摘要
机器人在执行各种任务中需要具备运动控制、定位、建图与路径规划的能力,针对传统^(*)算法在路径规划中容易陷入局部最优点和转折较多问题,提出了一种改进A^(*)(A-Star)和动态窗口法(Dynamic window approach,DWA)相结合的混合算法。移动机器人平台以激光传感器作为核心传感器,获得环境的二维信息,针对同步定位与建图问题,改进了Lazy Decision算法提高回环质量、减少计算量。针对机器人在路径规划中存在的问题,在A;算法中引入预测距离,设定动态衡量启发式A^(*)算法中的h(n)权重系数,解决了使用A;算法扩展节点多、容易陷入局部最优的问题,同时对路径规划中拐角进行修正,有效地将路径规划时间降低14%。并在由运动底盘、2D激光雷达传感器和计算机运算平台组成的自主移动机器人系统试验平台上验证了方法的有效性,不仅缩短运行时间、优化了路径复杂度,还有效克服了传统路径规划中转折角多、转折角度大的缺点。
Robots need to have the ability of motion control,positioning,mapping and path planning in performing various tasks.Aiming at the problem of traditional A^(*) algorithms that are easy to fall into local optimal points and turning points in path planning,an improved A;(A-Star)and dynamic window approach(DWA)combined hybrid algorithm.The mobile robot platform uses laser sensors as the core sensor to obtain two-dimensional information of the environment.Aiming at the problem of synchronous positioning and mapping,the Lazy Decision algorithm is improved to improve the loop quality and reduce the amount of calculation.In view of the problems in robot path planning,the prediction distance is introduced into the A;algorithm,and the h(n)weight coefficient in the dynamic weighing heuristic A^(*) algorithm is set to solve the problem of using A;algorithm to expand the number of nodes and easily fall into local optimal problem,while correcting the corners in the path planning,effectively reducing the path planning time by 14%.The effectiveness of the method is verified on an autonomous mobile robot system experimental platform composed of a moving chassis,2D LIDAR sensors and a computer computing platform.The method not only shortens the running time,optimizes the path complexity,and effectively overcomes disadvantages of a large turning angle and multiple turning angles of traditional path planning methods.
作者
杨明亮
李宁
YANG Mingliang;LI Ning(School of Mechanical Engineering,Taiyuan University of Science and Technology,Taiyuan 030024,China)
出处
《机械科学与技术》
CSCD
北大核心
2022年第5期795-800,共6页
Mechanical Science and Technology for Aerospace Engineering
基金
山西省重点研发计划(2020XXX010)。
作者简介
杨明亮(1976-),教授,博士,研究方向为移动机器人路径规划,yangmingliang@tyust.edu.cn。
