With help from AI and machine learning, Wi-Fi sensing detects movement in the Wi-Fi environment. While it sounds promising, the technology still has a long way to go.
Usually, new Wi-Fi developments include faster data rates, more bandwidth or fresh security options. But something else is brewing from the IEEE 802.11 folks that's worth considering: Wi-Fi sensing.
You might hear more about Wi-Fi sensing in the coming months and years. Therefore, Wi-Fi aficionados should have a good foundational understanding of Wi-Fi sensing and consider how this new facet of Wi-Fi can be used at home and in enterprise settings.
What is Wi-Fi sensing?
Before we delve into Wi-Fi sensing, remember that 802.11-based wireless networking is a networking technology that works at Layers 1 and 2 to extend LANs into the air for wireless clients to gain untethered network access. That greatly simplifies 802.11, but it works for explaining Wi-Fi sensing.
Layer 1 Wi-Fi signals in the radio frequency (RF) domain are constantly manipulated by motion in a given coverage area. That motion could be objects, people, animals or anything that is moving while Wi-Fi signals traverse the space. The frequently changing RF is what is being sensed.
Can Wi-Fi "sense" itself? It can with the help of machine learning algorithms that perform computations that equal the act of sensing changes.
How does Wi-Fi sensing work?
The 802.11 framework relies on channel state information (CSI) to dynamically know what shape the normal operation of a Wi-Fi transmission is in and how to self-adjust for the best chance of working well. Changes of varying magnitude are almost constant, and CSI gathers a slew of information that end users generally don't care about. It's part of the magic behind the Wi-Fi curtain.
But, with Wi-Fi sensing, CSI is the fuel for new analysis by AI to provide accurate information on motion detection, including gesture recognition, the presence of objects or people and velocity of movement.
Picture the Wi-Fi environment used to self-form sensing networks using existing nodes to help detect hospital or home-care patient falls, people where they should not be or even whether someone is breathing normally. The applications are potentially endless -- and fascinating -- considering no additional specialized hardware should be needed.
Wi-Fi sensing has the potential to be astoundingly transformative.
Are Wi-Fi sensing benefits and use cases real?
The IEEE's 802.11bf working group is working on delivering a new standard specifically for wireless LAN (WLAN) sensing, with an estimated completion by 2024. Considering the wide range of expected use cases -- especially in IoT, healthcare and smart home applications -- expect to see a flurry of pre-standard releases for Wi-Fi sensing-capable products.
Machine learning and AI are becoming mainstream additions to many network systems to improve control and monitoring. They are also the underpinning magic to Wi-Fi sensing. The notion of WLAN environments becoming cognitive systems is moving away from simply being marketing hype, and Wi-Fi sensing has the potential to be astoundingly transformative.
As a practitioner of business-grade WLAN systems, I hope Silicon Valley gets Wi-Fi sensing right when it's integrated into enterprise products, especially regarding code quality and pricing. Already, many market-leading WLAN systems can be quite buggy despite a single access point retailing for over $3,000.
The 802.11bf feature sets are a whole new paradigm to integrate to these codebases, and the WLAN industry has become comfortable with subscription models that border on extortion in some cases. While I look forward to all that Wi-Fi sensing might deliver, I'm jaded on its affordability and true usability until I see it firsthand.
