An AI-Powered RIS Technology for Hand Gesture Recognition in the Radiating Near Field

This paper focuses on using Reconfigurable Intelligent Surfaces (RIS) for gesture recognition through AI with experimental data. Using a mono-static setup of a RIS, 4.9 – 5 GHz radio waves can be directly beamed at a hand, which reflects the incident wave depending on the hand gesture. The received reflection signal from the hand is processed and given to a Convolutional Neural Network (CNN). The CNN performs with a prediction accuracy of 97.65%. Through this it is shown that it is possible to use an AI-powered RIS to recognise simple hand gestures, thereby adding to hand gesture recognition methods, and expanding on integrated sensing and communication (ISAC).