一类非线性不确定系统的最小方差神经控制

针对一类非线性不确定系统，提出一种新的基于神经网络动态补偿的最小方差控制方法（ＭＶＮＮＣ）。这种控制系统将传统的最小方差控制技术与神经网络优良的非线性逼近能力相结合，从而能有效地消除不确定性引起的控制误差，仿真实验表明，这种最小方差神经控制系统具有较强的鲁棒性和良好的动态性能。

Minimum Variance Neural Control for a Class of Nonlinear System with Uncertainties

A new neural network based minimum variance control technique (MVNNC) was proposed for a special class of nonlinear uncertain systems. This control scheme integrates classical linear minimum variance control technique and the excellent learning ability of neural network. It can eliminate system control error caused by nonlinear uncertainties. Simulation results show that this control scheme has a strong robustness with respect to system uncertainties and a good dynamic performance.