基于神经网络和PSO的机器人路径规划研究

提出一种神经网络和粒子群算法相结合的移动机器人路径规划方法。采用小波网络和RBF网络相结合的四层神经网络结构,克服了传统神经网络方法进行路径规划时对每个障碍均设计一些特定的隐节点,当障碍较多且环境动态时,网络结构庞大且神经元的阈值随时间的变化而需要不断改变的缺点。利用粒子群对神经网络的参数进行训练,在规定的代数内对网络参数优化,使得机器人在移动过程中能够快速响应环境的变化。通过对移动机器人在动、静态不同环境下的仿真实验,证明了方法的有效性。

Path Planning of Robot Based on Neural Network and PSO

A new method of neural network and particle swarm algorithm based mobile robot path planning was proposed. With combination of the advantages of wavelet network and RBF network, a four layers neural network was designed. In conventional method, many hidden cells should design for every obstacle according to information of blocks, and the scale of network was very large with many obstacles. So PSO was used to train the parameters of neural network with its character of quick optimization to make the robot respond quickly to the dynamic environment. At last, the effectiveness of the method was proved by simulation experiments of mobile robotic in dynamic and static environments.