基于粒子群优化算法的月面巡视器全局路径规划

Global Path Planning for Lunar Rover Based on Particle Swarm Optimization Algorithm

在月面巡视器遥操作系统中,路径规划分为任务级路径规划、全局路径规划和局部路径规划。根据巡视器全局路径规划的应用要求,引入粒子群优化算法应用于全局导航点的规划。针对粒子群算法在路径规划中容易造成不收敛或病态收敛的问题,对算法进行了修改,去掉了速度更新中的速度惯性因子,只保留自身认识因子和社会认识因子,使其在全局路径规划中能够快速收敛;同时引入经典遗传算法中的变异因子以增强算法的全局优化能力。仿真结果表明该算法具有计算简单、全局寻优能力强等特点,能够快速地找到优化的全局导航点。同时在不同的模拟月面地形上进行仿真试验,针对存在的问题提出了对应的二次优化方法,结果表明该方法较好地满足了巡视器全局路径规划的应用需求。

In the tele-operation system of lunar rover,the path planning contains three levels： mission-level path planning,global path planning and local path planning.Based on the requirements of the global path planning of the lunar rover,the Particle Swarm Optimization（PSO） algorithm is introduced in the global navigation point planning.Since the PSO algorithm may converge ill or not converge in path planning, the algorithm is modified.In the modified algorithm,the velocity inertial weight is deleted,but the cognitive and social coefficients are kept,with the aim at making the algorithm converge quickly in path planning.Also the variation coefficient in evolution algorithm is imported to enhance the global optimization ability.Simulation results show the improved algorithm is simple and has high ability to find the best path.Also simulation tests are done in several different simulated lunar terrain maps,and optimization methods are given to make the planning result better.