Feature extraction related to human gait for a continuous-wave radar

Human gait analysis with radar techniques is of increasing interest in today's research for the discrimination between different moving objects. The use of a Continuous-Wave (CW) Radar allows a simple acquisition of velocity information. The short-time Fourier transform (STFT) of the radar signal is a representation of the signal in both time- and frequency domain. The display of the absolute value in logarithmic form, the so called spectrogram, can be used for the extraction of velocity data. The movement of the legs is a candidate for a feature evaluation. The varying velocity is expressed In a Fourier series whose coefficients can be used as features. These features offer the possibility to discriminate between different objects.