机器人.pdf.doc

第 24卷　第 2期 吉林大学学报 (信息科学版 ) Vol. 24 No. 2 2006年 3 Journal of J ilin University ( Information Science Edition) M ar 2006 2 2 2 1 文章编号月:1671 5896 ( 2006) 02 0148 05 动态环境下机器人路径规划的混合蚁群算法 梅　昊 , 田彦涛 , 祖丽楠 (吉林大学通信工程学院 , 长春 130025) 摘要 : 为了实现动态环境下移动机器人的路径规划 , 将蚁群算法 (作为一种全局路径规划算法 ) 与人工势场 算法 (作为一种局部路径规划算法 ) 相结合 , 并对蚁群算法做了一些改变 , 使之适应栅格环境。仿真结果表 明 , 该混合蚁群算法路径性能较好 , 并能够满足实时要求。 关键词: 蚁群算法 ; 动态环境; 路径规划 中图分类号: TP18 文献标识码: A A lgorithm for Route Planning of Robot in Dynam ic Environment M E I Hao, TIAN Yan tao, ZU L i nan ( College of Communication Engineering, J ilin University, Changchun 130025, China) 2 2 2 Abstract: Ant co lony system is used as a global route p lanning algorithm , and A PF (A rtificial Potential Field) algorithm is used as a local route p lanning algorithm for mobile robot in dynam ic environm ent. Som e modifica 2 ? 2 2 tions were made to accommodate ant colony system to grid based environm ent. Simulation results showed that the hybrid algorithm can reach a p referable performance and could meet the real tim e demand. Key words: ant colony system; dynam ic environm ent; route p lanning In troduction O ne of the challenges for an intelligent robot is determ ining its fastest and safest route to its destination. This is what s known as the route p lanning p roblem. The route p lanning p roblem is an ordering p roblem , where a se quence of configurations is sought, starting from an initial location and ending at the goal location. The robot searches for an op timal or near op tim al route w ith respect to the p roblem objectives. The route p lanning p roblem criteria may include distance, time, energy and safety. Presently, there are m any search algorithm s fo r route p lanning, such as A PF (A rtificial Po tential Field ) [ 1 ] , FL ( Fuzzy Logic ) [ 2 ] , NN (Neural N etworks) [ 3 ] , GA ( Genetic A lgo