基于目标导向的双向RRT路径规划算法  

Goal-oriented BI-RRT Path Planning Algorithm

作　　者：韩丰键 邱书波[2,3] 冯超 曹启贺 李庆华 HAN Feng-jian;QIU Shu-bo;FENG Chao;CAO Qi-he;LI Qing-hua(School of Electrical Engineering and Automation,Qilu University of Technology(Shandong Academy of Sciences),Jinan 250353,China;School of Electronic and Information Engineering(Department of Physics),Qilu University of Technology(Shandong Academy of Sciences),Jinan 250353,China;Jinan Engineering Laboratory of Human-machine Intelligence Cooperation,Jinan 250353,China)

机构地区：[1]齐鲁工业大学(山东省科学院)电气工程与自动化学院,济南250353 [2]齐鲁工业大学(山东省科学院)电子信息工程学院(大学物理教学部),济南250353 [3]济南市人机智能协同工程实验室,济南250353

出　　处：《齐鲁工业大学学报》2021年第1期35-43,共9页Journal of Qilu University of Technology

基　　金：国家自然科学基金(61701270);齐鲁工业大学(山东省科学院)青年博士合作基金(2017BSHZ008)。

摘　　要：针对双向快速扩展随机树(BI-RRT)算法在路径规划中存在目标导向性差、收敛速度缓慢的问题,提出了一种目标导向的BI-RRT算法(goal-oriented BI-RRT,GOBI-RRT)。通过目标导向思想对随机树中采样点的产生进行改进,引导随机树偏向目标点生长,提高了收敛速度,降低了扩展过程的复杂度。同时提出了一种圆盘k点碰撞检测算法,对圆盘式机器人是否和障碍物产生碰撞进行检测。在多个场景中进行了圆盘移动机器人的路径规划仿真实验,实验结果验证了所提算法的可行性和有效性。To solve the problem of poor target orientation and slow convergence of BI-RRT algorithm in path planning,a goal-oriented BI-RRT(GOBI-RRT)algorithm was proposed.Base on goal-oriented thought,the generation of sampling points in the random tree is improved,the random tree is guided toward the target point to improve the convergence speed and the complexity of the expansion process can be reduced.At the same time,a disk k-point collision detection algorithm is proposed to detect whether the disk robot collides with obstacles.The path planning simulation experiments of the disk mobile robot are carried out in several scenarios,the experimental results verify the feasibility and effectiveness of the proposed algorithm.

关 键 词：路径规划 双向快速扩展随机树算法 目标导向 圆盘移动机器人 碰撞检测 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]