Indoor Through-the-Wall Passive Human Target Detection with WiFi

Abstract

Passive human target detection has a broad application prospect in security monitoring, intelligent home and humancomputer interaction. In through-the-wall scenario, due to the serious attenuation of signals caused by wall, the energy of target reflection signal decreases significantly and is submerged in the direct signal of the transceiver and the reflection signal of indoor static objects, making it difficult to be extracted. Therefore, the existing WiFi sensing system has some limitations in throughthe-wall scene, especially in detection of the stationary human target and the number of moving human targets. According to the above problem, we propose a detection system TWMD based on multidimensional signal features in this paper. Firstly, the received Channel State Information (CSI) data is preprocessed to eliminate the phase error and amplitude noise. Then, the multidimensional features are fully extracted from the correlation coefficient matrix by using time correlation and subcarrier correlation of CSI. Finally, the mapping between features and detection results is established by Back Propagation (BP) neural network. Our experimental results show that the recognition accuracy of TWMD in the environment with glass wall, brick wall and concrete wall are above 0.980, 0.900, 0.850, respectively. Compared with the existing detection system based on single signal feature, it improves about 0.450 in the detection of the number of moving targets.