubile, "Computer-mediated communication to support distributed requirements elicitations and negotiations tasks," Empirical Software Engineering, vol. 17, pp. 640-674, 2012. S. Ali and S. U. Khan, "SOFTWARE OUTSOURCING PARTNERSHIP MODEL," Science International, vol. 26, 2014. B. Kitchenham and C. Charters, "Guidelines for performing Systematic Literature Reviews in Software Engineering, Keele University and Durham University Joint Report," EBSE 2007-001, 2007. S. Ali and S. U. Khan, "Software Outsourcing Partnership (SOP): A Systematic Literature Review Protocol with Preliminary Results," International Journal of Hybrid Information Technology, vol. 7, pp. 377-392, 2014. M. Yaseen, S. Baseer, S. Ali, and S. U. Khan, "Requirement Implementation Model (RIM) in the Context of Global Software Development."
IJAAS Vol. 7, No. 3, September 2018: 303 – 308
Institute of Advanced Engineering and Science Indonesia
: D2, Griya Ngoto Asri, Bangunharjo, Sewon, Yogyakarta 55187, Indonesia
Malaysia
: 51 Jalan TU 17, Taman Tasik Utama, 75450 Malacca, Malaysia
COPYRIGHT TRANSFER FORM (Please compile this form, sign and send by e-mail) Please complete and sign this form and send it back to us with the final version of your manuscript. It is required to obtain a written confirmation from authors in order to acquire copyrights for papers published in the International Journal of Advances in Applied
Sciences (IJAAS) Full Name and Title Organisation Address and postal code City Country Telephone/Fax E-mail Paper Title Authors
Copyright Transfer Statement The copyright to this article is transferred to Institute of Advanced Engineering and Science (IAES) if and when the article is accepted for publication. The undersigned hereby transfers any and all rights in and to the paper including without limitation all copyrights to IAES. The undersigned hereby represents and warrants that the paper is original and that he/she is the author of the paper, except for material that is clearly identified as to its original source, with permission notices from the copyright owners where required. The undersigned represents that he/she has the power and authority to make and execute this assignment. We declare that: 1. This paper has not been published in the same form elsewhere. 2. It will not be submitted anywhere else for publication prior to acceptance/rejection by this Journal. 3. A copyright permission is obtained for materials published elsewhere and which require this permission for reproduction. Furthermore, I/We hereby transfer the unlimited rights of publication of the above mentioned paper in whole to IAES. The copyright transfer covers the exclusive right to reproduce and distribute the article, including reprints, translations, photographic reproductions, microform, electronic form (offline, online) or any other reproductions of similar nature. The corresponding author signs for and accepts responsibility for releasing this material on behalf of any and all co-authors. This agreement is to be signed by at least one of the authors who have obtained the assent of the co-author(s) where applicable. After submission of this agreement signed by the corresponding author, changes of authorship or in the order of the authors listed will not be accepted.
Retained Rights/Terms and Conditions 1. 2.
3.
Authors retain all proprietary rights in any process, procedure, or article of manufacture described in the Work. Authors may reproduce or authorize others to reproduce the Work or derivative works for the author’s personal use or for company use, provided that the source and the IAES copyright notice are indicated, the copies are not used in any way that implies IAES endorsement of a product or service of any employer, and the copies themselves are not offered for sale. Although authors are permitted to re-use all or portions of the Work in other works, this does not include granting third-party requests for reprinting, republishing, or other types of re-use.
Yours Sincerely,
Corresponding Author‘s Full Name and Signature Date: ……./……./…………
IAES Journals, http://iaescore.com/journals International Journal of Advances in Applied Sciences (IJAAS), email: iaes.editor@gmail.com, editor@iaesjournal.com
Guide for Authors International Journal of Advances in Applied Sciences (IJAAS) is a peer-reviewed and open access journal dedicated to publish significant research findings in the field of applied and theoretical sciences. The journal is designed to serve researchers, developers, professionals, graduate students and others interested in state-ofthe art research activities in applied science areas, which cover topics including: chemistry, physics, materials, nanoscience and nanotechnology, mathematics, statistics, geology and earth sciences. Papers are invited from anywhere in the world, and so authors are asked to ensure that sufficient context is provided for all readers to appreciate their contribution.
