平滑的
运动规划
任意角度路径规划
数学优化
算法
路径(计算)
最短路径问题
Dijkstra算法
计算机科学
计算
还原(数学)
避障
贪婪算法
数学
图形
人工智能
机器人
移动机器人
理论计算机科学
计算机视觉
程序设计语言
几何学
作者
Zhu’An Zheng,Shuangjian Xie,Zimo Ye,Xiangyu Zheng,YU Zhi-wei
标识
DOI:10.1177/09544062241305524
摘要
Aiming at the deficiencies of the original RRT-Connect path planning algorithm in dealing with obstacle avoidance, planning efficiency and path smoothing in static environments, an improved path optimisation method is proposed by fusing the RRT-Connect path planning algorithm with the greedy search strategy and using a third-order Bezier curve for path smoothing. Firstly, a greedy algorithm is added to the path planning process of the original RRT-Connect algorithm to guide the search direction, so as to make the path search more goal oriented, reduce the time of path planning and improve the efficiency. Secondly, the generated paths are smoothed with third-order Bezier curves to ensure the generation of smooth paths with continuous curvature and improve the quality of the planned paths. Finally, the improved RRT-Connect smoothing optimisation algorithm is simulated with the original RRT-Connect, A* and Dijkstra algorithms in different simulation environments to check the performance of the optimised algorithm. The results show that the success rate of obstacle avoidance in planning paths of the optimised algorithms are all 100%, and compared with the original RRT-Connect algorithm, the improved RRT-Connect smoothing optimisation algorithm has an overall reduction of 12.02% in planning path length, an overall reduction of 40.22% in computation time and an overall improvement of 69.15% in path smoothness. Compared to the A* algorithm, which is based on the graph search shortest path planning algorithm and takes the shortest time, there is an overall reduction of 62.46% in computation time. The improved RRT-Connect smoothing optimisation algorithm has shorter path planning lengths, smoother paths and shortest computation time in different simulation environments. Meanwhile, the stability of the optimisation algorithm is effectively improved in different environments and different running times.
科研通智能强力驱动
Strongly Powered by AbleSci AI