FMCW radar based Position Adaptive Gesture Recognition Method

This paper proposes a position adaptive gesture recognition method based on frequency modulated continuous wave (FMCW) radar. Specifically, we construct the range-Doppler and range-angle parameter maps respectively. Then, a position mapping algorithm is proposed to process the denoised parameter map. Then the dynamic time warping (DTW) algorithm based on feature fusion is used for gesture classification. Finally, the experimental results show that the average gesture classification precision is 97.6 %.