Dynamic gesture recognition using software defined radio based on smooth pseudo Wigner-Ville distribution

Abstract

Software defined radio (SDR) plays an important role in military and commerce because of its inherent flexibility. It also has great potential in home use, and it can be controlled by software installed on a personal computer or embedded system to achieve different purposes. In this paper, a novel method of dynamic gesture recognition based on support vector machine (SVM) and SDR is proposed. It is also an unconventional application of Wi-Fi signal processing. Different gestures performed between the transmitter and receiver affect the signal differently, so the received signal contains lots of information about the gestures. We use smooth pseudo Wigner-Ville distribution (SPWVD) to extract the feature vector of the signal and apply SVM to realize gesture recognition. The computer simulation results show that it is a feasible method for dynamic gesture recognition.