WiFi Sensing: Data-Aided Approach

This paper proposes a data-aided approach to address the insufficient number of channel state information (CSI) samples in conventional passive WiFi sensing methods using beacon frames to enable high WiFi sensing performance. It selectively uses pilot symbols or all symbols based on error detection to increase the available CSI for WiFi sensing, resembling IEEE 802.11bf's null data packet approach, where all the received symbols are used for CSI measurements. Reusing correctly detected data symbols increases CSI samples, improving signal-to-noise ratio and detection sensitivity. Experimental results using a software-defined radio show enhanced sensing in human presence detection, with potential applications in tasks like respiration rate estimation, fall detection, and people counting.