Sensor Network Approach to GPS RTK New Navigator Seminar th 20 June 2007
Nicholas Zinas Supervisors: Prof. Paul Cross Dr Marek Ziebart
Overview
Overview I) The concept of Network RTK - Network Correction Computation - Network Correction Interpolation - Network Correction Transmission
II) UCL RTK Software & Results - Concept of the RTK software - Models and Algorithms - Results
III) Current & Future Work - Theoretical Development - Research Experiment - Future Work
GPS Positioning
GPS Positioning
Absolute Positioning
PseudoRanges
Carrier Phases &Pseudoranges
Relative/Differential Positioning
One Reference Station
Static Fast Static Stand Alone
OTF Kinematic Stop & Go
Multiple Reference Stations
PseudoRanges
Carrier Phases & Pseudoranges
Precise Point Positioning CORS LADGPS
WADGPS Network RTK
Network RTK Benefits
1. 2. 3. 4.
Less Reference stations needed Low infrastructure Cost Improved Error Modeling increased availability and reliability Increased Reference Station – Rover distance separation Single Receiver cm positioning lower costs
Network RTK
Network RTK
Network Correction Computation
Fix Fix Network Network Ambiguities Ambiguities
Computation of Network Corrections
State Space
N
1i AB
Transmission of Corrections
Correction Interpolation
Observation Space
Linear Interpolation Algorithm
1i A ROVER
V
f 1i 1i 1i 1i 1i AB ( AB dTAB dI AB AB ) c f 1i 1i 1i 1i VAB AB AB N AB c
Linear Combination Model
Low Order Surface Model
Grid Based Parameteris ation
Broadcast (One way Communication)
Bilinear Communication
FKP
VRS
aX A ROVER bYA ROVER
RTK Software (2)
Models Troposphere ESA Zenith Delay model, Global Mapping Function (GMF - Boehm et al, 2006)
Ionosphere Klobuchar Broadcast Model (double differencing, ionospheric free linear combination)
GPS Antenna Model: IGS Antex file corrections (APC &PCV)
Geoid Model : Implementation of EGM96
RTK Software (3)
Algorithms Point Positioning SP3 (BRDC) orbit files implementation
LAMBDA method for Ambiguity Resolution Reference Station Ambiguity Resolution: spanning tree algorithm, closed loop approach
Single Epoch Carrier Phase Positioning Carrier phase and pseudorange (L1,L2,L1+L2 fixed solutions)
Multiple Epoch Carrier Phase Positioning: Helmert Blocking: Ambiguities Global Parameters, Rover position Local (L1+L2 fixed solutions)
Algorithms (3)
Results Multiple Epoch Carrier Phase Positioning (L1+L2) : Maximum 2 consecutive epochs Baseline : Barking – London (OS Stations) Baseline 18km (Models:None) 0.060000
Deviation from SKI-PRO coordinates
0.040000
0.020000
ΔLat 0.000000 00:00:00
04:48:00
09:36:00
14:24:00
19:12:00
00:00:00
04:48:00
ΔLong
-0.020000
-0.040000
ΔHeight
-0.060000
-0.080000
-0.100000
Time
Results (2) Multiple Epoch Carrier Phase Positioning (L1+L2) : Maximum 2 consecutive epochs Baseline : Barking – London (OS Stations) Baseline: 18 km (Models: Ionosphere,Troposphere,Antenna)
Deviation from SKI-PRO coordinates
0.060000
0.040000
ΔLat
0.020000
0.000000 00:00:00
04:48:00
09:36:00
14:24:00
19:12:00
00:00:00
ΔLong 04:48:00
-0.020000
ΔHeight -0.040000
-0.060000
-0.080000
Time
Results (3) Multiple Epoch Carrier Phase Positioning (L1+L2) : Maximum 4 consecutive epochs Baseline : Barking – London (OS Stations) Baseline 18km (Models: Ionosphere,Troposphere,Antenna) 0.080000
