摘要
批处理知情搜索树(batch informed trees,BIT*)作为一种先进的采样规划方法通常被应用于移动机器人的路径规划.针对BIT*在初始路径获得后存在路径代价降低速度不快、规划效率有待提高的问题,提出了一种基于偏置采样和包围优化的BIT*(wrapping-based biased BIT*,WB-BIT*)方法.该方法首先通过批量采样进行节点和连接的扩展,在获得可行路径后,利用包围优化策略从目标点到起始点逐步使路径靠近至障碍物周围,快速减少现有可行路径的长度.同时,根据可行路径上的路径点计算启发式函数以构建偏置采样区域,结合偏置采样和知情集采样,在保证均匀性的前提下有效运用现有路径信息,提高方法的规划效率.最后,将所提出的WB-BIT*方法与主流采样路径规划方法进行仿真实验对比,结果表明所提出的路径规划方法具备更高的规划效率.
As an advanced sampling-based planning method,batch-informed search tree(BIT*)is generally applied to mobile robots path planning.Considering problems of slow speeds of cost reducing and low planning efficiencies after BIT*finding an initial solution,this article proposes a path planning method based on biased sampling and wrapping optimization(Wrapping-based Biased BIT*,WB-BIT*).This method first expands nodes and connections through batch sampling.After obtaining a feasible path,the wrapping optimization strategy is adopted to allow the feasible path to closely approach obstacles from the target point to the starting point gradually.As a result,the length of the current feasible path is quickly reduced.Then,based on path points of the current feasible path,a novel heuristic function is introduced and calculated to construct the biased sampling area.Aided with the combination of biased sampling and informed set sampling,the current path information can be effectively utilized to improve the planning efficiency under the premise of sampling uniformity.Finally,WB-BIT*is compared with other popular sampling-based planning algorithms in simulation environments.Results show that the planning efficiency of the proposed WB-BIT*exceeds that of traditional algorithms.
作者
陈彦杰
梁景林
张智星
喻骁
王耀南
CHEN Yanjie;LIANG Jinglin;ZHANG Zhixing;YU Xiao;WANG Yaonan(School of Mechanical Engineering and Automation,Fuzhou University,Fuzhou 350108,China;School of Aerospace Engineering,Xiamen University,Xiamen 361102,China;College of Electrical and Information Engineering,Hunan University,Changsha 410082,China;National Engineering Research Center of Robot Visual Perception and Control Technology,Changsha 410082,China)
出处
《厦门大学学报（自然科学版）》
CAS
CSCD
北大核心
2022年第6期908-915,共8页
Journal of Xiamen University：Natural Science
基金
国家自然科学基金(62273098,62027810)
福建省自然科学基金(2021J01051)
天津大学-福州大学自主创新基金(TF2022-4)。
