SEPTEMBER 2016
SENSOR
FUSION Comes to RetailNext Interview with Alexei Agratchev, Co-Founder and CEO of RetailNext
» »
Wi-Fi in Sensor Applications Omron’s D7S Seismic Sensor
Join Today CLICK HERE
eeweb.com/register
CONTENTS
Sensor Technology
EDITORIAL STAFF Content Manager Karissa Manske kmanske@aspencore.com Digital Content Manager Heather Hamilton hhamilton@aspencore.com Director, Creative Development Jeff Chavez jchavez@aspencore.com Graphic Designer Carol Smiley csmiley@aspencore.com Audience Development Claire Hellar chellar@aspencore.com
4 6 8 14 18
PRODUCT WATCH Omron D7S Seismic Sensor Industrial Automation with Infineon TECH REPORT Wi-Fi in Sensor Applications The Sensor Loaded MATRIX Creator INDUSTRY INTERVIEW Sensor Fusion Comes to RetailNext Interview with Alexei Agratchev, CEO and Co-Founder of RetailNext
Register at EEWeb http://www.eeweb.com/register/
Published by AspenCore 950 West Bannock Suite 450 Boise, Idaho 83702 Tel | 208-639-6464
4
8
Victor Alejandro Gao General Manager Executive Publisher Cody Miller Global Media Director Group Publisher
14 EEWeb
18
Sensor Technology
OMRON’S
D7S Seismic Sensor
On average, 50 earthquakes strike each day, or about 20,000 a year. Many are weak enough to go un-noticed. Some are catastrophic. We can’t predict when catastrophe will strike, but new technology from Omron can help to minimize the damage. Introducing Omron’s D7S: the world’s smallest class size seismic sensor that’s helping high-tech companies develop groundbreaking solutions to help prevent secondary damage when an earthquake strikes.
Click image to view video.
4
As we go about our lives, we usually aren’t focusing on the potential dangers that surround us. We’re used to the normal daily routine, and that routine gives us a sense of stability and security. We focus on our homes, our families, and our work, and we forget that catastrophe can strike when we least expect it. We’re not used to dealing with an unpredictable major crisis like an earthquake, so when it happens, we’re unprepared—slow to react, and slow to respond. A seismic sensor, however, responds at the critical moment to warn us of danger, and Omron’s D7S sensor takes that response to the next level, with the ability to analyze, connect, communicate, and more. When an earthquake strikes, the D7S uses advanced algorithms to instantly determine if the seismic activity is from an earthquake or some other source. If an earthquake is detected, the D7S could send an alarm to an emergency control center in the cloud. The control center could determine areas with the greatest damage risk, and then alert homes and utilities in those areas, and beyond.
PRODUCT WATCH
Equipment that is extremely sensitive to vibrations, like computers and servers, could receive an automated shutdown command to protect data loss. Critical financial systems that react negatively to natural disasters, like the stock market, could be automatically suspended until infrastructure is re-established. Major transportation hubs, such as airports, could receive automated earthquake alerts, and then use that data to redirect flights and keep air travel moving as efficiently as possible. Alarms could be sent to schools to alert teachers and children to seek shelter without delay. In the event that family members are separated during an earthquake, an alert could be sent to warn a person that their loved one may be in danger. Elevators, normally the first choice to exit a building, pose a safety risk during an earthquake, and could be automatically disabled in high risk areas, forcing people to choose a safer exit. Inside homes, basic utilities that we use on a daily basis, like an open flame on a gas stove, could pose an extreme danger
during or after an earthquake. A smart gas meter could automatically sense an earthquake and shut off the gas supply, eliminating the danger of fire. On the road, earthquake damage and emergency responders can cause major traffic jams. Automated earthquake alerts could be used by navigation services to generate alternate routes for navigation applications. And traffic signs could receive messages warning drivers of immediate danger.
WE CAN’T PREVENT EARTHQUAKES, BUT NOW, WITH OMRON’S D7S, WE CAN MAKE A DIFFERENCE IN THE OUTCOME.
