Wi-Fi sensing adds motion detection and gesture recognition to your wireless network


This article is part of the Technology Insight series, made possible with funding from Intel.

Thanks to the newest smart camera systems and their powerful analytics applications, IDC says the video surveillance camera market will grow to $44 billion by 2025. But regardless of how many cool features those devices integrate, many privacy-conscious consumers are more comfortable with them standing guard outside, rather than peering in.

Wi-Fi sensing offers a compelling alternative for keeping an eye on the kids after school, caring for elderly relatives remotely, and collecting biometric measurements with a bit of anonymity. By processing signals from the myriad Wi-Fi devices transmitting and receiving all around you, Wi-Fi sensing detects environmental changes. Depending on the application, those changes might include motion when there shouldn’t be any, triggering a security alert. Or if there is no movement when grandma should be stirring, a notification might get pushed to her care provider.

The technology behind Wi-Fi sensing is relatively new, and there aren’t any standards governing its implementation. As a result, gaps in today’s infrastructure limit the range of what Wi-Fi sensing can do, leaving a lot of untapped value waiting for industry consensus. Current approaches to Wi-Fi sensing like Cognitive System’s 2017-era Aura and Linksys’ Aware platform proved the technology works. But the best is clearly yet to come.

KEY POINTS

  • Wi-Fi sensing builds upon the mechanisms already used in wireless networks to detect environmental changes.
  • Possible use cases of Wi-Fi sensing include motion/presence detection, security, elder care, home automation, and gesture recognition.
  • Work is underway to standardize Wi-Fi sensing, which will pave the way for increased functionality, interoperability, and reliability.

Wi-Fi sensing turns your wireless network into a radar array

A 2019 study from Deloitte found that U.S. households have an average of 11 connected devices. The wireless radios in those devices typically send information to and receive information from a central access point, creating a star-like topology.

An example home deployment might have one access point or multiple APs in a mesh configuration. Not every device on the network has to be Wi-Fi sensing-capable.

Above: An example home deployment might have one access point or multiple APs in a mesh configuration. Not every device on the network has to be Wi-Fi sensing-capable.

Wi-Fi sensing works by detecting environmental changes between devices. So, when you cross the communication path of a wireless router and game console for instance, an agent running on the router might perceive the disruption and, like radar, determine your location, size, and so on.

Every client added to the network creates additional opportunities to gather environmental data. And of course, all those sensing measurements need to be processed by a device with enough computational horsepower. Access points, or even edge/cloud servers, work well for this purpose.

A typical Wi-Fi sensing system comprises three main components: a Wi-Fi radio, a software agent for signal processing, and an application layer that turns context-aware information into services/features.

Above: A typical Wi-Fi sensing system comprises three main components: a Wi-Fi radio, a software agent for signal processing, and an application layer that turns context-aware information into services/features.

According to The Need for Enabling Touchless Technologies, the IEEE 802.11 Working Group is working on an amendment to the standard to support Wi-Fi sensing on the 2.4, 5, 6, and 60 GHz frequency bands. Applications that don’t rely on resolution but do need some range are best served by the 2.4, 5, and 6 GHz bands. Conversely, the fine resolution required by gesture recognition lends itself to millimeter wave signals.

How will Wi-Fi sensing be used?

The ubiquity of Wi-Fi makes the technology an attractive platform for new business opportunities, and there are already several exciting use cases for sensing-capable Wi-Fi networks.

Home monitoring is perhaps the simplest implementation of Wi-Fi sensing. In fact, it’s already available through Linksys’ Aware—a $3/month add-on for the company’s Velop Mesh systems. Aware offers adjustable sensitivity, programmable schedules to limit extraneous notifications, and a logged history. If something moves between two of the Velop’s nodes and your sensitivity setting is dialed in correctly, the Linksys App lets you know.

Motion detection through a service like Aware is fairly basic compared to more advanced Wi-Fi sensing use cases. But remember that Linksys layers this functionality on existing wireless hardware, not a chipset developed with sensing in mind. And compared to the upfront hardware and monthly subscription costs of a security system, Wi-Fi sensing is remarkably affordable.

As Wi-Fi sensing evolves, the technology’s performance will improve, allowing it to detect subtler movement from a greater number of sources, and then pinpoint their locations. That’s when the potential applications really get wild.

Take elder care, for example. Adding sensing to an existing Wi-Fi network protects the privacy of residents, since they’re not on camera, and affords them greater independence. At the same time, machine learning algorithms fed by higher-resolution detection data differentiate between someone sitting down to watch soaps or accidentally falling.

How about home automation? Connected light fixtures, multi-room audio, and climate control systems could come alive and then fade to black as family members enter and leave rooms—all without the cost or complexity of passive infrared sensors for motion detection.

The wake-on-approach and lock-on-walk-away use cases are similar in concept. Consumer electronics with Wi-Fi radios might detect a user in proximity and switch on from standby mode. The same mechanism could lock a device when a logged-in user saunters off. Enabling both features would be a boon for the battery life and security of mobile devices.

