摘要
针对传统蚁群算法进行机器人路径规划时具有收敛性差,易陷入局部最优等不足,提出一种多步长蚁群算法,该算法使机器人可根据其所处的环境选择步长,并找到一条长度短、行走步数少的路径.通过栅格法对机器人的工作环境进行建模,确定机器人的可选步长,改进启发信息的构成及信息素的更新方式,同时采用最大-最小蚂蚁思想来限制信息素强度防止早熟收敛现象的产生.大量的仿真结果表明改进的蚁群算法能够找到一条比传统蚁群算法路径更短,行走步数更少的路径,改进的算法更智能、高效,能够显著的提高机器人的路径规划性能.
The problems of traditional ant colony algorithm are searching so long and getting into local optimal easily, we prevent a multi-step ant colony algorithm, the robot have many step size can choose in accordance with the environment, find a shorter and less moves path. Modeling the environment with grid method, determining the robot can choose the length of step-size,improving the way of updating pheromone, adapting max-rain ant system which can prevent the precocious by limiting global pheromone intensity, many simulation results show the improved ant colony algorithm can find a shorter and less moves path,and the algorithm is more intelli- gent, efficient, which improves the performance of path planing.
出处
《小型微型计算机系统》
CSCD
北大核心
2016年第2期366-369,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(614663032)资助
江西省科技厅项目(20141BBA10035)资助
关键词
多步长
蚁群算法
路径规划
机器人
multi-step
ant colony algorithm
path planning
robot
