Hand Gesture Recognition Using FMCW Radar in Multi-Person Scenario
Hand Gesture Recognition Using FMCW Radar in Multi-Person Scenario
复制标题
DOI：
10.1109/wisnet51848.2021.9413794
复制
发表时间：
2021-01
期刊：
影响因子：
--
通讯作者：
Davi V. Q. Rodrigues;Changzhi Li
中科院分区：
文献类型：
--
作者：
Davi V. Q. Rodrigues;Changzhi Li
使用射频（RF）波的远程手势识别是用于用户与智能设备之间的交互的新兴方法。本文研究了在同一传感器照射下，利用调频连续波（FMCW）雷达识别多个近距离物体手势的可行性。建议使用四种手势。现有的基于雷达的手势识别主要依赖于传感器视场中一个运动引起的多普勒效应。此外，他们没有考虑位于同一范围内的目标的随机身体运动，这在现实生活中是不切实际的。利用FMCW雷达的距离分辨能力和特征信号提取，实现了FMCW雷达的抗干扰。实验结果证实了基于FMCW雷达的远程控制方法在存在多个干扰的情况下的有效性。
Remote hand gesture recognition using radio frequency (RF) waves is an emerging approach for the interaction between users and smart devices. In this paper, the feasibility of using frequency-modulated continuous-wave (FMCW) radar to recognize hand gestures when multiple closely spaced subjects are illuminated by the same sensor is investigated. Four hand gestures are suggested. Existing radar-based hand gesture recognition mainly relied on the Doppler effect caused by one motion in the sensor’s field of view. Besides, they did not consider the random body motion of targets situated at the same range, which is not practical in real-life scenarios. Leveraging the range discrimination capability of FMCW radars and characteristic signal extraction, interference rejection is achieved in this paper. Experimental results confirm the effectiveness of the FMCW radar based remote control approach in the presence of multiple interferences.