非线性系统模糊神经网络变结构自适应控制

在非线性系统的模糊动力学模型基础上,提出一种模糊神经网络变结构自适应控制器;网络的结构根据非线性系统特性动态构成,基于该网络提出非线性预测器,基于梯度法提出了一种网络参数学习算法,并分析了收敛性及其性质。将网络预测器与参数学习算法相结合,构成自适应控制算法,证明了算法的收敛性。仿真结果证实了算法的有效性。

Variable Structure Fuzzy Neural Network Adaptive Controlof Nonlinear Systems

In terms of a fuzzy dynamic model of nonlinear systems. the paper proposes a nonlinear adaptive control algorithm using fuzzy neural network. The network structure is automatically generated while the system is running. A nonlinear predictor is constructed by fuzzy neural network. A learning algorithm of the network parameters based on gradient method is introduced. Its convergence and properties are proved. Let-ting the prediction output equal to the desired output, the predictor obtains the control signal. By combining the control signal production method with the parameter learning algorithm of the network predictor, the adaptive control algorithm is developed. Its convergence is proved. The simulation results illustrate the effec-tiveness of the proposed algorithm. The algorithm can deal with the nonlinear system control without math model.