Noncontact Human Gait Analysis and Limb Joint Tracking Using Doppler Radar

A full understanding of normal motion from human limb joint trajectory tracking could be essential to develop and establish a scientific basis for correcting any abnormalities. Practicality of using continuous wave (CW) radars, being simple and low cost, has been successfully investigated for noncontact gait monitoring, but it is still challenging to identify the motions of different limb joints. Here, we investigate the feasibility of extending the 1-D block processing algorithm to distinctly track specific limb joints and discuss the advantages and limitations of the technique for CW radar. To establish a repeatable reference data, we run a full wave EM analysis on customized Boulic human model emulating specific video recorded body motions. Results based on measured data are in agreement with simulated ones for simple motions such as swinging one hand or one leg only. The proposed technique is also successful in extracting lower body parts while the whole body is in motion; however, it is still hard to clearly extract upper body parts such as swinging hands.