摘要
针对目前蚁群算法在路径规划策略中存在的易陷入凹陷、迭代时间长、转角多、避碰时等待时间长等缺点,提出了一种基于视野域的动态快速路径规划的蚁群算法。首先,采用栅格填充法对建立的栅格图中复杂形状障碍物存在的过多凹陷进行填充;其次,采用改进的负反馈的蚁群算法来减少蚁群在无法找到有效终点方向的后续搜索;再次,模拟给蚂蚁加上视觉以便在特定区域搜索而减少过多的转折点,从而更快速地找出最优路径。同时,在行走过程中若存在障碍物动态变化,则结合动态窗口算法重新规划局部的最优路径。实验表明采用该规划算法可使路径长度减少11%,转角和行走时间减少45.4%和32.3%。所以该智能算法能够为机器在动态环境的自主规划与导航中提供一种可行的解决方法。
In view of the shortcomings of current ant colony algorithm in path planning strategy,such as easy to fall into a depression,long iteration time,many corners,and long waiting time for collision avoidance,a dynamic fast path planning ant colony algorithm based on field of vision is proposed.The grid filling method is used to fill the too many depressions of the complex shape obstacles in the grid graph.The improved ant colony algorithm with negative feedback is used to reduce the subsequent search when the ant colony cannot find the effective destination direction.The simulation adding vision to ants is conducted to search in specific areas and reduce too many turning points,so as to find the best route more quickly.At the same time,if there are dynamic changes of obstacles during walking,the dynamic window algorithm is combined to re-plan the local optimal path.The experimental results show that the path length can be reduced by 11%,and the turning angle and the walking time can be reduced by 45.4%and 32.3%by means of this planning algorithm.This intelligent algorithm can provide a feasible solution for the autonomous planning and navigation of machines in dynamic environment.
作者
钟志峰
周霖
柯伟
ZHONG Zhifeng;ZHOU Lin;KE Wei(College of Computer and Information Engineering,Hubei University,Wuhan 430062,China)
出处
《现代电子技术》
2023年第24期85-93,共9页
Modern Electronics Technique
基金
湖北省技术创新专项(2018ACA13)。
关键词
动态路径规划
视野域
蚁群算法
栅格填充法
负反馈
动态窗口算法
dynamic path planning
vision field
ant colony algorithm
grid filling
negative feedback
dynamic window algorithm
