遗传算法在机器人路径规划中的应用研究

采用栅格法对机器人工作空间进行划分，用序号标识栅格，并以此序号作为机器人路径规划参数编码，应用遗传算法对机器人路径规划作了研究。文中引入间断无障碍路径新概念以简化初始种群产生，定义了插入算子和删除算子以保证路径的连续性和简明性。应用遗传算法工具箱ＮＰＵＧＡＴｏｏｌｂｏｘＶｌ．０对所提方法作的仿真研究结果表明，该方法可行。

Robot Motion Planning Using Genetic Algorithms

The paper presents a new method for robot motion planning based on the genetic algorithms (GA). The robot is supposed to move in a twodimensional workspace with some obstacles in it. The grids are used to discrete the twodimensional workspace. Our method is believed to be unique in the following aspects compared with the existing methods in China. First, sequence number of the grid, rather than the usual binary coding, is used to code the moving path of the robot. The sequence number is so defined that one grid corresponds to only one sequence number. Secondly, a new concept about noncontinuous but obstacle free path is put forward to simplify the generation of the initial population of GA. That is, the initial chromosomes are generated by using a string of sequence numbers, with the corresponding grids not continuous in space but free of obstacle. Finally, an insertion and a deletion operator of GA are defined to ensure that each final path is continuous and nonredundant for the robot to follow. A criterion based on the distance that the robot moves among the grids is used as the objective function. Simulation results using NPU GA Toolbox Vl.0 show that the method is effective in optimization and good in performance.