Analysis of multi-human micro-Doppler signatures based on generalized S transform

Micro-Doppler signature is a unique signature of an object in motion, and can provide reasons for object classification and recognition. Human walking is a typical non-rigid motion, and there are many components with different frequencies overlapping seriously with each other in its radar echoes. So it is difficult to analyze walking human's micro-Doppler signatures. First, the human walking model and its radar echo model are constructed, then the micro-Doppler signatures of main body parts, single human walking and two human walking are analyzed with the generalized S transform (GST) as the time-frequency analysis tool, and the view that taking the micro-Doppler signatures of feet, calves, and torsos as the main signatures of human walking recognition is proposed, which lies a foundation for further study of human micro-Doppler signatures in future.