基于改进Elman神经网络的非线性预测控制

Nonlinear Model Predictive Control Based on Improved Elman Neural Network

为了提高非线性预测控制中预测模型的精度,提出一种基于递归神经网络建模的预测控制方案。采用改进Elman神经网络在线建立预测模型,用递推最小二乘法在线修改神经网络权值,并引入误差补偿环节,从而达到改善预测模型精度的目的,使控制系统的控制性能得到提高。仿真实验表明了该方法的有效性。

A predictive control method is proposed based on the neural network due to improving the accuracy of the nonlinear model predictive model .The improved neural network- Elman is used to build the predictive model of the process online.The method used to correct the NN weight value is the recursive least square.The tache of error compensation is introduced to the neural network,it meets the accuracy of the predictive model and the control performance of the control systems is improved.The results of simulation show the effectiveness of the control algorithm.