The Types of papers The Types of papers we publish. The types of papers that may be considered for inclusion are: 1) Original research; 2) Short communications and; 3) Review papers, which include meta-analysis and systematic review
How to submit your manuscript All manuscripts should be submitted online at http://iaescore.com/journals/index.php/IJAAS
General Guidelines 1) Use the IJAAS guide (http://iaescore.com/journals/ourfiles/IJAAS_guideforauthors_2019.docx) as template. 2) Ensure that each new paragraph is clearly indicated. Present tables and figure legends on separate pages at the end of the manuscript. 3) Number all pages consecutively. Manuscripts should also be spellchecked by the facility available in most good word-processing packages. 4) Extensive use of italics and emboldening within the text should be avoided. 5) Papers should be clear, precise and logical and should not normally exceed 3,000 words. 6) The Abstract should be informative and completely self-explanatory, provide a clear statement of the problem, the proposed approach or solution, and point out major findings and conclusions. The Abstract should be 150 to 250 words in length. The abstract should be written in the past tense. 7) The keyword list provides the opportunity to add keywords, used by the indexing and abstracting services, in addition to those already present in the title. Judicious use of keywords may increase the ease with which interested parties can locate our article. 8) The Introduction should provide a clear background, a clear statement of the problem, the relevant literature on the subject, the proposed approach or solution, and the new value of research which it is innovation. It should be understandable to colleagues from a broad range of scientific disciplines. 9) Explaining research chronological, including research design and research procedure. The description of the course of research should be supported references, so the explanation can be accepted scientifically. 10) Tables and Figures are presented center. 11) In the results and discussion section should be explained the results and at the same time is given the comprehensive discussion. 12) A good conclusion should provide a statement that what is expected, as stated in the "Introduction" section can ultimately result in "Results and Discussion" section, so there is compatibility. Moreover, it can also be added the prospect of the development of research results and application prospects of further studies into the next (based on the results and discussion). 13) References should be cited in text by numbering system (in Vancouver style), [1], [2] and so on. Only references cited in text should be listed at the end of the paper. One author should be designated as corresponding author and provide the following information: • E-mail address • Full postal address • Telephone and fax numbers Please note that any papers which fail to meet our requirements will be returned to the author for amendment. Only papers which are submitted in the correct style will be considered by the Editors.
International Journal of Advances in Applied Sciences (IJAAS) Institute of Advanced Engineering and Science (IAES) e-mail: ijaas@iaesjournal.com
IJAAS Journal Order Form Volume
Number
Amount
Price (Rp)
Price (USD)
Total
Name
:
Company
:
Address
:
City / State
:
Zip
:
Telephone/Fax
:
: ........................, ..........................
Signature: ..................................
Order form for subscription should be sent to the editorial office by fax or email
Payment by Bank Transfer Bank Account name (please be exact)/Beneficiary: LINA HANDAYANI Bank Name: CIMB NIAGA Bank Branch Office: Kusumanegara City: Yogyakarta Country :Indonesia Bank Account # : 5080104447117 Bank Swift : BNIAIDJAXXX >>> Please find the appropriate price in the price list on next page >>>
The price list for domestic and foreign subscribers
Volume
Number
Year
1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 8 8 8 8
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2012 2012 2012 2012 2013 2013 2013 2013 2014 2014 2014 2014 2015 2015 2015 2015 2016 2016 2016 2016 2017 2017 2017 2017 2018 2018 2018 2018 2019 2019 2019 2019
Price (IDR) for domestic subscribers 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00
Price (USD) for foreign subscribers 36 36 36 36 36 36 36 36 40 40 40 40 40 40 40 40 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44
The price included the printing, handling, packaging and postal delivery fees of the hardcopy to the address of the authors or subscribers (with Registered Mail). For foreign subscribers, an additional fee is charged if you would your order is mailed via Express Mail Service (EMS): - $25 for ASIA Continent - $35 for AUSTRALIA Continent - $35 for AFRICA Continent - $39 for AMERICA Continent - $39 for EUROPE Continent (No additional fee for delivering your order by Registered Mail).