Wi-Fi SIMO Radar for Deep Learning-Based Sign Language Recognition

This study explores the use of Wi-Fi signals to recognize gestures and translate sign language. It employs an advanced passive radar system based on an injection-locked quadrature receiver (ILQR) in a single-input multiple-output (SIMO) setup. The system effectively detects 3-D hand motions for sign language using 2.4-GHz Wi-Fi signals from an access point (AP). Experimental data processing involves sampling pairs of baseband I- and Q-channel signals to create multiple output time series. These series are transformed into images using the Gramian angular field (GAF) method for deep learning. The images capture temporal and spatial information while minimizing noise interference. A deep learning model, combining convolutional neural networks (CNNs) and long short-term memory (LSTM) networks, achieves over 90% classification accuracy for ten Chinese sign language gestures from 10,000 labeled samples. Ultimately, the study successfully demonstrates a real-time sign language recognition prototype system using the proposed Wi-Fi sensing technology.