Social distancing is a cornerstone of COVID-19 mitigation; it continues to play a vital role in reducing the risk of virus exposure and spread. While world health authorities have established that 6 feet (2 meters) is a safe distance, designing devices to assist consumers with social distance awareness and alerts has proven challenging because their core functionality relies on accurate, low-latency distance measurements.
In a recent collaboration, Altran worked together with semiconductor company Renesas to develop an intelligent wearable device/platform and prototype a social distancing wristband based on ultra-wideband (UWB) technology. The wristband alerts the wearer when a second device is detected within a user-specified “safe” distance. This article shares insights from phase one of that project: the process of evaluating wireless protocols to meet requirements for accurate distance measurement while keeping other key platform requirements, such as power efficiency, size and user experience, in balance.
A small device with a large list of requirements
In this project, our goal was to create an embedded social distancing platform suitable for wearable applications that leveraged Renesas IC technologies. As proof of concept, a wristband prototype based on this platform was also designed and manufactured in low volume to demonstrate the functionality (monitoring and alerts) and user experience in a social distancing use case (Figure 1).
Figure 1. A wristband prototype alerts the wearer when a second device is detected within a user-specified safe distance. (Source: Altran)
A wearable form factor dictated the need for one or more wireless technologies, the choice of which centered around a few basic requirements.
- Accurate distance measurement – for accurate alerts and no false alerts. By far, for our use case, the most important criterion for choosing a wireless protocol is its ability to measure distance with a level of accuracy capable of distinguishing between safe and unsafe distances. Measurement accuracy is also key to eliminating (or severely reducing) the number of false alerts caused by imprecise distance measurement. Receiving alerts that may or may not equate to unsafe distancing makes it challenging for users to decipher real vs. false threats.
- The impact of the physical environment. The wireless protocol should be minimally impacted by the physical environment of typical use scenarios. The device, in other words, needs to be capable of delivering accurate and repeatable measurements whether it’s used indoor or outdoor, in line of sight (LOS) or non-LOS (NLOS) situations, and in dynamic environments, such as those with many moving objects or changing LOS.
- Low latency. To be effective, response time between threat detection and user alert must be fast enough that the user has time to take preventative action and/or needed precautions.
- Form factor. In a wearable device, wireless technology must be lightweight and small.
- Power efficiency. Wearables are battery operated, but detection – of an object, a person, a signal, etc. – typically involves sensors, components not known for their power efficiency. For our use case, it was crucial to design a wireless solution with exceptional power efficiency, in all operating modes, to deliver expected battery life between charges.
- Scalability. A social distancing use case, by definition, involves multiple people and often crowds, so the wireless solution must be able to provide reliable and accurate distance measurement for multiple simultaneous targets.
In general, each wireless technology supports distance and location measurement using some combination signal capture (using time-based, angular position, or received signal methods) and positioning techniques (using triangulation or trilateration methods) (Figure 2).
Figure 2: Typical distance/location measurement technique. (Source: Altran)
Evaluating wireless technologies
We evaluated several commercially available wireless protocols to assess how well they could meet our requirements for a social distancing wearable. Our candidates included Wi-Fi, cellular, Bluetooth Low Energy (BLE) and ultra-wideband (UWB). In general, the known distance/ positional accuracy specifications of each eliminated many protocols (Figure 3), but there are merits worth noting here.
Figure 3. Distance measurement accuracy of typical wireless technologies. (Source: Altran, using published references )
We looked at Wi-Fi first, simply because of its ubiquity. Its wide deployment in indoor environments made it a promising solution for the social distancing use case inside buildings, particularly in complex structures such as airports, alleys and parking garages, or underground locations where GPS and other satellite technologies may not be available or provide low accuracy.
Pros: Due to the widespread adoption of Wi-Fi and convenience of setting up Wi-Fi networks, solutions could be deployed quickly for user positioning with very low cost and effort. In addition, with recent advances in Wi-Fi-based indoor positioning, Wi-Fi can provide reliable and more precise location services (than older Wi-Fi technology) suitable for some social distancing applications.
How it works: In a Wi-Fi system, a wireless transmitter, known as a wireless access point (AP), is required to transmit radio signals to communicate with user devices in its coverage area. The most common and easiest way to support indoor positioning is to calculate the user’s location based on the received signal strength indicator (RSSI) of signals from the user device. RSSI accuracy is in the range of 10+ meters, reduced to 1-3 meters approx. 75-85% of the time when utilizing newer Wi-Fi round-trip time (RTT) technology.
Summary: With the current advances of Wi-Fi, such as RTT, the accuracy of localization systems has significantly improved, resulting in its adoption for many indoor positioning applications. But distance accuracy down to 1 meter is still insufficient for our social distancing use case. In addition, Wi-Fi may not be effective in dynamic and complicated indoor environments due to the effects of NLOS environments, where signals can be scattered by obstacle shadows or people.
