基于神经模糊势场法的足球机器人路径规划  

Soccer Robot Path Planning Based on Neural Network Fuzzy Potential Field Method

作　　者：林志雄[1] 张莉[1] 

机构地区：[1]西安工程大学,陕西西安710048

出　　处：《计算机仿真》2014年第1期416-420,437,共6页Computer Simulation

基　　金：陕西省教育厅专项基金项目(2010JK570)

摘　　要：针对足球机器人路径规划问题的实时性和对抗性的难点,为实现足球机器人在动态比赛环境下自主避障,解决传统人工势场法存在的局部极小点问题,提出一种自适应动态神经模糊人工势场法。首先为解决机器人路径规划所需高实时性及达到自主避障问题,采用具有实时性优势的人工势场法,其次针对本方移动机器人、敌方机器人和足球之间的特殊动态关系下出现的极小值点,引入受相对位置和相对速度影响的势场力分量,并设计了调节势场力大小的自适应神经模糊控制器实现局部路径规划,使得足球机器人在动态比赛环境下规划出最优最快的到达目标点的无碰撞路径,并通过仿真验证了改进方法在动态路径规划中的可行性和有效性。For the difficulty of the real-time and confrontational characters in the path planning problem of robot soccer, in order to solve the local minimum paint problem which exists in the traditional artificial potential field meth- od, this paper gave an adaptive dynamic neural-fuzzy artificial potential field method which can make the robot soccer avoid the obstacles indepently in the dynamic envirosnment. Firstly, to improve the high real-time character and make it avoid the obstacles indepently, this paper put forward an artificial potential field method which has the advan- tage of real-time character. Secondly, this paper introduced two potential field force vectors effected by relative posi- tion and relative speed for the minimum point happened in the special dynamic relationship among our robot soccer, enemy robot soccer and the soccer itself. Then we designed an adaptive neural-fuzzy controller which can adjust the field force to realize the local path planning and let the robot soccer planning the best and fastest way to achieve the target without collisions. At last, this paper verified the feasibility and effectiveness of the advanced method through simulation.

关 键 词：足球机器人 路径规划 人工势场 模糊控制 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]