Universal Output Feedback Control Of Uncertain Nonlinear Systems

Over the past decades,tremendous progress has been achieved to control a class of nonlinear lower triangular systems with uncertainties via state feedback through a systematic design way.When only a part of the state or the output of the system is measurable,how to design the controller to globally asymptotically stabilize nonlinear systems becomes more realistic.In the case of linear systems,the separation principle allows output feedback problems to be solved by combining state feedback controller with state observers.However,the separation principle does not hold for nonlinear systems^[46].In many practical problems,the assumptions that all states of the nonlinear systems can be measured and can be used in feedback design are often unrealistic, such that it is difficult to apply state feedback control law here.Sometimes,even if the states of the nonlinear systems can be measured directly,but taking into account the costs of the implementation and other factors,the output feedback control approach are adopted more,if it can achieve the stabilization of the closed-loop system.This dissertation studies a class of nonlinear systems which can be transformed into the lower triangular systems,with the aid of the differential geometric approach^[3].The intuitive idea behind this adaptive system is the tuner increases the controller gain as long as the output of the system is not zero,then eventually,the controller gain becomes sufficiently large and the system is stabilized.Performs has partially been determined offiine.Partially in the sense that it has been decided that the controller gain will be monotonically nondecreasing.The data only determine how fast the controller gain will indeed increase.In such a case,the idea of universal control to design adaptive control for the systems with a nonidentifier based tuner will address the problem which is hard to identify these unknown parameters,as the unknown parameters occur nonlinearly.In our thesis contents are organized as follows:In chapter 1,introduce the background of this topic,internal and overseas research situations, theoretical and Practical significance,the main contributions of this dissertation.In chapter 2,introduce thesis involved basic theory,including basic concepts,the main lemma and inequality important.Universal control,universal adaptive regulation control,Barbalat's lemma,Young's inequality,and so on.In chapter 3,for a class of large-scale uncertain nonlinear systems interconnected by the uncertain states.Each sub-system is a lower triangular system under linear growth conditions, with unknown parameters and unknown rate.We show that under linear growth conditions,there is a decentralized universal output feedback controller rendering the closed-loop system globally stabilization.Use global decentralized control approach,the introduction of dynamic gain which will be driven by the estimation error.In chapter 4,for a class of nonholonomic uncertain systems in chained form,with uncertainties was dominated by a smooth function multiplied by an unknown constant.Using of the state-scaling^[78][79] technique,we transform it into a lower triangular system.Using the idea from universal control,in order to deal with the unknown constant and smooth function,dynamic gains are introduced.Then the universal output feedback domination design is applied to design a reduced-order adaptive observer and an universal output feedback controller,such that all the state of non system can be regulated to zero.In chapter 5,we summarize the dissertation and discuss open problems for future research.