Device Free Wireless Gesture Recognition by 5G-NR Signal

Device free wireless sensing is an important research direction for next generation communication systems, which attracts increasing attention recently. Hand gesture recognition is one of the most popular applications. Although gesture recognition via commodity WiFi or Long Term Evolution (LTE) signal is well studied, seldom work exploits New Radio (NR) signal despite NR is already extensively commercialized. Compared with previous wideband communication system, the maximum bandwidth of NR is further improved to 100 MHz in Frequency Range 1. On this basis, we propose an NR based hand gesture recognition scheme in the presence of interference activities. We analyze the effect of time domain channel path leakage, and propose a multi-path extraction and combination method, which can separate the expected gesture from background movements. A prototype system is built to demonstrate the performance of gesture recognition in the presence of interference, and the influence of signal bandwidth is studied as well.