基于变步长蚁群算法的移动机器人路径规划

PDF下载HTML

Title:

Mobile robot path planning based on variable-step ant colony algorithm

作者:

徐玉琼¹, 娄柯², 李志锟^2,3

1. 广州大学松田学院 电气与汽车工程系，广东 广州 511370;
2. 高端装备先进感知与智能控制教育部重点实验室，安徽 芜湖 241000;
3. 安徽工程大学 安徽省电气传动与控制重点实验室，安徽 芜湖 241000

Author(s):

XU Yuqiong¹, LOU Ke², LI Zhikun^2,3

1. Department of Electrical and Automotive Engineering, Songtian College, Guangzhou University, Guangzhou 511370, China;
2. Key Laboratory of Advanced Perception and Intelligent Control of High-End Equipment, Ministry of Education, Wuhu 241000, China;
3. Anhui Provincial Key Laboratory of Electric Transmission and Control, Anhui Polytechnic University, Wuhu 241000, China

关键词:

传统蚁群算法; 双层蚁群算法; 路径规划; 变步长; 信息素; 启发函数; 收敛; 移动机器人

Keywords:

traditional ant colony algorithm; double-layer ant colony algorithm; path planning; variable-step; pheromone; heuristic function; convergence; mobile robot

分类号:

TP242.6

DOI:

10.11992/tis.202004011

摘要:

针对传统蚁群算法以及双层蚁群算法在路径规划中存在搜索效率低、收敛性较慢以及成本较高的问题，本文提出了变步长蚁群算法。该算法扩大蚁群可移动位置的集合，通过对跳点的选择以达到变步长策略，有效缩短移动机器人路径长度；初始化信息素采用不均匀分布，加强起点至终点直线所涉及到栅格的信息素浓度平行地向外衰减；改进启发式信息矩阵，调整移动机器人当前位置到终点位置的启发函数计算方法。试验结果表明：变步长蚁群算法在路径长度及收敛速度两方面均优于双层蚁群算法及传统蚁群算法，验证了变步长蚁群算法的有效性和优越性，是解决移动机器人路径规划问题的有效算法。

Abstract:

To address the problems of the traditional and double-layer ant colony algorithms, such as their low search efficiency, slow convergence, and high path–planning cost, in this paper we propose a variable-step ant colony algorithm. The proposed algorithm expands the set of mobile locations of the ant colony, and uses the variable-step strategy of selecting the hopping points, thus effectively shortening the path length of the mobile robot. The initialization pheromone adopts an uneven distribution, which increases the pheromone concentration of the grid in a straight line from the start to end points, with the pheromone decaying outward in parallel. The heuristic information matrix is improved and the method used to calculate the heuristic function of the mobile robot from the current to the end positions is adjusted. The experimental results show that the performance of the variable-step ant colony algorithm is superior to those of the double-layer and traditional ant colony algorithms with respect to path length and convergence speed, which proves its effectiveness and superiority. Thus, the proposed algorithm is effective in solving the path-planning problem of mobile robots.