两种基于Narendra方案的混合自适应修正方案

Two Kinds of Modified Hybrid Adaptive Schemes Based on Narendra's Scheme

本文在Narendra等人所提出的适用于理想对象的混合自适应控制方案的基础上提出了两种混合自适应方案,本文首先将Narendra等人提出的仅适合于理想对象的控制器方案利用规范化方法推广到对象具有未建模动态情况,提出了一种修正方案,然后又针对修正方案在某些情况下收敛速度缓慢这一缺点进行分析,发现是Narendra方案的固有缺点,于是又提出了一种变换误差模型方案,采用两种不同的敏感函数和误差测度模型交替使用的方法去估计控制参数,既利用了Narendra方案的优点,也克服了Narendra方案的固有缺点,使系统的跟踪性能有很大改进。

In this paper we proposed two kinds of modified hybrid adaptive control schemes based on the scheme for ideal plant proposed by Narendra and others. First, using normalization we extend Narendra's scheme only adapted to ideal plant to the plant which contains unmodeled dynamics, and propose a modified scheme. Then, we analyse the cause of low convergence rate of modified scheme, and discover that it is an inherent shortcoming of Narendra's scheme. So we propose an Interchanging error model scheme, in which two sorts of sensitive, functions and error measure models are interplaced to estimate the control parameters. This scheme not only has the advantages of Narendra's, but also overcomes the inherent shortcoming of Narendra's scheme, and it evidently improves the tracking performance of the system.