Utilization of Wi-Fi Signal for Validation of Micro-Doppler Model in a Person Falling Scenario

With the demand for medical care among the elderly in the growing aging population, it is increasingly important to have intelligent detection of falling incidents. Understanding the micro-Doppler signature during a fall event with the accurate channel model is important for future development of the effective falling detection algorithm. In this work, the micro-Doppler channel model is presented based on a combination of radar cross-section (RCS) scattering and spheroid human models. A simplified 3D human fall model is created by creating spheroids between skeleton joints to represent different body parts. The scattering wave from each spheroid body component is computed on the basis of the scattering geometry and the RCS. The channel is then synthesized by superposition of these scattering waves. The measurement system that consists of the Wi-Fi-based channel sounder and the Kinect motion capture system is utilized to simultaneously acquire channel state information (CSI) and human skeleton data, respectively. The proposed model with the captured skeleton is validated with the measured Wi-Fi CSI by analysis of a joint time-Doppler spectrum.