The potential applications for Omron’s D7S are nearly limitless, and as its integration spreads through products, industry, and infrastructure, we will begin to see a trend in the reduction of secondary earthquake damage and a faster return to normalcy after a crisis. We can’t prevent earthquakes, but now, with Omron’s D7S, we can make a difference in the outcome. Get more details about Omron and the D7S and ways you can integrate this seismic sensor into your application…visit Omron.
5
Sensor Technology
Industrial
Compute & Communicate
AUTOMATION
with Infineon
T
he future holds a growing number of smart factories powered by smart semiconductor solutions. For Infineon, continually pushing for higher bandwidth and increasing efficiency while reducing the footprint is the key goal. The automation platforms today sport greater digital connectivity. Compute, communicate, actuate, and sense and security nodes form the backbone of industrial automation systems. This critical backbone is running 24 hours a day, seven days a week, and requires high levels of reliability, huge gains in resource efficiency, product individualization, greater adaptability, and security to ensure continuity of service.
Click image to view video.
6
The PLC functions as the brains of the system, making intelligent and logical decisions based on input from other systems and delegating tasks at a control, supervisory, and field level. Infineon’s XMC4000 & Imotion™ family of products feature rich peripherals dedicated to motor control, which enables the optimization of the system behavior with respect to the specific application’s requirements. For example, the EtherCAT & CAN bus interface. They bring powerful digital and analog power silicon together with algorithms, development software and design tools, speeding up your time to market.
PRODUCT WATCH Actuate
Sensing
Security
Actuate literally and figuratively means revving up a motor, a valve or a solenoid. In industrial automation (IA), these electromechanical systems require ultimate reliability. Infineon’s IGBT’s and OPTIMOS™ 5 Power MOSFETs, along with gate drivers such as the 2EDN MOSFET EiceDriver™ are ideal for higher power, higher voltage platforms. PROFET smart MOSFETs, on the other hand, offer a higher level of integration and the protection required by the harsh IA environment. Other integrated solutions offered by Infineon Technologies are IPM (intelligent power modules) like the µIPM™ family and NovalithICs, half-bridges, which integrate the gate driver & switches, as well as offering short circuit, current sensing, and over-temp protection.
The ability to sense various types of feedback is a crucial part of a machine’s operation, similar to how humans rely on their senses. In control elements, magnetic sensors provide reliable operation in harsh environments. Infineon sensors have built-in mechanical stress compensation. Infineon’s Hall (TLI/TLV496x), Angle (TLE5012B), and 3D sensors (TLV/TLI493D-x1B6) enable torque & speed control and more.
While initiatives like IoT promise huge gains, they also present new challenges—especially in terms of security. Systems that were originally designed to be isolated are now exposed to attacks of ever-increasing sophistication. For example, by compromising and taking control of the cloud, external control is very likely achieved, resulting in complete compromise of the security process. Infineon is the leader in predictive maintenance, safety and cyber security. OPTIGA™ grade security products, which are scalable across platforms, are an example of what Infineon offers for security.
Infineon offers products for all your industrial automation needs with technology that achieves more, consumes less, and is accessible to all. For more information, click here.
7
Sensor Technology
8
TECH REPORT
Wi-Fi
in Sensor Applications Sensors are used for measurements and for acquisition of data; but they require an effective data transfer mechanism to enable full-fledged applications that utilize the data they collect. A popular method of data transfer is through wireless means. Among the wireless mechanisms, 802.11 Wi-Fi stands out for a number of reasons. Redpine’s informative paper on the use of Wi-Fi in sensor applications describes these advantages; and also covers the implementation of such application scenarios.
Submitted by Redpine Signals
9
Sensor Technology
A Wireless Sensor Network Architecture A set of sensors monitored at a central facility is commonly referred to as a ‘sensor network’ as shown in Figure 1—arising from them being organized as a mesh of interconnected nodes. Sensors are deployed in a variety of environments and for a number of purposes—including building automation, facility management, environmental monitoring, industrial automation, military zones, asset management, and many others. Legacy sensor networks have used proprietary wireless transport mechanisms, and lately standards based wireless transport like Zigbee or Bluetooth. These wireless mechanisms are characterized by low operational power, low cost, low range, and largely proprietary network and data transfer protocols, including mesh networking. Sensors can potentially act as relays and be capable of adapting to changing scenarios.