Wi-Fi-based technology is also used mainly for indoor and indoor-adjacent environments since it requires several APs for localization which may not provide seamless transitions in indoor-outdoor environments or be feasible in outdoor environments. Wi-Fi APs also require additional infrastructure such as power and protection from the elements, making them more complex to deploy.
With the explosive growth of Bluetooth-enabled devices in both indoor and outdoor environments, we also considered BLE technology for our solution.
Pros: BLE is used for short-range wireless communications (2.4 to 2.485 GHz); and its localization technology has several advantages when compared to Wi-Fi. BLE signals have a higher sample rate (i.e., 0.25 Hz to 2 Hz), providing more data from which to estimate distance. BLE technology is also more power efficient, thus more suitable for wearable devices. And BLE signals can be obtained from most smart devices, while Wi-Fi signals can be obtained from only APs. Finally, BLE beacons are capable of being battery powered, and thus are more flexible and easier to deploy than Wi-Fi access points.
How it works: Bluetooth-based localization is considered a practical approach in indoor and indoor-adjacent (outdoor patios, decks, etc.) environments. Indoor localization schemes collect RSSI measurements to detect the user’s location by using the triangulation mechanism with data from other Bluetooth devices.
Even though BLE-based indoor localization can achieve better performance than similar Wi-Fi localization systems, BLE technology is strongly affected by fast fading and interference leading to low distance accuracy when detecting another device. Accuracy is also strongly impacted by BLE advertising channels, human movements, and human obstacles. Methods proposed to improve the accuracy have achieved results down to 2 meters.
Summary: Promising for some social distancing applications, Bluetooth technology did not offer the consistency and accuracy of distance measurement for our social distancing wearable. Combining Bluetooth and Wi-Fi technologies was also explored, but this also did not result in the needed accuracy.
The cellular network infrastructure being widely deployed today can be used to help locate a person (or more accurately, an active SIM- or E-SIM-enabled smart device) within an outdoor environment. Although cellular connectivity is available within indoor environments, it does not currently produce accurate, reliable or fast enough measurements for our use case. Social distancing is relevant in both indoor and outdoor settings, so our discussion of cellular localization continues focused on outdoor applications.
Over the past few years, we have seen tremendous technological growth in cellular technology, some of which makes it a key candidate for use in location positioning applications. With current cellular networks supporting assisted GPS (A-GPS), enhanced Cell ID (E-CID) and observed time distance of arrival (OTDOA), cellular’s location accuracy has significantly improved.
Pros: One of the biggest advantages of cellular-based distance measurement is that it doesn’t require additional hardware infrastructure; it can operate on existing networks. In addition, most of the global population owns at least one cellular-equipped smart device, so deployment requires only a mobile app and some data processing capacity in the network.
How it works: In outdoor environments, cellular localization techniques use the algorithms mentioned above, namely A-GPS, E-CID and OTDOA. Here, E-CID enhances CID accuracy by adding reference data such as RSS levels and RTT information which is used to triangulate and calculate location coordinates. E-CID is also capable of using angle-of-arrival (AoA) information to improve the overall accuracy. Through these techniques, current LTE-based cellular protocols (3/4G) are capable of distance measurement accuracy in outdoor settings down to a range of 5-10 meters. Adequate if you lose your phone, but not accurate enough for our use case.
Many telcos worldwide are actively deploying new 5G cellular networks, and 5G has performance characteristics that could make it an excellent candidate for next-generation social distancing platforms. Further testing for our use case will bear this out, but given the state of 5G deployment, it was not considered for our project.
5G includes key technologies such as mmWave communications, device-to-device (D2D) communications and ultra-dense networks (UDNs), which contribute to its capacity for high-precision localization. Positioning techniques exploiting the mmWave communications are based on validation of triangulation measurements and angle of differences of arrival (ADOA). Simulations show that triangulate-validate and ADOA methods can achieve sub-meter accuracy with a probability of 85% and 70%, respectively, in an indoor, 18 x 16 m area . Localization accuracy can be further improved by implementing Kalman filtering algorithms.
Next-generation 5G technologies will also enable directional or linear array antennas, which will help make cellular-based positioning techniques viable for indoor applications as well. Here, the basic principles of AoA and time of arrival (ToA) are used for location measurement.
Summary: Suitable for outdoor environments where cell network infrastructure is fully deployed, existing 3/4G cellular protocols can only deliver down to 10-meter distance accuracy, unsuitable for our use case. While future generations of 5G are on track to achieve sub-meter distance accuracy – possibly lower with new techniques – deployment coverage is not sufficient at this time to make 5G solutions a viable choice for our need. And 5G’s suitability for indoor localization is still untested.
