Living Object Detection of Electric Vehicle Wireless Charging Systems Based on a Single Motion Feature

The electromagnetic field generated during the wireless charging of electric vehicles is likely to cause electromagnetic injury to nearby living objects. Although some living object detection methods have been proposed, they are unable to accurately distinguish an organism from other non-living moving objects. In this paper, a method based on millimeter wave radar and motion feature is proposed to exactly detect the living objects rather than the moving objects. The standard velocity deviation is served as a single motion feature of moving objects. A machine learning model, namely support vector machine, is designed to classify the living objects and other non-living moving objects based on the selected single motion feature. Experimental results show that the average accuracy of human body identification is higher than 97%, and the false positive rate of foreign objects is about 10%.