Figure 1. A Wireless Sensor Network
10
Recently, however, wireless networks based on 802.11 ‘Wi-Fi’ have become pervasive in enterprise and industrial environments, among others, where sensors are often deployed. The presence of a well established and standardized wireless network has brought about a new possibility in the deployment of sensors—the use of Wi-Fi as the transport mechanism. 802.11 was, of course, not designed with sensor applications in mind, but innovations in implementations have enabled the use of 802.11 while satisfying all sensor requirements and more. A typical Wi-Fi deployment would have a set of access points positioned so as to provide coverage over the operational area—which might be a large factory floor, multiple offices, a hospital, a campus, and many more. In almost all cases, the network is a star configuration, with the possible provision of client roaming or hand-off. The Access Points are connected to the enterprise’s network through wired connections, or in part through a wireless distribution system. One of the advantages in this scenario is the location of a central controller or coordinator—this can be anywhere on the network, and with internet connectivity, potentially anywhere in the world. With standardized IP based transport, no additional network infrastructure needs to be added to transfer sensor information to any part of the network. Wi-Fi may also be used to determine and provide the location of a client device. Typically the device’s transmissions are received by several access points
TECH REPORT or receivers, or the device receives beacons from several access points. The relative signal strengths seen, or the relative times of arrival measured, provide the means to locate the client device. Access points in these scenarios are carefully located so as to provide the best ranging measurements, and a thorough measurement is carried out in advance of the signal propagation nuances of the deployment landscape.
Sensor Node Requirements A ‘sensor’ itself is, of course, the device that gathers data—the transducer used for measurement of temperature, humidity, pressure, proximity, position, flow, sound, presence of liquids or gases, etc. A wireless communication facility now enables the sensor to be deployed at any location with ease. The resultant ‘sensor node’ is a selfcontained device that gathers data, processes it, and communicates it to a control unit wirelessly. Successful deployment of a cost-effective wireless sensor network benefits from the availability of the following characteristics of the sensor node: •
Extremely low power consumption, with years of battery life
•
Standardized wireless transport
•
Good wireless range for ease of deployment
•
Ability to coexist with other wireless devices in the vicinity
•
Easy configuration
•
Low cost
Figure 2. A Sensor Node
We will now see how all of these requirements are provided for in a Wi-Fi based sensor solution. A Wi-Fi Sensor Node. Figure 2 depicts the typical configuration of a sensor node. The microcontroller is the main programming element in the design. It configures the sensors and the wireless subsystem; and it handles the battery or power management and the conditioning of sensor data as required. It may also be the device or functional block that provides the ultra-low power timed sleep mode. The Wi-Fi subsystem may be partitioned in various ways, based on the chipset used, but in general consists of a MAC, baseband processor, RF transceiver, and RF front-end. The battery drain of the system depends on several factors. They include the power consumption of the sensors, the active mode power consumption of the microcontroller and wireless subsystems, and the power consumed in the sleep mode. With typical duty cycles of operation of a few milliseconds ON and several minutes OFF, overall energy
11
Sensor Technology
drain could be dominated by the sleep mode power, with other influencing factors being the time taken to reach an operating point upon being woken up, and of course, the peak active mode consumption. Sensors used for critical monitoring are also often kept active at all times, and in these cases the sensor’s power consumption becomes important. A fundamental advantage with Wi-Fi based sensor networks is that no sensor node is burdened with the need to relay data from other nodes. The ON or active time therefore becomes highly regulated and predictable.
Points would not handle or respond to packets sent by unassociated clients unless they are a special class of packets like probe requests. The advantage with this mode, however, is that the sensor node is awake for a shorter duration, thereby extending battery life. A significant advantage in either mode is that all transmissions can fully adhere to the collision avoidance mechanisms of the 802.11 protocol. And in the crowded ISM bands, it is advantageous for all transmitting devices to follow the same protocol, ensuring that overall avoidance of collisions is maximized.
