摘要
智能车辆在多障碍物非结构化场景中,传统Informed RRT*算法存在计算效率低、路径平滑性差的问题。为此,提出了一种基于密度聚类算法与三次B样条的Informed RRT*路径规划方法。首先,设计基于DBSCAN算法的障碍物聚类方法,简化多障碍物非结构化场景。然后,提出基于三次B样条的平顺性方法,使生成的路径平均曲率降低,路径平滑点增加。最后,基于MATLAB进行仿真,搜索出一条更平滑的路径。结果表明,在多障碍物非结构化场景中,本文提出的改进Informed RRT*算法相对传统算法效率提高了28.54%,路径最大曲率由0.14优化到0.12,取得显著效果。
In the unstructured scenario of multiple obstacles,the traditional Informed RRT*algorithm has the problems of low computational efficiency and poor path smoothness.In this paper,we propose an Informed RRT*path planning method based on density-based clustering(DBSCAN)and cubic B-splines.Firstly,an obstacle clustering method based on DBSCAN algorithm was designed to simplify the unstructured scenario of multiple obstacles.Then,a smoothness method based on cubic B-spline is proposed,which reduces the average curvature of the generated path and increases the smoothness point of the path.Finally,the simulation based on MATLAB was carried out to find a smoother path.The results show that the improved Informed RRT*algorithm proposed in this paper is 28.54%more efficient than the traditional algorithm in the unstructured scenario with multiple obstacles,and the path maximum smoothness optimized from 0.14 to 0.12,which achieves remarkable results.
作者
王维强
黄婷婷
WANG Weiqiang;HUANG Tingting(School of Automotive and Traffic Engineering,Wuhan University of Science and Technology,Wuhan,Hubei 430065,China)
出处
《计算技术与自动化》
2025年第4期1-6,共6页
Computing Technology and Automation
