Optimization Strategy for Electric Vehicle Charging Port Identification and Location based on Improved Point Cloud Registration

抄録

With the rapid rise of the electric vehicle industry, the gap between electric vehicle ownership and available charging pile is becoming increasingly large. In order to ensure the automatic charging efficiency of electric vehicles, it becomes crucial to achieve identification and location of electric vehicle charging ports efficiently and accurately. However, existing technologies face numerous challenges, such as noise interference, large data volumes, and low registration efficiency, which lead to suboptimal performance in charging port identification and positioning. Existing point cloud data noise reduction, feature point extraction and registration techniques for charging port identification and location have problems such as low noise reduction accuracy, poor quality of extracted points and low registration efficiency. Therefore, this paper proposes an optimization strategy for electric vehicle charging port identification and location based on improved point cloud registration. Firstly, the adaptive K-dimensional tree (K-D Tree) method is used to reduce the noise for point cloud data by dynamically selecting the optimal splitting dimension and value. Next, using the geometric feature information of the point cloud data, high quality feature key points are extracted by clustering analysis. Then, a feedback updating mechanism based on the registration loss function is proposed, which updates the K-D Tree model in real-time by the calculation results of the loss function to improve the registration efficiency as well as the charging port identification accuracy. Finally, simulation experiments are conducted to verify the performance of the proposed method in the identification and location of electric vehicle charging ports. The simulation results indicate that, compared with baseline 1 and baseline 2, the intersection over union (IOU) of proposed algorithm is increased by 43.54% and 55.46%, respectively.