A new technique for Doppler frequency analysis of radar signals

A longstanding performance goal for Army ground surveillance radars is the automatic classification of moving ground targets on the battlefield. Limited classification has been achieved by splitting the radar Doppler spectrum into lower and higher frequency bands, thus distinguishing between targets having different radial velocities. While this is an important feature, since only certain targets (vehicles) are capable of obtaining high radial speeds, it does not provide a distinction between a moving man and vehicles having low radial speed components. That there are distinct differences in the Doppler signatures between man and vehicle is easily established by listening to typical Doppler return signals. One can clearly distinguish the step modulation of a walking man from the relatively constant or slowly changing tone of a vehicle. As to be expected, a spectrogram of a walking man shows this repetitive behavior. Most obvious is a sinewave like frequency modulation at the step frequency of the main body return power appearing as a black band near 100 Hertz. This oscillatory temporal characteristic in the Doppler shifted frequency is practically never seen in vehicular return and is the discriminant chosen for automatic classification.