A Trajectory-Driven SISO mm-Wave Channel Model for a Human Activity Recognition

In this paper, we propose a trajectory-driven three-dimensional (3D) non-stationary channel model for millimeter wave (mm-Wave) single-input single-output (SISO) systems. The proposed channel model is designed to model and simulate the micro-Doppler characteristics of human walking activities in indoor environments. We present an expression of the time-variant (TV) channel transfer function (CTF), where its temporal variation is determined by the trajectories of moving human body segments. As moving human body segments, we study the influence of the head, chest, and hips. We investigate the TV Doppler characteristics of the proposed channel caused by the motion of the person. To do so, we focus on the TV spectrogram and the TV mean Doppler shift. To confirm the validity of the proposed channel model, we conducted a measurement campaign for human walking activities by using the Ancortek 2400T2R4 software defined radar kit operating at 24 GHz. The TV trajectories of human body segments are collected during the experiment by using the Rokoko smartsuit. The results for the micro-Doppler signature obtained from the measurements are compared with those of the trajectory-driven channel model. The findings show a good match between the proposed channel model and the real-world measurements.