摘要
为解决复杂多变环境下,常规路径规划算法生成路径长、不连贯、存在多余转折点和大转折角度、容易碰撞障碍物等问题,提出一种改进遗传算法的动态避障算法.结合行间随机选择策略和A*算法插入中间点策略改进初始化算法;在传统遗传算法中设置障碍物安全距离,引入删除和优化算子,优化适应度函数;采用动态窗口法结合改进遗传算法得到路径最优解.结果表明:改进后算法生成的路径平滑,与A*算法、某改进遗传算法相比,路径更短、碰撞次数更少,在复杂动态环境下具有良好的避障性能,为移动机器人在现实场景中的安全、高效导航提供了可行的解决方案.
In order to solve the problems of path length,incoherence,redundant turning points and large turn-ing angles,and susceptibility to collisions with obstacles generated by conventional path planning algorithms in a complex and changeable environment,a dynamic obstacle avoidance algorithm based on improved genetic al-gorithm was proposed.An interline random selection strategy and A*algorithm were combined to insert inter-mediate point strategy to improve the initialization algorithm.The safety distance of obstacles was set in the tra-ditional genetic algorithm,and the fitness function was optimized by introducing deletion and optimization op-erators.The dynamic window method combined with the improved genetic algorithm was used to obtain the op-timal path solution.The results show that the path generated by the improved algorithm is smooth;and com-pared with A*algorithm and some improved genetic algorithm,the path is shorter and the number of collisions is less.The improved algorithm has good obstacle avoidance performance in complex dynamic environment,and provides a feasible solution for safe and efficient navigation of mobile robot in real scene.
作者
翁伟
陈龙
郑祥盘
陈力雄
WENG Wei;CHEN Long;ZHENG Xiang-pan;CHEN Li-xiong(School of Intelligent Manufacturing,Fujian Information Vocational and Technical College,Fuzhou,Fujian 350002,China;College of Physics&Electronic Information Engineering,Minjiang University,Fuzhou,Fujian 350108,China;School of Mechanical and Automotive Engineering,Fujian University of Technology,Fuzhou,Fujian 350108,China)
出处
《宁德师范学院学报(自然科学版)》
2024年第2期133-142,共10页
Journal of Ningde Normal University(Natural Science)
基金
福建省技术创新重点攻关及产业化项目(校企联合类)(2023XQ018)
福建省科技厅对外合作项目(2023I0039)
校级“揭榜挂帅”项目(ZD202303)。
关键词
移动机器人
路径规划
遗传算法
融合算法
mobile robot
path planning
genetic algorithm
fusion algorithm