基于Hopfield神经网络的自适应控制系统的设计与仿真

Design and Simulation of Self-Adaption Control System Based on Hopfield Neural Network

根据连续型Hopfield神经网络的特性,设计基于Hopfield神经网络自适应控制系统,解决当前大多数控制系统需要外界参与的问题.设计一个三元组的Hopfield神经网络,并通过自反馈机制更新神经元的权重,完成自适应控制的任务.通过MATLAB平台仿真建立Hopfield神经网络,构建神经网络输出与参考标准输出之间的对比实验.结果表明,Hopfield能够在有限次数内逼近参考标准输出,从而完成控制任务.基于Hopfield神经网络的自适应控制系统有较高的精度,能够完成常见的设备控制,具有较强的可行性和便捷性.

A novel control system was designed by means of Hopfield neural network to solve the problem of control systems which are need external parameters. A Hopfield neural network was designed by three neurons,and the selffeedback mechanism to update neurons＇ weights was used for the Hopfield neural network to implement the assignment of controlling. The simulation of Hopfield neural network was built by MATLAB platform,and the comparison of neural network output and reference standard output was built for the results. The results showed that the self-adaption control system based on Hopfield neural network has better precision of controlling. And it can complete common device control and has better feasibility and convenience.