摘要
针对室内或地下等狭窄而复杂环境下的移动机器人全局路径规划,提出了一种基于Dijkstra算法的改进遗传算法路径规划策略,以解决传统遗传算法在狭窄环境下难以有效初始化的问题。首先借助Dijkstra算法得出基准路径,然后以此基准路径为基础,通过改进的编码方式与搜索空间进行初始种群的编码,最后通过遗传算法获得最优路径。提出了全局通行度和路径安全度的概念,用来评估当机器人不可视为质点时的环境状态与路径优劣。仿真实验结果表明,与传统遗传算法和人工势场法相比,本方法在保证路径距离较短的情况下,能使路径安全度提高50%以上,或者将时间复杂度降低一半以上,表明了所提方法的实用性和有效性。
To solve the global path planning problem in narrow and complex environments, the paper presented a new GA- based strategy. The initialization of path planning using the traditional GA in narrow space was difficult. First, base path search used Dijkstra algorithm; then, initial population coding by a new genetic code scheme and improved searching spaces; finally, path optimization used genetic algorithm. For the robots could not be scaled as points in narrow space, the paper proposed the global pass degree and path safety to evaluate the environment and the path. Simulation results show that compared with traditional GA and artificial potential field, this method ensures the path distance is short, increases path safety by more than 50% or reduces the time complexity by more than half. Results demonstrate that the practicality and effectiveness of the proposed method.
出处
《计算机应用研究》
CSCD
北大核心
2015年第2期413-418,共6页
Application Research of Computers
基金
国家博士后科学基金资助项目(20090460211)
国家自然科学基金资助项目(50904007)
中央高校基本科研业务费资助项目(FRF-TP-12-055A)
