基于神经网络和遗传算法的移动机器人路径规划  

Path Planning of Mobile Robot Based on Neural Network and Genetic Algorithm

作　　者：徐美清[1] 刘洞波[2] Xu Meiqing;Liu Dongbo

机构地区：[1]湖南铁路科技职业技术学院,湖南株洲412006 [2]湖南工程学院,湖南湘潭411101

出　　处：《湖南铁路科技职业技术学院学报》2020年第2期87-90,共4页Vocational Education Research on Rail Transit

摘　　要：本文提出一种将遗传算法和神经网络综合起来以完成未知动态环境下移动机器人路径规划的方法。该方法中,机器人的移动由神经网络来控制,先通过前端装载的五个传感器来获取所处环境信息,将获取的障碍、位置及目标信息作为神经网络输入,然后利用遗传算法训练调整神经网络权值,最后将训练调整后的神经网络输出作为机器人的驱动控制力。仿真结果表明采用该方法,机器人能在较短的时间内寻找到一条平滑避障路径,完成从起始点到目标点的平滑高效移动。This paper proposes a method that integrates genetic algorithm and neural network to complete the path planning of mobile robot in unknown dynamic environment.In this method,the movement of the robot is controlled by a neural network.Firstly,the environment information is obtained by five sensors loaded in the front end.Then,the information of obstacles,positions and targets is taken as the input of neural network.Then,genetic algorithm is used to train and adjust the weight of neural network.Finally,the output of trained neural network is used as the driving force of robot.The simulation results show that using the method in this paper,the robot can find a smooth obstacle avoidance path in a short time and complete the smooth and efficient movement from the starting point to the target point.

关 键 词：神经网络 遗传算法 移动机器人 路径规划 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]