避障
运动规划
序列二次规划
数学优化
避碰
平滑的
弹道
障碍物
计算机科学
平滑度
二次规划
势场
功能(生物学)
路径(计算)
领域(数学)
控制理论(社会学)
算法
碰撞
机器人
移动机器人
数学
人工智能
计算机视觉
地质学
天文
物理
程序设计语言
政治学
进化生物学
控制(管理)
生物
数学分析
法学
地球物理学
纯数学
计算机安全
作者
Ping Qin,Fei Liu,Zhizhong Guo,Zhe Li,Yuze Shang
标识
DOI:10.1177/01423312231186684
摘要
To enable autonomous vehicles to generate smooth and collision-free trajectories and improve their driving performance on structured roads, this paper proposes a hierarchical trajectory planning algorithm based on an improved artificial potential field method. To improve the applicability of the algorithm to complex scenarios, the Frenet coordinate system was established to address these limitations. First, the safety distance model is applied to the risk assessment of the improved artificial potential field method. Then, the hierarchical solution is carried out, and the road solvable convex space and the rough path solution are solved by combining the artificial potential field method. On this basis, the potential field term and the smoothing term cost function are established, and the sequential quadratic programming (SQP) algorithm is used to solve the exact path that meets the requirements of safety and smoothness. Hierarchical planning shortens the solution time by quickly determining the bounds of the convex space. Finally, in the speed planning, in order to take into account the comfort and safety of the occupants, the speed curve is solved by considering the dynamic constraints of the vehicle. The obstacle avoidance effects of the algorithm on static and dynamic obstacles are tested in different simulation scenarios. The results of the simulation experiment show that the proposed algorithm can successfully achieve obstacle avoidance on complex structured roads.
科研通智能强力驱动
Strongly Powered by AbleSci AI