Detecting Aged Person's Sliding Feet from Time Series Data of Foot Pressure

摘要

This paper focuses on sliding feet which is one of major causes of a fall accident of aged persons and proposes a sliding feet detection system based on ordinary piezo-electric sensors, which are embedded under the carpet. Specifically, the proposed system (1) starts to extract a wave occurred from one step (composed of a time series of voltage data measured by the piezoelectric sensors), (2) transforms it to frequency components by Fast Fourier Transform (FFT), (3) extracts the characterized spectrum which includes the sliding feet feature, and finally (4) detects whether a user on the sensor slides his/her feet by comparing differences between the features of ordinary walking and that of sliding feet. In order to verify our proposal system, we conduct the human subject experiment by collecting the ordinary walking step data of two persons and the sliding feet of one aged person, and the experimental results show that the system successfully detects the users sliding feet.