Moving Foreign Object Detection and Track for Electric Vehicle Wireless Charging Based on Millimeter-Wave Radar

In recent years, the electric vehicle wireless charging system (EV-WCS) has been widely concerned because it is safe, reliable, automatic, and environment-friendly. The high-frequency and high-power electromagnetic field between the transmitting coil and receiving coil of an EV-WCS may be harmful to an intruded living object. Therefore, a reliable and sensitive living object detection device has to be equipped to reduce or close the charging power as living objects entering the wireless charging region. Existing studies mainly focused on the presence/absence detection of living objects, but lacked research on multiple object detection and moving trajectory track, which are both benefit for practical applications. This paper develops a 77GHz millimeter-wave radar-based moving foreign object detection method based on Kalman filter and data association algorithms. Simulation and experiments are both implemented to verify the effectiveness of the proposed method. Firstly, principle of the frequency modulated continuous wave radar is introduced. Formulations for calculating distance, velocity, and angle, are deduced. Secondly, a method that combines a Kalman filter with target clustering and data association algorithms is developed to track the moving trajectories of multiple objects. In particular, the Kalman filter is employed to reduce the influence of measurement noises. The DBSCAN algorithm is used for target clustering. The joint probabilistic data association algorithm is utilized to track multiple object trajectories. Finally, the proposed method is evaluated with simulation and experiments on TI IWR1642 77GHz radar in terms of measurement accuracy and multiple objects track. The results show that the proposed method can simultaneously measure the distance, velocity, angle, and number of moving objects with an average distance error of 2cm. Particularly, it can distinguish multiple objects that are very close to each other. With the DBSCAN algorithm, the proposed method can cluster five simulated objects with a maximum error of 3cm, and average errors of 0.63cm, 0.66cm, 0.48cm, 0.60cm, and 0.71cm, respectively. In addition, the proposed method is able to track the trajectories of the five objects within 15 calculating cycles, and the average errors are 0.35cm, 0.39cm, 0.39cm, 1.2cm, and 0.92cm, respectively. Also, experimental results indicate that the proposed method can track the moving object with a position error band of 5cm. Conclusions of the paper can be summarized as follows: ① The proposed method is able to detect the position and velocity of moving foreign objects accurately, and the maximum error for distance measurement is within 2cm. ② The proposed method can realize presence detection and trajectory track of multiple adjacent moving objects. ③ With the capability of moving object track, the proposed method is valuable for improving reliability of living object detection, and realizing more reasonable protection. For instance, the wireless charging system can determine to lower or close the output power by considering the position range and moving trend of objects. © 2023 Chinese Machine Press. All rights reserved.