基于可视图与改进遗传算法的机器人平滑路径规划  

Smooth Path Planning for Robot Based on Visibility Graph and Improved Genetic Algorithm

作　　者：黄荣杰 王亚刚[1,2] HUANG Rongjie;WANG Yagang(School of Optical-electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;Shanghai Engineering Research Center of Assistive Devices,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海康复器械工程技术研究中心,上海200093

出　　处：《控制工程》2024年第4期678-686,共9页Control Engineering of China

基　　金：国家重点研发计划项目(2020YFC2007502)。

摘　　要：针对传统遗传算法在路径规划中存在收敛速度慢、易早熟和路径质量差等缺点,提出一种基于可视图与改进遗传算法的路径规划算法。首先,利用可视图法压缩地图信息,减少搜索节点;然后,对路径个体采用浮点数编码,引入模拟二进制交叉(simulated binary crossover,SBX)算子和多项式变异算子,并采用精英保留策略和轮盘赌相结合的选择算子以防止优质个体丢失;之后,将贝塞尔(Bezier)算子引入遗传算法,改善路径的平滑性;最后,分段优化贝塞尔控制节点,防止优化路径与障碍物碰撞。在仿真地图中进行测试,实验结果表明,所提算法相比于其他算法可以规划出一条更平滑、更短的路径。将算法应用在康复助行机器人中进行测试,实验结果表明,所提算法能有效解决机器人的全局路径规划问题,提升全局路径规划的效率。Considering the shortcomings of the conventional genetic algorithm(GA)in path planning,such as slow convergence speed,easy maturity and poor path quality,a path planning algorithm based on visibility graph and improved GA is proposed.Firstly,visibility graph is used to compress map information and reduce the number of search nodes.Secondly,the floating point representation is used for individual path,the simulated binary crossover(SBX)operator and polynomial mutation operator are introduced,and the selection operator combined with elite retention strategy and roulette is adopted to prevent the loss of high-quality individuals.Thirdly,the Bezier operator is introduced into genetic algorithm to improve the path smoothness.Finally,segmented Bezier control nodes are optimized to prevent collisions between the optimized paths and obstacles.The experimental results based on the simulation maps show that the proposed algorithm can plan a smoother and shorter path compared with the other algorithms.The algorithm is applied to a walking-aid rehabilitation robot,and the experimental results show that the proposed algorithm can effectively solve the global path planning problem of the robot and improve the efficiency of global path planning.

关 键 词：康复助行机器人 遗传算法 浮点数编码 贝塞尔曲线 路径规划 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]