Unlike its Bluetooth and Wi-Fi counterparts, UWB operates at a broad spectrum of GHz frequencies, from 3.1 to 10.6 GHz. While UWB isn’t as widely deployed as the other protocols, it has some unique properties that made it an excellent candidate for our social distancing project as well as future indoor positioning use cases.
Pros: UWB can be used to capture highly accurate spatial and directional data and can sustain measurement accuracy at the centimeter level in short to medium distance ranges. UWB measurement accuracy is capable of distance accuracy down to 5-10 cm depending on the use case. Due to its unique characteristics such as high time-domain resolution, immunity of multipath, low-cost implementation, low power consumption, good penetration and wide bandwidth UWB signals (at least 500 MHz as specified by FCC), the impulse radio UWB technology has the ability to generate very short-duration Gaussian pulses in time domain, which enables some advantages when compared with other wireless RF technologies. Its wide bandwidth also gives it comparatively better immunity to multipath propagation and narrowband interferences prevalent in other communication technologies, since these types of interferences only affect part of the spectrum.
UWB has good penetration in solid materials, such as walls and other structures, so it can perform more consistently in NLOS environments. And a key advantage for our small form factor design, UWB allowed us to use smaller antennas due to increase in operating frequency and the RF circuitry was simpler, even though data transfer rates are higher.
How it works: In UWB communication, ultrashort pulses are used to communicate the data, which permits high-precision estimation of a two-way distance using the duration or time of flight (TOF) for the signals. The higher the spectral density provides more robustness in multi-path environments and hence more accurate ranging (distance measurement) capabilities.
As part of our UWB evaluation, we were provided with a UWB low-rate pulse (LRP) chipset from Renesas. LRP’s main advantage is down to 10 times lower power consumption than other standard UWB solutions, and hence was an ideal fit for our battery-operated wearable. For example, in transmit mode, typical power consumption for UWB high-rate pulse (HRP) ranges from 100 – 120mA, where UWB LRP typically draws 10-20mA. LRP standard based devices are not normally used for distance ranging applications, but the latest standard IEEE 802.15.4z enables them to operate in ultra-low power consumption mode while enabling secure ranging capabilities using round-trip TOF mechanisms we used in distance calculations.
In this first phase of our project, we typically measured UWB LRP’s distance accuracy within 20-30 cm. For clear LOS environments, closer to 20 cm; and for NLOS environments, closer to 30 cm. In next project phase, both distance accuracy and reliability will be tuned further to achieve closer to the needed 10 cm.
When compared to BLE and Wi-Fi, UWB works on short-burst impulse radio from Tx to Rx. In combination with its wide bandwidth, this reduces latency down to sub-ms since no decoding or modulation is required.
Summary: Based on an evaluation of key factors such as distance measurement accuracy, reliability, form factor/size, performance in the typical deployment environment, latency, low power consumption, scalability and reduced sensitivity to interference, we concluded that UWB LRP – leveraged in the new chipset from Renesas – was the best wireless technology for accurate distance measurement our social distancing project.
We finalized the social distancing platform utilizing both BLE and UWB in combination (Figure 4). This gave us the advantages of UWB’s high-precision distance measurement and consistency as well as the power efficiencies of BLE for always-on proximity detection when sensing a device within the local environment. In our application, BLE also supports pushing the historical alert data and actual distance measurements to a mobile app.
Figure 4: The final POC platform uses a combination of BLE and UWB LRP for optimal power utilization. (Source: Altran)
A clear choice for a social distancing wristband
Social distancing and mask wearing remain humanity’s first line of defense against the spread of COVID-19 and other diseases spread through contact or airborne transmission. In this project, Altran and Renesas teamed up to develop an embedded platform for a social distancing use case using a Renesas MCU and UWB LRP chips. While this project included design and (small-volume) manufacturing of a wristband prototype, the platform itself can be easily be adapted to enable social distancing (and contact tracing), as well as other indoor and outdoor location/position-based functionality in many types and form factors of IoT products where distance and location accuracy are essential. The option of using UWB LRP chips further extends the range of use cases to include those where power efficiency is critical.
 Wireless protocol distance accuracy data:
 Simulation results
|Nitya Verma is a Director of Embedded Solutions at Altran, part of the Capgemini Group. He is responsible for architecting many types of embedded products, from IoT to electronics, healthcare, industrial, and automotive applications, and leads product innovation for wearables. He also collaborates with Altran’s Research & Innovation team to provide emerging technology strategies and solutions across 5G, edge, wearables, IoT, and automotive domains. He has deep experience in deploying embedded wireless technologies to drive growth across use cases and industries.
For more Embedded, subscribe to Embedded’s weekly email newsletter.
The post Assessing technologies for a social distancing wearable appeared first on Embedded.com.