基于Elman网络的非线性系统神经元自适应预测控制

提出在非线性系统的E1man网络辨识模型的基础上，用单神经元设计预测控制器的方案。Elman网络在BP网络的基础上，加入反馈信号，利用内部状态反馈来描述系统的非线性动力学行为，提高了学习速度，适合于动态系统的实时辨识。神经元结构简单，且有很强的自学习和自适应能力，它根据系统的期望输出与一步超前预测输出之间的偏差，并通过某种特定的学习算法在线调整控制器的参数，使控制器能够适应对象参数的变化，从而实现对一类非线性系统的有效控制。仿真实验证明了该方案的有效性。

An Adaptive Neuron Predictive Control Based on Elman Networks for Nonlinear System

In this paper ,the scheme of an adaptive neuron predictive control based on Elman networks for nonlinear system is presented. Compared to BP network , Elman network is formed by adding internal feedback signal into BP networks .It uses internal state feedback to describe the nonlinear dynamic rules of system ,so that its study speed is improved and it is suitable for real-time identification of dynamic system. A single neuron has simple structure, good adaptability and self-studying ability. According to errors , between target and prediction, and certain algorithm, it can tune parameters of the controller. So the controller is able to be applied to control nonlinear system effectively. The simulation proved the effectiveness of this scheme.