Advanced Control Strategy And Applications To Uncertain Nonlinear Systems With Constraints

It is well-known that there exist nonlinearity and uncertain parameters and functions in most of nonlinear physical systems.Actually,there are mainly two reasons caused the rise of nonlinearity and uncertainty: one is the nonlinearity,which cannot be ignored during the control design;the other is the added nonlinear components to improve the closed-loop system performance,or simplify the structure of control systems.Ignoring those nonlinearities and uncertainties may degrade the system performance,instability,or damage the whole system.For this reason,this thesis focuses on the research on the control of nonlinear systems in the presence of high nonlinearity and unknown system dynamics.Due to the complexity and nonlinearity of nonlinear systems,it results in remarkable difficulties in the control design and performance analysis.Although the research of nonlinear system control is difficult,many effective approaches have been proposed: adaptive control,robust control,neural network control and so on.This thesis is mainly to solve the analysis and control of nonlinear systems with the constraint and unknown control direction,and relative applications.Moreover,the stability of the whole closed-loop system can be proven by Lyapunov direct method.Under the support of Lyapunov direct method,adaptive neural network control for nonlinear systems is proposed to stabilize the whole system.Besides,the designed method is applied to a practical control problem of robotic systems.Both state feedback and output feedback controllers are proposed,and the effectiveness of the proposed control is verified by simulation examples.Firstly,barrier Lyapunov functions are used for the control of nonlinear systems with constraints.Since the value of barrier Lyapunov function will go to infinity,when the state converges to the boundary.Adaptive updating laws are also designed to ensure that the state of nonlinear systems remains in the constrained set.Then,aiming at nonlinearity: the unknown control direction,the Nussbaum function is used to analyze and design the control method for compensating the effect of unknown control direction.It is proven theoretically that all signals in the closed-loop system are uniformly bounded.Next,the proposed adaptive neural network control framework is applied to a practical nonlinear system control case: the tracking control problem of roboticmanipulators.The mathematical dynamics of robotic manipulators is analyzed in this thesis.Further,both state feedback and output feedback control methods are proposed to stabilize the robotic manipulators and address the tracking problem.To sum up,the main contribution of this thesis is to design adaptive neural network control of uncertain nonlinear systems with constraints and unknown control direction.The control parameters can update in real time to achieve good control performance.Furthermore,this control method is applied to specific application examples to verify the effectiveness of the proposed control method.