Millimeter Wave Radar Gesture Recognition Based on ECC

Gesture recognition of millimeter-wave radar plays an important role in human-computer interaction within the current Internet of Everything context. Based on the signal pre-processing algorithm of MUSIC, this paper proposes a gesture recognition algorithm using error-correction code(ECC) module. Firstly, the gesture signals collected by millimeter-wave radar are processed into three sets of images of azimuth, elevation and depth by using signal processing algorithms such as MUSIC; then they are sent to the feature extraction network and output the discriminant code; finally, the discriminant codes are sent to ECC module to output the final verdict result. The addition of error correction code makes the whole decision mechanism more logical and robust, and it may also significantly enhance the real-time performance of gesture detection since it avoids the complicated network.