Hand Gesture Recognition using a Three-dimensional 24 GHz Radar Array

This paper proposed a hand gesture recognition method using a three-dimensional radar array. This array consisted of three radars to transmit and receive 24 GHz continuous electromagnetic (EM) wave. The integrated mixer heterodyned the scattered EM wave to the intermediate frequency (IF) with a local oscillator. Temporal and frequency signatures including I/Q signal magnitude difference, phase difference and spectra power integral were proposed as the features of recognition. A decision tree algorithm was constructed as the recognition classifier. 2000 training samples and 4000 untrained samples were used in the experiment. Results showed that this system could achieve a high recognition accuracy rate higher than 92%.