分片线性神经网络逆系统方法的内模控制

Inverse system method for internal model control based on piecewise-linear neural networks

为提高传统逆系统方法的跟踪精度和抗干扰能力,提出了基于连续分片线性神经网络α阶逆系统方法的非线性内模控制方法。利用标准连续分片神经网络逼近非线性系统的α阶逆模型,将它串连在原系统之前,得到复合的伪线性系统,对该伪线性系统应用内模控制策略进行控制,并分析了闭环系统的性能。仿真结果表明:该方法跟踪效果好、抑制干扰能力强,且设计简单,是解决非线性系统控制的一种可行的方法。

An internal model control strategy was developed based on a standard piecewise linear neural network (SCPLNN)αth-order inverse system method to improve the tracking precision and disturbance rejection of traditional inverse system methods.The method combine anαth-order inverse model approximated by a piecewise-linear neural network with the original system to get a composite pseudo-linear algorithm.Then,the internal model control method is introduced into the pseudo-linear system and the performance of a cl...