Transmission of data over the wireless channel can be in one of two modes. In the first mode, the node wakes up and associates to an access point or BSS before transmitting its data, with full compliance to the Wi-Fi network’s protocol. In the second mode, the sensor node wakes up and transmits data without going through the association procedure with the BSS. In this case, the infrastructure systems or controller would need to be customized to receive this data. Normal or off-the-shelf Access
The power advantage of 802.11 arises from the use of high transmit data rates in order to minimize active time. With modulation schemes up to 64QAM, Wi-Fi’s spectral efficiency is also higher than most other protocols. For example, sensor information amounting to 1k bytes can be transmitted in 802.11g or 802.11n format in as little as 160 microseconds, or a more usual 100 bytes of data can be sent in just about 30 us. Often, thus, the time taken to prepare the wireless subsystem to transmit the
Figure 3. Operational Profile for Power Consumption
12
TECH REPORT data is a significant energy drain. The subsystems therefore are optimized for these applications in enabling as quick a turn-on as possible. Ease of configuration before deployment, as well as after deployment in some cases, is an important requirement. Factory configuration of devices normally covers RF calibration, MAC address, and some configuration specific information. Other parameters like wireless network configuration, IP address, reporting intervals, and other operational parameters are commonly configured through a serial interface. In some cases they are also configured through the wireless interface itself, through a special mode. These configuration interfaces may also be used to deliver firmware upgrades.
The 802.11n Advantage The IEEE 802.11n standard primarily provides for high throughput, high-efficiency, and long range data connectivity, and includes the use of multiple antennas and transmit-receive chains. However, the standard also includes a single-stream mode that is intended to provide the benefits of 11n to low-power small formfactor devices including sensor nodes. The use of single-stream 802.11n WLAN in these client devices provides the following benefits: •
Higher throughput and lower transmit times—achieved through better efficiency in PHY and MAC.
•
Longer range—through use of multiple antennas at the access point
•
Preservation of 802.11n network capacity— the presence of legacy 802.11a/b/g clients forces the 11n nodes to use protection mechanisms and results in overall drop in network capacity. 802.11n helps avoid this.
RS10002-NBZ-S0M in a Sensor Node Redpine Signals RS10002 WiSeMCU™ device is the industry’s first Wireless MCU with a multi-protocol wireless (WiFi+Bluetooth+Zigbee) sub-system. It is a fully integrated module with an ultra-lowpower ARM Cortex M4F microcontroller as an application processor, a built-in wireless subsystem, advanced security, power-management, frequency reference, RF power-amplifiers, diversity RF Frontend and passives. RS10002 device is ideal for sensor applications as it can collect the data from multiple sensors using different wireless protocols (Wi-Fi, BT, BT LE, ZigBee, BT) and transfer the data over the network using Wi-Fi. The system-level and profile driven fine-grained power control offers a battery life of over 5 years. It offers higher throughputs which result in lesser “awake” periods, thus reducing power consumption substantially. The RS10002-NBZ-S0M supports TCP/UDP over IPv6/IPv4, making it extremely easy to integrate into an already existing WLAN. It supports wireless provisioning and offers several digital and analog interfaces to work with a variety of sensors.
13
Sensor Technology
14
TECH REPORT
The Sensor Loaded
MATRIX Creator
is One of the Most IoT-Ready Dev Boards to Date Development boards are a staple of the Maker movement, and this year’s Washington DC Maker Faire showcase did not fail to impress. Promising an IoT-ready platform that lets users build computer vision, home automation, robotics, and other custom hardware solutions, the Miami-based AdMobilize fielded a fancy Raspberry Pi shield called the MATRIX Creator, which comes readymade with 15 built-in sensors. From a hardware perspective, MATRIX Creator is an absolute bombshell, consolidating an ultraviolet sensor, pressure sensor, 3D accelerometer, 3D gyroscope, 3D magnetometer, humidity sensors, 8 MEMs microphone arrays, a temperature sensor, 36 LEDs, as well as a Xilinx Spartan 6 FPGA and an Atmel ARM Cortex M3 microcontroller. Connectivity and communication standards are equally abundant, consisting of Zigbee, Thread, Z-wave, NFC, IR RX/TX, 2 ADC Channels, 17 digital GPIOS, SPI, 12C and UART.
