An Investigation into the Non-Uniformity of Channel State Information in WiFi Sensing Systems

Wireless sensor networks are becoming increasingly widespread due to the expansion of the Internet of Things (IoT). With the omnipresence of wireless signals, Integrated Sensing and Communications (ISAC) has become possible, enabling rich profiling of commercial environments using IoT devices. Prior work has demonstrated that Channel State Information (CSI) collected from commodity WiFi-enabled IoT devices can sense various channel phenomena, including the movement and activities of unsuspecting patrons. As with any sensor system, these sensing applications depend on a sufficient rate of CSI for the system. This paper accordingly studies the impact of wireless channel sampling rate selection on the accuracy of detecting human movement speeds using WiFi. Our experiments reveal insights into the non-uniformity of CSI samples generated at the transmitter for various WiFi packet rates. To further investigate this non-uniformity in the captured CSI, we analyzed the propagation characteristics of IoT packets at different WiFi transmission rates. Our investigation of probability distributions showed that the inter-packet time differences become more predictable at higher packet rates, resulting in more uniformly measured CSI. These observations highlight weaknesses in existing WiFi sensing systems, and impact the choice of transmission frequency in future deployments.