WiFi perception of human gestures based on multi-dimensional association method research

Abstract

This paper studies the WiFi perception in indoor perception, studies the human gesture tracking system based on channel state information (CSI), and researches the CSI quotient model, static component elimination algorithm, and gesture tracking algorithm to realize human gesture perception based on WiFi. Through the derivation of the channel state information, it is found that the phase information contains the distance, speed, orientation, and other information of the target, which confirms the feasibility of WiFi sensing. Aiming at the core problem of how to extract the target information from the channel state information, this paper pays attention to the disturbance of the phase information caused by non-coherent noise and multipath clutter, uses the CSI quotient model for denoising, and constructs a static component elimination algorithm to remove the clutter, and the effect is remarkable. Finally, a gesture tracking algorithm based on a complex plane phase is constructed, and an actual system is built for verification. The results show that the system is capable of drawing figures, letters, and other figures by gestures.