15
Sensor Technology
With its robust hardware and a circular form factor of approximately four-and-ahalf inches in diameter, the board present a compact form factor, a significant step up from the size of boards whose hardware was built from scratch. AdMobilize programming and R&D analyst Alfred Gonzalez-Cuzan pointed out to Electronic Products that the MATRIX Creator packs an additional row of general purpose input/outputs (GPIO) “in case you want to add anything else like a servomotor, or something we haven’t thought of.”
from the company’s app store, or use it as a foundation for creating their software; alternatively, those who prefer coding their hardware from scratch may also do so.
MATRIX Creator aims to satisfy both novices and experts alike, offering a simplified Linux-based OS that runs on the PI and is programmable in JavaScript, with Python and C++ later to come. Users simply select a pre-existing app
If you’re ambitious enough to create a hardware application reliant on multiple Raspberry Pis running in unison, the MATRIX CLI offers a hardware simulator for that purpose too. For more information, visit creator.matrix.one.
ZF Electronic Systems
Energy Harvesting Products Now Available! No Wires.
No Batteries.
No Boundaries! 16
Users interested in building more advanced sensing applications with a face, gesture and object detection, can access a pre-existing computer vision library for free. “Since we have an MCU, all these sensors operate in real-time, so you’re not relying on the Raspberry Pi’s processor to do much of the sensor work,” Alfred told Electronic Products.
Your Circuit Starts Here. Sign up to design, share, and collaborate on your next project—big or small.
Schematics.com Click Here to Sign Up
Sensor Technology
SENSOR FUSION Comes to RetailNext Interview with Alexei Agratchev, CEO and Co-Founder of RetailNext
18
INDUSTRY INTERVIEW
R
etailers know that the key to success lies in the continual improvement of customer experience. Recognizing what needs to be improved comes from data supplied regarding their customers. However, when it comes to analytics and data gathering, most brick and mortar stores fall short compared to their e-commerce competition. That is the gap RetailNext strives to bridge. By focusing entirely on the optimization of the shopper’s experience, RetailNext helps retailers bring about an increase in sales, a reduction in theft and the elimination of unnecessary costs. The ability to garner a large amount of data and interpret that into an action plan sets RetailNext apart. RetailNext began improving analytics for retailers by attaching a low-energy Bluetooth beacon to in-store devices. Now, they’ve introduced sensor fusion technology with their new product, Aurora. EEWeb recently met up with Alexei Agratchev, CEO and co-founder of RetailNext, to discuss how RetailNext’s sensor fusion data gathering is changing how companies interact with their inventory, employees and customers.
19
Sensor Technology
Tell us a bit about your background and how you came to co-found RetailNext. My background is in international relations. I worked for Cisco for about nine years managing different product and research and development teams. During my time with Cisco, I helped with an internal startup which dealt with video analytic security products focused on the casino markets. Cisco purchased Sypixx Networks, who had done a lot of the encoding, decoding, and hardware sold to casinos. My team and I built the software application on top of that technology. We also started looking into other sensors’ abilities to capture data in physical spaces. All of our work to that point had been for catching criminal activity in casinos. Arun Nair, Marlie Liu and I started RetailNext in November of 2007, anticipating that the technology around capturing data was going to improve dramatically over the next few years. We shifted our focus to retail stores, given the vast number of stores
20
and the trillions of dollars spent inside. At the time, there was a complete lack of visibility of what went on inside the stores regarding a shopper’s experience, especially when compared to online shopping. We took the opportunity to create this visibility and have been working in the retail arena ever since.
