具有降阶模型的 MRAC 系统的鲁棒控制器设计

在自适应控制中被控对象的模型往往与实际对象有一定的误差,表现为对象的未建模动态特性和外来干扰.而适应控制的设计者只能以理想模型为基础来设计自适应控制器.然而有些自适应算法不能保证问题对外来干扰或未建模特性的稳健性,甚至很小的干扰也会使得自适应环不稳定.自80年代初,C.E.Rohrs 等提出这种具有未建模

A NEW ROBUST CONTROLLER OF MRAC SYSTEMS WITH REDUCED MODELS

This paper proposes a now direct adaptive algorithm truncate modification,which is de-signed on the basis of the reduced order plant,but guarantees the robustness of the trackingerror between the outputs of the reference model and the with unmodelled dynamics.It is shownthat,even for some high frequency reference inputs,the algorithm guarantees the boundednessof all signals in the adaptive loop and the expected precision of the tracking error,if the un-modelled dynamics are sufficiently small.In the absence of unmodelled dynamics,the residu-al tracking error is shown to be zero.