Research On Passive Person Perception Based On Channel State Information

With the widespread deployment of Wi-Fi,we can realize the perception of people in complex indoor environments through the use of wireless Wi-Fi resources,without the need to wear any equipment and set up other sensing devices.By extracting the channel state information(CSI)of the Wi-Fi physical layer and analyzing the CSI,it can reflect the status of the real indoor environment,improve the accuracy and accuracy of human perception,and enhance robustness,stability,and universality.The main research work of this paper is as follows:(1)Aiming at the fading effect of multipath effects on Wi-Fi propagation signals,a new non-line-of-sight path identification method is proposed.Based on the wavelet packet transform,the coefficients of the subcarriers,energy spectral density,and entropy are used as the input of the classification algorithm.The support vector machine algorithm is used to effectively identify non-line-of-sight situations and improve the recognition accuracy.(2)Aiming at the problem of perception of people in a complex indoor environment,a new passive perception method is proposed.A method based on kernel density estimation is used to distinguish between static and dynamic environments,to obtain the relationship between human activities and CSI,and to reconstruct and reconstruct amplitude information through wavelet transform,which not only preserves the useful environmental information of CSI but also eliminates environmental noise.The characteristic eigenvalue-based phase difference and amplitude-based eigenvalue high-order cumulants are extracted and classified using the optimized particle swarm support vector machine method.This method can improve the granularity of perception and can more accurately sense the status of people.(3)The method proposed in this paper is tested in an actual environment.The experimental results show that CSI can effectively identify non-line-of-sight signals,with a recognition accuracy of 96.23% in a dynamic environment and 94.75% in a static environment.Human perception performance can reach 92% in a static environment and 94% in a dynamic environment.