This paper presents the results of our experimental investigation into how different environments impact the classification of human motion using radar micro-Doppler (MD) signatures. The environments studied include free space, through-thewall, leaf tree foliage, and needle tree foliage. Results on presented on classification of the following three motions: crawling, walking, and jogging. The classification task was designed how to best separate these movements. The human motion data were acquired using a monostatic coherent Doppler radar operating in the C-band at 6.5 GHz from a total of six human subjects. The received signals were analyzed in the time-frequency domain using the Short-time Fourier Transform (STFT) which was used for feature extraction. Classification was performed using a Support Vector Machine (SVM) using a Radial Basis Function (RBF). Classification accuracies in the range 80-90% were achieved to separate the three movements mentioned.

Date Published: 12 May 2016
PDF: 10 pages
Proc. SPIE 9829, Radar Sensor Technology XX, 98291B (12 May 2016); doi: 10.1117/12.2228397

Published in SPIE Proceedings Vol. 9829:
Radar Sensor Technology XX
Kenneth I. Ranney; Armin Doerry, Editor(s)