基于神经网络的反馈式自适应逆控制系统  

Feedback Adaptive Inverse Control Systems Based on Neural Networks

作　　者：周辉[1] 董正宏[1] 朱仁峰[1] 

机构地区：[1]装备指挥技术学院电子工程系

出　　处：《控制工程》2006年第3期244-246,249,共4页Control Engineering of China

摘　　要：由于在建立非线性逆模型时采用带有复杂非线性函数的滤波器来完成,而由此带来了结构复杂、运算量大等缺点,在实际运用上受到了很多限制。为此,采用较为简单的线性逆控制方式与单层神经网络相结合构成逆控制结构,其良好的非线性特性使系统具有逼近任何非线性模型的能力,且结构简单实用。应用举例表明这种方法在运算量和控制性能上均取得了非常好的效果。The general method of building a nonlinear inverse model is adopting a filter that consists of many complex functions. In this ease, the complexity of construction and algorithm limits its application. Aiming at that, a method consisting of linear inverse model and neural networks is presented. Compared with traditional method, it is excellent at algorithms and construction. Simulation results show that it requires less calculation and makes better performance.

关 键 词：自适应逆控制 非线性 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]