Intelligent Hand‐Gesture Recognition Based on Programmable Topological Metasurfaces

Intelligent recognition of human positioning, posture, and hand‐gesture based on wireless electromagnetic (EM) induction has recently garnered widespread attention and is anticipated to find wide applications in various smart scenarios. Here a novel and robust hand gesture recognition method based on programmable topological metasurfaces is presented. By exploiting the programmability of surface waves by the topological metasurfaces, multi‐dimensional near‐field coupling pathways are flexibly constructed to extract the EM feature information corresponding to different hand gestures. An intelligent EM acquisition system that can rapidly capture the EM information of the topological metasurface for different gesture coupling conditions is built. The EM transmission data for thousands of gestures to train the neural network, which can achieve a recognition accuracy of more than 99% for 5 single‐hand gestures, and 25 two‐hand gesture combinations are experimentally collected. It is expected that the proposed scheme will advance the research in human body wireless sensing and promote the intelligent sensing applications of the topological metasurfaces. A novel and robust hand gesture recognition method based on programmable topological metasurfaces is presented. By exploiting the programmability of surface waves by the topological metasurfaces, multi‐dimensional near‐field coupling pathways are flexibly constructed to extract the EM feature information. A recognition accuracy of more than 99% is experimentally achieved for 5 single‐hand gestures and 25 two‐hand gesture combinations.