Deviation from SKI-PRO Coordinates
0.060000 0.040000
ΔLat
0.020000 0.000000 00:00:00
04:48:00
09:36:00
14:24:00
19:12:00
00:00:00
ΔLong 04:48:00
-0.020000 -0.040000
ΔHeight
-0.060000 -0.080000 Time
Results (4) Multiple Epoch Carrier Phase Positioning (L1+L2) : Maximum 4 consecutive epochs Baseline : Barking – London (OS Stations) Number of Satellites used in Multiple Epoch Carrier Phase Positioning 12
Number of Satellites
10
8
6
Satellites
4
2
0 00:00:00
04:48:00
09:36:00
14:24:00 Time
19:12:00
00:00:00
04:48:00
Results (5) Single Epoch Carrier Phase Positioning (L1+L2) (carrier phase & pseudorange) Baseline : Barking – London (OS Stations)
Baseline 18km (Models: Ionosphere,Troposphere,Antenna)
Deviation from SKI-PRO computed Coordinates
0.06 0.04 0.02 0 00:00:00 -0.02
DLat 04:48:00
09:36:00
14:24:00
19:12:00
00:00:00
04:48:00
DLong DHeight
-0.04 -0.06 -0.08 T i me
Results (6)
ΔLatitude
1σ Single Epoch Carrier Phase 0.0091 Positioning Maximum 4 Consecutive Epochs (All Models) Maximum 2 Consecutive Epochs (All Models) Maximum 2 Consecutive Epochs (No Models)
ΔLongitude
2σ
1σ
ΔHeight
2σ 1σ
0.0037 0.0182
0.0086
0.0162 0.0074
0.0038 0.0173
0.0083
0.0159
0.0036
0.0102
57 % 0.0318
0.0151 0.0072
0.0046 0.0204
91% 0.0323
0.0076
0.0167
Ambiguity Success Rate 2σ
43% 0.0302
0.0186 0.0092
13% 0.0372
Research Experiment (1)
CORS : 3 - 1 ZMAX Net (JGC) - 1 Trimble NetR5 (Geot) - 1 Trimble 4000SSI (Dionysus Satellite Observatory)
GPS Receivers : 14 Reference Station Networks a) JGC-DION-S1 b) JGC-DION-GEOT c) GEOT-DION-S1 d) JGC-DION-GEOT-S1
- 12 Trimble R8/5800 - 2 Trimble 5700
Research Experiment (2)
Dionysus – dion: 159.713m
Dionysus – JGC: 11604.133m
Dionysus Satellite Observatory
Dionysus – Geot : 9771.100m
Dionysus – S1 : 22001.990m
Research Experiment (3)
Dionysus – R12 : 789.034m Dionysus – R11 : 20331.590m Dionysus – R6 : 13676.251m Dionysus – R8 : 13125.525 Dionysus – R9 : 10801.607 Dionysus – R7 : 9295.421m Dionysus – R4 : 8459.406m Dionysus – R1 : 8309.600m Dionysus – R2 : 7594.345m Dionysus – R10 : 7327.178 Dionysus – R3 : 5530.964m Dionysus – R5 : 5108.805m
Dionysus Satellite Observatory
Research Questions
How many users needed in order to see improvement in the computation of the Rover position?
What is the magnitude of the improvement, if any? Does the geometry of the rovers affect the solution?
What is the Ambiguity Success Rate we can achieve in Single Epoch Carrier Phase Positioning? How can we take advantage of the redundancy in the equations in terms of modelling various error sources? Since we know the ambiguities between the users can we use this system to determine the atmospheric influence on GPS signals in a regional level?
Future Work
Implementation of the concept in the UCL-RTK software Test the results against GIPSY computed rover positions Compare the position time series of R12 against single baseline RTK positioning Generate VRS stations for each of the rovers and compare against VRS positioning Aim is to develop a robust centralized Network RTK positioning approach where all the appropriate steps will be carried out at a central processing facility, transmitting to the user just its final position.