Metasurface-Enabled Multi-Target WiFi Sensing

As an emerging technology, WiFi sensing has garnered widespread attention in recent years. However, due to limitations in WiFi bandwidth and hardware capacity, multi-target sensing with WiFi remains an unresolved issue. In this paper, we present <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlingShot</small>, a first-of-its-kind approach for multi-target WiFi sensing enabled by metasurface. <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlingShot</small> exploits a metasurface's ability of beam scanning to periodically scan among targets in a high frequency, so their sensing signals can be separated in the time domain. However, the WiFi-initiated and the metasurface-initiated signals will interfere with each other, which could significantly degrade the sensing accuracy. Moreover, the WiFi devices and the metasurface operate in a distributed manner, so the inherent clock offset and clock drift among them make it difficult to accurately separate the sensing signals of the targets. To address the above challenges, <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlingShot</small> leverages the metasurface's ability of phase shifting to cancel interfering signals and designs a passive clock synchronization scheme to synchronize the metasurface and WiFi devices. We implement <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlingShot</small> and evaluate it under various settings. Results show that <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlingShot</small> can sense up to 8 targets. Respectively compared to the state-of-the-art methods, <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">SlingShot</small> has 6.75 % higher mean accuracy in activity recognition and 58.34 % lower mean error in respiration monitoring.