Speaker
Description
A through-the-wall detection (TTW) detection system using opportunistic WiFi signals in real-world environments is of great interest for non-invasive sensing in security, rescue, and monitoring applications. The existing approaches rely on controlled access points or complex MIMO configurations. To address this, in this paper, a TTW detection system that integrates a highly customized software-defined radio platform (SDR), BlueBottle, and complex signal processing techniques to extract and process Channel State Information (CSI) from all ambient Wi-Fi packet types is presented and tested. By tackling the irregular and sparse arrival of Wi-Fi packets, which violates the fundamental assumption of standard time-frequency analysis techniques, via adaptive interpolation to CSI temporal samplings, this system can provide denser data coverage, improved detection sensitivity and robust feature abstract performance. Experimental results demonstrate reliable TTW detection performance using opportunistically captured Wi-Fi data. Future work will focus on adaptive interpolation methods, improved sampling strategies to support deployment in diverse environments and classifications using cutting-edge techniques.
