基于进化神经网络的移动机器人路径规划方法  

A method based on evolutionary neural network for mobile robot path planning

作　　者：宋勇[1,2] 李贻斌[1] 刘冰[2] 

机构地区：[1]山东大学控制科学与工程学院,济南250061 [2]山东大学威海分校信息工程学院,山东威海264209

出　　处：《中国科技论文在线》2010年第1期1-5,共5页

基　　金：国家自然科学基金(60675044)

摘　　要：针对进化机器人路径规划的不可移植性进行了研究:在某个环境里进化好的行为对于变化了的环境不能适应,需要重新进化与学习。利用神经网络构建移动机器人传感器输入和执行器输出之间的映射关系,提出了一种基于神经网络的进化机器人的路径规划算法。该算法结合反应行为与行为学习的复合体系结构,将距离传感器信息和决策量作为样本库完成反应式行为,利用改进进化算法优化神经网络权值,在进化过程中,将新出现的样本不断地加入样本库,从而使机器人的高层智能行为逐步转换为低层本能行为。给出了算法的具体步骤,仿真实验结果证明该算法行之有效。To improve the adaptability for the behavior of mobile robot in unknown environments,the mapping relation was constructed between input of sensors and output of actuator based on neural network.An algorithm of path planning was presented based on neural network for evolutionary robots combined the composite architecture with responsive behavior and learning.The responsive behavior was derived from the samples composed with sensor information and decision.The weights of neural network were evolved via the improved evolutionary algorithm.The new samples were affiliated to the database constantly.The high intelligent behaviors of robot were transmitted to responsive behaviors.The detailed process to apply the algorithm was presented.The experimental results indicated that the proposed approach was feasible.

关 键 词：进化机器人 路径规划 神经网络 进化算法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]