基于改进粒子群算法的移动机器人路径规划方法研究  

Research of the Path Planning Method for Mobile Robots Based on Active Particle Swarm Optimization

作　　者：胡玉兰[1] 姜明洋[1] 赵慧静[1] 

机构地区：[1]沈阳理工大学信息科学与工程学院,辽宁沈阳110168

出　　处：《计算机工程与科学》2009年第6期139-141,共3页Computer Engineering & Science

基　　金：国家863计划资助项目(2003AA411340);国家自然科学基金资助项目(60473134)

摘　　要：针对移动机器人传统路径规划算法效率不高、寻优能力差等问题,本文提出一种基于改进粒子群优化算法(PSO)的移动机器人路径规划方法。该方法采用神经网络训练碰撞罚函数,得到无碰撞路径,然后采用粒子群优化算法解决路径的最优问题。利用神经网络实现大量的并行和分布计算,发挥PSO简单、容易实现的优点,提高了路径规划的计算效率和可靠性。仿真结果表明,这种新路径规划方法是可行且有效的。The quality and efficiency of calculation are two puzzling problems in the traditional algorithms for robot path planning. In this paper, a new method of path planning for robots based on active particle swarm optimization is proposed. It adopts the neural network training collision penalty function to search for a collision-free path, and then uses particle swarm optimization to resolve the best path issue. This method utilizes neural networks to realize massively parallel and distributed computing, which improves the computational efficiency and reliability. As it is proved by analysis and simulations, a better result is obtained by the proposed algorithm.

关 键 词：粒子群优化算法 神经网络 路径规划 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]