基于改进混合A*与融合平滑策略的无人车路径规划OA
Unmanned Vehicle Path Planning Based on Improved Hybrid A*and Integrated Path Smoothing
针对无人车在复杂环境中存在的搜索效率偏低、路径平滑性不足等问题,提出一种改进混合A*算法与融合平滑策略的路径规划方法.提出分级三分法节点扩展策略,通过评估初始扩展节点优劣动态调整搜索方向与角度,提升复杂环境下的寻优能力.提出融合Douglas-Peucker算法与B样条曲线的平滑策略,在剔除路径冗余特征点的同时进行关键点拟合,生成曲率连续且满足车辆运动学约束的平滑路径.仿真结果表明,在复杂障碍物环境中,改进的节点扩展策略使无人车计算时间缩短了19.61%,路径长度减少了13.11%;融合平滑策略相比B样条平滑路径平均曲率降低了28.57%,累积曲率变化降低了68.76%.相比于改进RRT算法,本文算法在密集障碍物地图中的扩展点数减少了约43.72%,有效提升了无人车路径规划的效率和路径平滑性.通过ROS 阿克曼小车试验平台验证了算法在真实物理环境中的规划能力,生成的路径具有更好的几何连续性和运动学可行性.
To address the low search efficiency and insufficient path smoothness of unmanned vehicles in complex environments,a path planning method combining an improved Hybrid A* algorithm with an integrated smoothing strategy is proposed.First,a hierarchical ternary node expansion strategy is developed,in which the search direction and steering angle are dynamically adjusted by evaluating the quality of initially expanded nodes,thereby improving the search and optimization capability in complex environments.Second,an integrated smoothing strategy combining the Douglas-Peucker algorithm and B-spline curves is proposed.This strategy removes redundant feature points from the path and fits the remaining key points to generate a smooth path with continuous curvature that satisfies vehicle kinematic constraints.Simulation results show that,in complex obstacle environments,the improved node expansion strategy reduces the computation time by 19.61%and the path length by 13.11%.Compared with B-spline smoothing,the proposed integrated smoothing strategy reduces the average curvature by 28.57%and the cumulative curvature variation by 68.76%.Compared with an improved RRT algorithm,the proposed algorithm reduces the number of expanded nodes by approximately 43.72%in dense-obstacle maps,effectively improving both path planning efficiency and path smoothness for unmanned vehicles.In addition,experiments on a ROS-based Ackermann-steering vehicle platform verify the planning capability of the proposed algorithm in a real physical environment,and the generated paths exhibit better geometric continuity and kinematic feasibility.
刘永涛;赵哲宇;王淞;纳林奇;王泰琪;白凤鹏;宋贵云
长安大学 汽车学院,西安 710064长安大学 汽车学院,西安 710064武警警官学院 信息通信系,成都 610213长安大学 汽车学院,西安 710064长安大学 汽车学院,西安 710064抚顺市交通运输综合行政执法队,辽宁,抚顺 113009新疆美克化工股份有限公司,新疆,库尔勒 841000
交通工程
路径规划混合A*算法B样条曲线Douglas-Peucker算法
path planninghybrid A* algorithmB-spline curveDouglas-Peucker algorithm
《汽车工程学报》 2026 (5)
779-791,13
教育部基础学科和交叉学科突破计划项目(JYB2025XDXM104)咸阳市重大科技创新专项(L2025-ZDKJ-ZDGG-RGZN-003)咸阳市渭城区科技发展计划项目(2025VCCN-03)西安市自然科学基金项目(2025JH-ZRKX-0617)
评论