Short time state space method for human motion identification

In this paper, UWB radar system is introduced and the applicability of the short time state space method (STSSM) is investigated for the extraction of micro-Doppler (m-D) features from a walking subject. A full wave model of human motion and the Boulic model are used to validate the proposed STSSM and calibrate the parameters before it was applied to experimental data. It is experimentally demonstrated that STSSM can successfully distinguish the motion of different body joints by identifying corresponding m-D frequencies.