Wearless and Non-Invasive Gait Speed Observation System Based on Channel State Information of Wi-Fi

This paper intends to extract the Doppler spectrum generated by human walking from the Wi-Fi channel state information, and then estimate the human gait speed, so as to realize the wearless and non-invasive gait speed observation system based on Wi-Fi. In this study, we adopt a two-fold approach to address the research objectives. Firstly, a hardware platform is constructed to extract channel state information (CSI) from a commercially available Wi-Fi network card. Subsequently, an analysis of the subcarrier amplitude information within the extracted channel state information is conducted, leading to the design of an overarching system architecture. Lastly, signal processing techniques are employed to extract and enhance the Doppler spectrum in order to estimate human gait speed. At the same time, this paper collects the walking video data, uses the video tracking analysis and modeling software to extract the real gait speed of the human body, and then compares the experimental results of the gait speed observation system based on Wi-Fi with the real speed to verify the accuracy of the gait speed observation system realized in this paper.