基于蚁群算法与参数迁移的机器人三维路径规划方法

Robot 3D Path Planning Method Based on Ant Colony Algorithm and Parameter Transfer

为解决机器人进行三维路径规划时路径规划效率受算法影响较大的问题,以最短距离为目的,提出了一种基于蚁群算法参数迁移的机器人三维路径规划方法。在使用栅格法建立的机器人三维环境模型中,使用蚁群算法寻找最短路径。针对蚁群算法的参数选择问题,使用参数迁移算法得到最优参数。将已知的环境模型和其对应的蚁群最优参数作为源任务,将源任务映射到高维空间,通过迁移参数连接不同源任务,根据图论的知识建立参数迁移图,将参数迁移图扩展,使其包含目标任务,为随机未知环境模型分配一组蚁群最优参数。仿真实验表明,基于参数迁移的蚁群算法可以快速有效地完成机器人三维路径规划。相比传统的参数选择方法和其他智能优化方法,蚁群参数迁移算法可以大幅减少路径规划所需时间,提高了路径规划效率。

In the process of three-dimensional( 3 D) path planning for robots,the efficiency of path planning is greatly affected by the algorithm itself. For the purpose of the shortest distance,a robot 3 D path planning method was proposed based on ant colony parameter transfer algorithm. Ant colony algorithm was used to find the shortest path in the robot environment model which was established by grid method. For the parameter selection problem of the ant colony algorithm,the parameter transfer algorithm was used to obtain the optimal parameters. The known environment model and its corresponding ant colony optimal parameters were used as source tasks,and the source tasks were mapped to highdimensional spaces. Connecting different source tasks through transfer parameters,the parameter transfer graph was established based on the knowledge of graph theory. The parameter transfer map was extended to include the target task,and a set of ant colony optimal parameters was assigned to the random unknown environment model. Simulation results showed that the ant colony algorithm based on parameter transfer can complete the robot 3 D path planning quickly and effectively. Compared with the traditional parameter selection method and other intelligent optimization methods,the ant colony parameter transfer algorithm can greatly reduce the time required for path planning and improve the path planning performance.