Gesture Classification with Handcrafted Micro-Doppler Features using a FMCW Radar

This paper deals with gesture recognition using a 77 GHz FMCW radar system based on the micro-Doppler (μ D) signatures. In addition to the Doppler information, the range information is also available in the FMCW radar. Therefore, it is utilized to filter out the irrelevant targets. We have proposed five micro-Doppler based handcrafted features for gesture recognition. Finally, a simple k-nearest neighbor (k-NN) classifier is applied to evaluate the importance of the five features. The classification results demonstrate that the proposed features can guarantee a promising recognition accuracy.