What type of company is RetailNext? What is the key focus of RetailNext? RetailNext is a retail company with a technology focus who helps retailers measure and improve customers’ in-store experiences. We place systems in the store and collect large quantities of data showing what customers are doing inside the store. With these systems, we capture data about a variety of things: how many people walk past the store, how many come in, gender, approximate age, new or repeat visitors, how much time they spend in the store, where do they go within the store, what merchandise are they spending time with, do they
INDUSTRY INTERVIEW
Aurora
interact with sales associates, did they utilize the fitting rooms, what was the wait on lines to check out, and so on. Then, what they ultimately buy ties into all of the other information. If you look at RetailNext, a lot of what we do lies in our ability to capture, in an accurate and scalable way, vast amounts of data. We’re able to do this across tens of thousands of stores across hundreds of brands in almost 70 countries. We install our hardware in over one thousand stores monthly, each installment requiring hundreds of thousands of sensors. The core of our company lies in our ability to know what to do with all of the data collected, and creating an action plan. Historically, RetailNext has been purely a software company, with offthe-shelf hardware. We’ve installed and re-installed hardware from many different manufacturers around the world. Our responsibility lies in installing, supporting, and capturing data. Over the last year, we’ve given more focus to creating a sensor of our own. We based this sensor on everything we’ve learned over the years to optimize what we need for the retail environment.
This sensor you mention is the Aurora, is that correct? Yes.
And RetailNext advertises the Aurora as a sensor fusion product. Can you explain what sensor fusion means? Sensor fusion refers to the added tasks the Aurora performs. Before Aurora’s creation, we had to separately install stereo vision cameras, Wi-Fi sensors, and Bluetooth sensors. The Aurora has all of these devices in one package, which simplifies installation and configuration. RetailNext has some patents granted on tagging video streams with information received from connected Wi-Fi or Bluetooth devices. Traditional cameras can tell a business if somebody came into the store, which displays they stopped at, and which products they sought out. Our tagging process takes it a step further because we can differentiate between customers and sales associates. Collecting this information with a single sensor has been a huge breakthrough for us.
Before Aurora’s creation, we had to separately install stereo vision cameras, Wi-Fi sensors, and Bluetooth sensors. The Aurora has all of these devices in one package...
Aurora
21
Sensor Technology
How is all of the data analyzed? It is very important for everything RetailNext does to be scalable, and so all analysis of data is automated. Historically, much of the information was converted in a server sitting in the back of the store, with the track then being sent to the cloud. With the Aurora device, a lot of the processing takes place within the device. We build them around Qualcomm’s Snap Dragon chip set, and we convert videos into tracks and underline pertinent information. As tracks are created, they are sent to the cloud, where most of the automated analysis happens.
Although we use video as a way to generate data, discernable features get automatically blurred, and later, the video is discarded.
Since the information is first videorecorded, how does RetailNext deal with privacy concerns? Protecting privacy is one of our top concerns. RetailNext complies with each country’s regulations and laws. We have installations in Germany, Switzerland, Australia, and Japan, where some of strictest privacy laws exist. Although we use video as a way to generate data, discernable features get automatically blurred, and later, the video is discarded. Video monitoring has been done in stores for years as a safety measure. Much of our focus in the past eight years has been to ensure that we comply with all legal aspects, ensuring that nothing will violate the client company’s contracts with their customers.
22
How do you protect your devices against hacking? Everything regarding our devices is encrypted. Often, our systems are an upgrade for retailers concerning their security. Preventing hacks is very important, especially considering the number of highprofile data breaches in recent years. Those data breaches cost retailers both money and customer trust, and retailers have since woken up and upgraded their infrastructure. Everything we do is extremely safe and is one of our biggest selling points.
How is your technology influencing both your direct customers and, in turn, their customers? Our job is to gather data to enable retailers to improve the in-store experience for their customers. The data we provide influences store layout and staff training. In many cases, a double-digit sales increase is seen by simply making the store easier to navigate. Also, ensuring that stores are sufficiently staffed at different times throughout the day often changes after we present our data. We can break it down even further. While we don’t differentiate each customer as an individual, we can narrow it down to say that in this particular store between 2:00 and 3:00 pm there were 20% more females between the age of 25 and 35, and that 60% of them are repeat customers. This information influences store merchandising and day-to day-staffing. In the past, some retailers ran their stores with a cookie cutter execution. Now, these
INDUSTRY INTERVIEW same retailers realize the importance of looking at each store, with different peak times and a different variety of clientele. They can customize parts of each store to fit individual needs and see increased performance. The final layer of data acquisition happens when customers opt-in to retailers’ apps. These apps generally include special prices or coupons as a reward for downloading the app. The more a customer uses the app, the more customized the app becomes to a customer’s preferences, which encourages customer retention.