Intel’s IoT group is especially interested in sensing’s application to hospitality. “In a hotel, housekeeping goes door to door checking for occupancy before servicing each room,” says Dr. Carlos Cordeiro, CTO of wireless connectivity at Intel and an IEEE fellow. “But imagine if they had the ability, through their Wi-Fi infrastructure, to know if there was someone in the room without being invasive.” The idea here would be to deploy housekeeping resources efficiently, keep from disturbing guests, and maintain privacy.

Given enough resolution, Wi-Fi sensing may even be used for gesture recognition, displacing many of today’s touch controls, particularly in public places. It goes without saying that COVID-19 has everyone acutely aware of shared surfaces. So there’s an immediate need for touchless alternatives to subway turnstiles, elevator buttons, and airport kiosks.

Opportunities abound in the enterprise, residential, and retail markets for radar-, ranging-, and sensing-based use cases able to leverage Wi-Fi for more than communication.

Above: Opportunities abound in the enterprise, residential, and retail markets for radar-, ranging-, and sensing-based use cases able to leverage Wi-Fi for more than communication.

Image Credit: Intel

What do we need to make Wi-Fi sensing happen?

Before we’re able to master our appliances with pinches, swipes, and waves, Wi-Fi sensing must experience a renaissance. Right now, a limited number of use cases are possible with commercially available hardware based on existing wireless standards. But there are gaps in the technology begging to be bridged by a broader industry-wide effort.

Work toward that effort is already underway. Last year, the Wireless Broadband Alliance published a white paper detailing the building blocks needed to create Wi-Fi sensing systems. Part of the analysis led by Cognitive Systems, Intel, and the Centre for Development of Telematics (CDOT) included guidance for overcoming apparent gaps.

As an example, Wi-Fi’s physical layer (PHY) protocols already perform certain measurements for sensing the surrounding environment. But those measurements weren’t exactly designed for the applications targeted by Wi-Fi sensing. If a future standard could specify additional measurement data, sensing accuracy would improve. Meanwhile, the technology’s efficiency stands to benefit by exposing each device’s sensing capabilities at the medium access control (MAC) layer.

Beyond a handful of other hardware-oriented improvements suggested by the WBA’s paper, standardized application programming interfaces (APIs), security considerations, and interoperability tests are also advised.

“Over the last 20 years, Wi-Fi has been used primarily as a communication service,” says Chris Beg, a senior mixed-signal architect at Cognitive Systems. “Sensing, although complementary, is a completely different usage of Wi-Fi. As a result, there is a gap when it comes to testing Wi-Fi sensing features, in terms of knowledge, procedures, and tools. To overcome these limitations, it became clear to Cognitive very early on that industry collaboration and standard support will be required.”

Standardization is key to Wi-Fi sensing’s future

Today’s standards and adherent Wi-Fi platforms do limit the scope of Wi-Fi sensing. However, home monitoring systems like Linksys’ Aware demonstrate the technology’s viability, even before it sees the surge of momentum standardization promises. Now, it appears the work that Cognitive Systems, Intel, and the CDOT put into its whitepaper is bearing fruit.

Cognitive’s Beg continues, “Shortly after publication, major momentum within the IEEE 802.11 community began around sensing, seeing the formation of a Topic-Interest-Group, followed by a Study-Group. Hopefully later this year, we will see the formation of the Task-Group (802.11bf), which will become responsible for defining the 802.11 standard support necessary to achieve both today’s use cases and also futuristic applications.”

Companies like Cognitive Systems are to be lauded for driving Wi-Fi sensing using existing hardware, says Intel’s Cordeiro. For the market to really take off, though, more needs to be done. “The standards work is important because that’s what gets us the right solutions at the radio level, the protocol level, and the API level for innovation to be able to happen.”

Combining measurement data from a laptop’s accelerometer, light sensor, gyroscope, and millimeter wave radio could enable health analytics and remote health monitoring functionality.

Above: Combining measurement data from a laptop’s accelerometer, light sensor, gyroscope, and millimeter wave radio could enable health analytics and remote health monitoring functionality.

With everything going on right now, we’re all spending more time in front of connected devices. Imagine if the PC you work and game in front of was able to measure your heart rate, your breathing rate, evaluate your stress level, and remind you to take breaks. “Use cases like that are really resonating with people,” continues Cordeiro. “We have these Wi-Fi devices out there for communications. But look at all the other use cases that go beyond communications and how they stand to benefit society. For us to be able to bring this to market at scale, with the right reliability and quality, the industry needs to come together. That happens through the standards process.”

Expect Wi-Fi sensing to become a more familiar concept as standardization progresses. Most recently, Cognitive Systems and its partnering WBA members defined a second phase project for the Wi-Fi sensing group, which involved developing a test methodology around Wi-Fi sensing and home monitoring. The group has defined several KPIs that can be used to evaluate a system and some test procedures for how they can be measured. Since sensing is a very different Wi-Fi application, the group hopes that this work will form the foundation of how to test and eventually certify a Wi-Fi sensing system.



Source link