面向无人驾驶车辆的局部路径规划研究.pdf

技创新与应用 2023 年5 期 Technology Innovation and Application 新前沿 DOI：10.19981/j.CN23-1581/G3.2023.05.006 面向无人驾驶车辆的局部路径规划研究 1 1 2 2 2 徐华全 ，叶德超 ，谢振东 ，董志国 ，于洁涵 1.广东工业大学，广州 510000 ；2.广州市公共交通集团有限公司，广州 5 10000） 摘 要 针对无人车单车智能出现的因路侧设备损坏、障碍物遮挡导致的行驶安全问题，开展无人机与无人车协同的路径规划 研究。首先，提出一种基于无人机与无人车协同的无人智能系统模型。其次，提出基YOLOv3 算法 改进人工势场法相结合的无人 车局部路径规划方案，验证YOLOv3 算法对城市交通中常见人、车、交通标志等路况信息的检测具有良好的实时性 准确性。然后， 针对传统人工势场法，通过引入无人车与目标点的相对距离 设定无人车距离来优化人工势场模型。最后，模拟障碍物位置信息， 进行仿真实验。实验结果表明，改进后的人工势场法显著提高路径平滑度，有效地规划出车辆的行驶路径。 关键词 路径规划；YOLOv3 ；人工势场法；大数据；无人驾驶 中图分类号 U463.6 文献标志码 A 文章编号 2095-2945 渊2023 冤05-0021-05 Abstract: Aiming at the driving safety problems caused by roadside equipment damage and obstacle occlusion caused by the intelligence of unmanned vehicles and bicycles, the research on the coordinated path planning of unmanned aerial vehicles and unmanned vehicles is carried out. First, an unmanned intelligent system model based on the collaboration of UAVs and unmanned vehicles is proposed. Secondly, a local path planning scheme for unmanned vehicles based on the combination of the YOLOv3 algorithm and the improved artificial potential field method is proposed, which verifies that the YOLOv3 algorithm has good real- time and accurate detection of road condition information such as common pedestrians, vehicles, and traffic signs inthe urban traffic system. Then, for the traditional artificial potential field method, the artificial potential field model is optimized by introducing the relative distance between the unmanned vehicle and the target point and setting the distance of the unmanned vehicle. Finally, the position information of obstacles is simulated, and simulation experiments are carried out. The experimental results show that the improve