Are you ever met with resistance from retailers regarding your product? How does RetailNext provide a service beyond what other retail technology companies offer?
RetailNext began in 2007 and for the first four years had no competitors. Our company had a great head start to work on the most effective technology and glean our knowledge about retail culture before the competition showed up. Today, there are around 50-60 startups with a similar vision of bringing online e-commerce-style analytics to brick and mortar stores. Many of these companies underestimate the challenges associated with building this type of business, so RetailNext has made some acquisitions. We have also seen other companies consolidate or go out of business. The value proposition associated with partnering with RetailNext is an easy sell, especially after we’ve been in the business for a while. We can show hundreds of successful examples. When RetailNext was first beginning, we
23
Sensor Technology
were met with more resistance from insisting they didn’t need any help, and they knew everything going on in their store. However, as companies see what our data provides in terms of sales increases and more efficient staffing, they are sold.
As part of our partnership with businesses, our retail professionals provide clear, defined reports to managerial staff.
24
Our technology can recognize and separate sales associates and customers. Using only video makes it impossible to make that differentiation in a practical way. We came up with the solution of attaching a low-energy Bluetooth beacon to our devices, and that technology is now built into Aurora. We also created an algorithm to automatically detect interactions between customers and employees and map out where those interactions take place.
What are your expectations for RetailNext in the next five years? Our growth has happened fast, and it’s accelerating in spite of the fact that the retail market never adopts anything overnight. We’re expecting to double or even triple our business each year for quite a long time to come. Retail commerce is a massive market, and we have a strong leadership position, with an ability to roll out our services and our standing credibility with customers. We are still investing much of our efforts in our products. Aurora is a good example of something that came out of our research, and it’s just the beginning of what we can do on a hardware side to make this easier. We’re also dedicating time to the software side as well as working with a lot of partners on our sensors.
Any final thoughts? Beyond the initial investment of data gathering, RetailNext takes a final step. We have teams of ex-retailers who have experience running highly successful companies. As part of our partnership with businesses, our retail professionals provide clear, defined reports to managerial staff. On top of the reports, we create training and merchandising plans built from the data we collect. So not only do we collect the most comprehensive data, but we also provide action plans.
Physical retail is not going anywhere because people enjoy shopping in physical stores. However, I think the store experience will dramatically change within the next 5 to 10 years. Brick and mortar retail stores are a massive industry, dating back thousands of years. From what we’ve seen, retail stores are going through one of the most dramatic transformations ever. It’s an exciting place to be.
MYLINK
146
01 20 Ju Malyy2
NEW issues are coming!
: O G A Wigning for a Des rket a M l a Glob chle Jim Ba r, w with ge Intervie dustry Mana In O WAG et g Mark n ti h ig L
Until then, enjoy these current magazines.
–
Ds tial LE n e d i s ➜ RFeeature-Richd, nt, an Efficie ic m Econo ns g Optio Garages n i t h g i ➜ Lfor Shops and
August 2016
Precision Drives
Power
A New Generation of Prosthetics
Debugging I2C Buses in Embedded Systems
‹
The MicRed Power Tester 600A
‹
Securing
the Airspace with
DroneShield
AUGUST 2016
Interview with John Franklin, Chief Scientist at DroneShield
Considering the ARM Architecture for Industrial PCs Using Ferroelectric-RAM to Build Smart Airbag Automotive Applications
Building a Platform to Launch the Future
Pilot Earphones Translate Foreign Languages in Real-Time
➥
Meet Naked, the World’s First 3D Fitness Tracker
➥
Interview with Ryan Panos, Vice President and General Manager at Cobham AvComm
July 2016
Read more EEWeb magazines
News » Tools » Resources » Ideas
New issues are coming up!