Research On Dynamic Surface Control Of Two Types Of Constrained Nonlinear Systems

State/output constraints are commonly encountered in real systems,and the violation of constraints often result in system performance deterioration and system instability.Hence,the research of constrained nonlinear systems has received extensive attention.This paper discusses the dynamic surface control of two classes of constrained nonlinear systems.Main results are:1.Adaptive dynamic surface control for pure-feedback nonlinear systems with output constraint and unknown control directionTo deal with a class of uncertain pure-feedback nonlinear systems with output constraint and unknown control direction,an adaptive dynamic surface control(DSC)scheme is provided in this part.The effects caused by output constraint and unknown control direction are eliminated by employing barrier Lyapunov function(BLF)and Nussbaum function,respectively.By introducing first-order filters into control design,a novel statefeedback tracking controller is designed to overcome the repeated differentiations on virtual controllers in backstepping methods.It is shown that output constraint is not violated,all the signals in the closed-loop system are semi-globally bounded,and the tracking error converges to a small region around zero.2.Adaptive fuzzy dynamic surface control for full-state constrained highorder nonlinear systems without feasibility conditionsFeasibility conditions are sufficient conditions for the existence of BLF-based stateconstrained control.In this part,the DSC problem of state-constrained high-order nonlinear systems is investigated without involving feasibility conditions.Fuzzy logic systems are employed to approximate unknown system nonlinearities.By integrating nonlinear transformed functions and first-order filters into control design,full-state constraints can be handled,and the repeated differentiations on virtual controllers can be circumvented in control design.It is rigorously proved that the states of fuzzy systems remain within a specific compact set,on which fuzzy approximation is valid.Besides,asymmetric full-state constraints are not violated,all the signals in the closed-loop system are semi-globally bounded,and the tracking error converges to a small region around zero.