基于毫米波雷达和 MATLAB 的手势识别实验设计

A gesture recognition teaching experiment is designed by using the linear frequency modulation millimeter wave radar with one transmitter and two receivers and MATLAB. First, different gesture samples are collected by using the linear frequency modulation millimeter wave radar, and then the fast time Fourier transform and the slow time short-time Fourier transform are performed on the collected single channel signals by using MATLAB software to generate the range and radial velocity spectra. After coherent processing of dual channel signals, fast time Fourier transform and slow time short-time Fourier transform are performed to generate angle and angular velocity spectra. Extracting empirical features from the spectrum, based on which classification of hand gestures can be achieved by supervised machine learning algorithm. Or the spectrum can be directly input into the neural network for unsupervised learning and classification. This paper describes in detail the experimental principle and process of gesture recognition using radar and machine learning algorithms, which is helpful for students to master the basic methods and application flow of radar signal processing.