自适应避障蚁群算法的移动机器人路径规划  

Mobile Robot Path Planning Based on Adaptive Obstacle Avoidance Ant Colony Algorithm

作　　者：刘双双[1,2] 詹京吴 黄宜庆 LIU Shuangshuang;ZHAN Jingwu;HUANG Yiqing(Key Laboratory of Advanced Perception and Intelligent Control of High-end Equipment Ministry of Education,Wuhu 241000,China;Anhui Key Laboratory of Electric Drive and Control,Wuhu 241000,China)

机构地区：[1]高端装备感知与智能控制教育部重点实验室,安徽芜湖241000 [2]安徽省电气传动与控制重点实验室,安徽芜湖241000

出　　处：《安徽工程大学学报》2021年第3期27-33,共7页Journal of Anhui Polytechnic University

基　　金：安徽工程大学中青年拔尖人才基金资助项目(2016BJRC004);安徽省高校协同创新基金资助项目(GXXT-2020-069)。

摘　　要：传统蚁群算法在解决机器人路径规划的问题上存在收敛慢,容易收敛于次优路径的不足。针对以上不足,研究在状态转移概率、信息素更新方面进行改进。在状态转移概率上引入了一个避障函数因子,减小蚂蚁进入地图陷阱的概率,从而有效减少陷入死锁的蚂蚁数量,加快收敛速度。在信息素方面,加入了随迭代次数变化的信息素挥发因子,增加算法进行全局搜索的可能性,避免算法陷入局部最优。设计三种栅格障碍图对两种算法进行仿真,结果对比均表明,经过改进得到的AOA蚁群算法的初次收敛次数更少,迭代速度大大增强,并且最终规划得到的最小路径长度均比原算法小,与其他的改进算法相比,在找到最优路径的同时,其初次迭代的收敛次数也大大降低。Traditional ant colony algorithm has the disadvantage of slow convergence and easy convergence to suboptimal path in solving robot path planning problem.In order to solve the above problems,the state transition probability and pheromone update are improved.In the state transition probability,an obstacle avoidance function factor is introduced to reduce the probability of ants entering map traps,thus effectively reducing the number of ants in deadlock and accelerating the convergence speed.In the pheromone aspect,the pheromone volatilization factor which changes with the iteration times is added to increase the possibility of global search and avoid the algorithm falling into local optimum.Three kinds of grid obstacle diagrams are designed to simulate the two algorithms.The results show that the improved AOA ant colony algorithm has less initial convergence times and higher iteration speed,compared with other improved algorithms,the optimal path is found and the convergence times of the first iteration is greatly reduced.

关 键 词：蚁群算法 路径规划 避障 信息素挥发因子 移动机器人 自适应 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]