Advertisement

Automated presence detection for social distancing in retail applications

There is a variety of presence-detection solutions, including occupancy-density indication and absolute social-distancing measurement for social distancing in retail spaces that range in complexity and cost

Presence-detection solutions are not one-size-fits-all. The core sensors and technologies required to accomplish the variety of features and use cases are as diverse as the applications and environments in which they are used. This series of articles on presence detection will feature examples of use cases across commercial, industrial, and retail applications and will highlight core sensing capabilities, networking and processing considerations, and technology solutions that can be architected together to achieve the desired feature set.

In each case, system architectures must take into consideration environmental factors; the shape, structure, and layout of the space; power availability; node-to-node communications requirements and constraints; and data and information security. The first installment focuses on presence detection for social distancing in retail applications.

STMicroelectronics-social-distancing-retail

By Andrea Berry, market development manager, Industrial group, STMicroelectronics

Social distancing is a concept that few could have anticipated becoming such a pervasive part of our daily lives in such a short time frame. In the early stages of the 2020 pandemic, there has been a great deal of anxiety and stress over social distancing. Individuals who previously did not have to think too much about the perimeter around their physical presence are now forced into a state of hyper-awareness. What exactly does 6 feet look like anyway? And how do you respond when you and the people physically near you do not have the same estimation of 6 feet?

Even in open, outdoor spaces, social distancing has proven problematic. It has been even more complicated for shoppers who have had to continue to procure groceries and necessities for their households. The stores were called upon to implement new policies and procedures overnight to help reinforce social distancing behavior within their very narrow, often very congested aisles. Many shoppers may not even be aware of the new behavioral guidelines, thus rendering the new processes ineffective, resulting in increased tension.

As it is very likely that social distancing will be a mainstay of the retail experience for the foreseeable future, it is clear that stores will need to implement better, more automated means of social distancing for shoppers to rely on. Here, we will consider technology solutions that could be used to facilitate social distancing in a retail space. We will discuss various options ranging from density indication to absolute social-distancing measurement, providing a variety of choices to consider based on complexity and expense.

Occupancy-density indication
Indicating, and limiting, the number of people in a given area of a retail space is a simple but effective tool in retail social distancing. For any given section of a store, there is a maximum number of people that can fit in that section where a 6-foot distance between each person can be maintained.

In the case of grocery stores and other shops with aisle-format layouts, most have already implemented one-way aisles in an effort to reduce the flow of traffic from multiple directions, but this does very little to reinforce a 6-foot distance of shoppers coming into the aisle from the aisle entrance. In fact, many shoppers have failed to notice or practice the one-way protocol anyway, rendering it largely ineffective.

There are some very simple technical solutions that can, at the very least, provide indication that a shopper is entering from the wrong side or that the aisle is at capacity. Simple motion-detection technology, similar to what is used in home security systems (typically based off of passive infrared, time-of-flight, microwave, mirror optic, or ultrasonic sensors, or a combination thereof), could be used at each end of the aisle to determine when someone is entering or exiting.

However, this would require not only knowing that the person is present but also in what direction they are traveling. This could be implemented by placing a series of sensors at the “entrance” and “exit” of the aisle and monitoring the sequence of triggers. If the outermost sensor triggers but the innermost does not, then perhaps a person began to enter an aisle but backed out. If the outermost sensor triggers and then the innermost, the person has entered the aisle, and if the innermost sensor has triggered before outermost, the person has exited.

Such triggering information can be centrally processed to determine both directionality and level of occupancy, and warnings can be presented indicating that the shopper is entering from the wrong direction or that the aisle cannot accommodate additional shoppers.

It would also be possible to create an aisle “stoplight” using this approach, similar to the on-ramp indicators on interstates that are intended to space out the entrance of vehicles onto the highway. By estimating the amount of time that it might take a shopper to move down the aisle 6 feet inclusive of shopping activities, each aisle could be equipped with a stoplight intended to space out the entrance of each shopper.

This type of solution would have limited application in a store that was not arranged in aisle format. Also, it would not be effective in enforcing social distancing once the shoppers were in the aisles. However, this could be a cost-effective approach to enforcing directionality and capacity limits to each aisle.

STMicroelectronics-Presence-Detection-block-diagram

Designing a variety of proximity sensing building blocks requires a range of sensor, power management, and connectivity solutions.

Absolute social-distancing measurement
There is truly a vast array of technologies that can be used to monitor and measure the exact position of each shopper in a retail space. In fact, these solutions would not necessarily be much more complex or cost-prohibitive than the occupancy-density indicators described above. The cost of each solution must be considered as not just the relative cost of the core technology but by how many sensor nodes would need to be deployed. The complexity of the system is also related to the number of nodes, as well as how the data is aggregated, processed, and disseminated.

