mm wave based Gesture Recognition through FMCW Radar using Deep Learning Techniques

This research work employed machine learning models for hand gesture recognition using mmwave radar signals to classify different ergonomic and intuitive hand gestures. The proposed system uses IWR6843 AOP EVM frequency modulated continuous wave radar that helps in collecting the hand gestures data for the scalability purpose. The dataset consists of 2718 samples having six classes named as button press, circle, hand down, hand up, swipe left, and swipe right. These samples are collected from four individuals to train the deep neural network, recurrent neural networks, and transformer network. This study investigates the impact of various deep learning approaches on captured mmwave radar based data for gesture recognition. Moreover, a comparison is performed between the proposed methodology and state of the art study. The proposed methodology achieves a promising accuracy of 99.4% which shows that adopted model is robust, scalable and outperforms all other adopted models.