A Robust Passive Motion Detection System Based on Frequency-Space Diversity

Emerging wireless Passive Motion Detection (PMD) techniques enable traditional wireless network to sense target information and enhance user experiences such as smart home and elderly care. However, most existing PMD systems require channel stationarity and easily suffer from false alarms with limited channel resource. In this paper, we propose the design and implementation of R-MoDe, a robust PMD system based on frequency-space diversity. Firstly, a signal trend elimination method is presented for mitigating the impact of non-stationary channel. Secondly, we leverage the frequency-space diversity provided by the multi-subcarrier and multi-antenna, which are supported by the commodity WiFi systems, and propose a clustering-vote based feature extraction algorithm to boost the detection performance. Finally, based on the distribution of the extracted features we develop an adaptive motion detection rule to enable accurate and robust detection. We implement R-MoDe on commodity WiFi devices under real-life indoor environment and these experimental results show that R-MoDe can achieve a high detection rate of 99% using only one pair of transceivers.