Optical sensors
An obvious means of tracking the exact location of people in any given space is by using optical sensors. In fact, as most stores already have security cameras located throughout the store, this may actually be a relatively easy solution to deploy, depending on how much coverage the existing camera infrastructure currently provides. If current infrastructure is insufficient, small optical sensor nodes could be outfitted around the store, whereby the number of nodes would primarily depend on the field of view, as well as any visual obstructions that exist between the node and the space being monitored.

Image processing may be performed at the edge by the sensor node, with real-time alert at the sensor if there was an occupancy exception in that area. The processing would need to be centralized in order to provide occupancy exceptions between two different nodes (for instance, if two people are too close to each other but are each being captured by separate and adjacent optical sensors). Both visual and thermal optical sensors could be used in such an application.

Radar/LiDAR
Similar to optical sensors, radar and LiDAR could be used to accurately pinpoint the location of every person within a given space. The long range of each system has the advantage of potentially reducing the number of sensor nodes that would be necessary in certain applications, as compared with the number of nodes that would be necessary for an optical system with a shorter range.

However, particularly in stores with aisle formats, the obstructions within the store would likely prove to be problematic. In this regard, radar would require a higher degree of customization and synthesis with the physical structure of each specific deployment, which would limit the ability of radar to be modular or scalable and would complicate the installation.

Pressure sensors
Another possible approach to monitoring retail social distancing is pressure-sensing flooring. Flooring equipped with pressure sensors (MEMS, strain gauge) could detect the introduction or presence of a weighted object (person, cart) and would thus also provide accurate location information based on the known location of the sensor itself. There already exist some pressure-sensing flooring solutions for retail applications as related to tracking and analyzing shopper behavior. However, this approach requires a significant infrastructure deployment and may be excessive in the context of social-distance monitoring.

Microphone arrays
Ultrasonic microphones have proven to be very effective in detecting the presence of a person. Their high degree of sensitivity can detect the sound of breathing, and they can isolate the sound even in the presence of increased background noise. This approach could be used for social-distancing purposes as well, although there are limitations in microphone arrays being able to differentiate more than one person. This approach would likely require substantial levels of signal conditioning and intelligence processing.

RFID localization or zoning
In environments where use of a shopping asset such as a cart or basket is likely, such as grocery or superstore environments, it is possible to use RFID technology to determine asset location. Each cart or basket could be equipped with an RFID tag, and the location of the cart or basket could be determined by readers positioned around the store.

In RFID localization, tag locations are triangulated based on relative signal strengths of the tag as read by multiple readers. In RFID zoning, an RFID reader is calibrated to a specific zone, and thus, indicated tags must be located inside of that zone. It would be possible for either of these techniques to be effective in determining social distancing in a retail environment; however, this would reflect only the location of the cart or basket and not necessarily the location of the person using the asset. Obviously, this will not work for instances when a person is not using such an asset.

Wi-Fi/Bluetooth
There is ongoing activity to evaluate the deployment of Wi-Fi and Bluetooth location determination for social distancing and contract tracing using mobile phone connectivity. Some stores track shoppers’ entrances and exits based on their cellphones searching for a Wi-Fi signal. Apps relying on Bluetooth have been developed that allow users to customize a “personal space” radius, which alerts them when there is another Bluetooth device inside their radius. These technologies have vast applicability not just in retail environments but anywhere where social distancing is of concern. Wi-Fi would have limited use in outdoor situations, but cellular communications could be used in this regard.

This approach requires that an individual has a networking device (typically a cellphone), and that the connectivity mechanisms are enabled, and is thus limited to the user’s employment of these factors. Additionally, some users may have privacy concerns about being traced in this fashion, as well as concerns about the data that could be exposed from their device. However, this method is certainly more effective than having no protocol at all and could be deployed fairly rapidly with limited infrastructure investment for a retail operation.

Summary
Innovative solutions are critical to continued adoption of the new norms of daily activity. There is a number of technologies that can be implemented to assist the practice of social distancing, which is especially important for retail activities that even at-risk people cannot avoid. The type of sensing technology selected will determine the capabilities of the solution, from basic indication of acceptable occupancy levels to precise measurement of socially distant limits.

Additionally, by combining some of the techniques described herein, it is possible to customize a solution for nearly any application. Many of these techniques can be implemented with very little development time, and as social-distancing parameters evolve, the technology solutions will also evolve to better accommodate a variety of applications and use cases.

Related articles

Part 2: Intelligent presence detection for commercial building automation

Advertisement



Learn more about Electronic Products Magazine

Leave a Reply