Adaptive Neural Network Tracking Control For Constrained Nonlinear Interconnected Systems

In practical engineering,many physical systems can be modeled as nonlinear interconnected systems,such as electromechanical system,coupled inverted pendulums,mass-spring-damper system and helicopter system.As nonlinear characteristic of the controlled systems,saturation,quantization and dead zone constantly appear in lots of industrial processes.The presence of nonlinearities may severely deteriorate system performance and lead to instability.Therefore,the nonlinear interconnected systems with nonlinear input/output also need extensive attention and in-depth research.On the other side,to implement certain quality of performance indices,the states,outputs and errors of the controlled systems are often constrained to some extent,and the scope of their constraints are varying with time.Therefore,it is necessary to design an effective control scheme which not only achieves the predefined stability performance of the controlled system,but also ensures the constraints are not violated.So far,although many effective control strategies have been presented for nonlinear interconnected systems,there are still many control problems to be solved.Based on the development status at home and abroad,combined with the decentralized adaptive control method and backstepping technology,the main researches and works of the constrained nonlinear interconnected system in this paper are as follows:For time-varying output-constrained nonlinear large-scale systems preceded by input saturation,Chapter 2 proposes a reliable adaptive finite-time decentralized control scheme.In the backstepping control design process,time-varying barrier Lyapunov functions,effective controllers,and adaptive laws are constructed.By fusing with the finite-time stability criterion,it is proved that the proposed control scheme not only ensures that the system achieves stability in finite time,but also ensures that the output constraints are not violated.Finally,the effectiveness of presented control scheme is validated by the simulation results.Chapter 3 focuses on the problem of decentralized adaptive neural network finite-time control for pure-feedback nonlinear interconnected systems with input quantization and time-varying state constraints.In order to satisfy state constraints,a time-varying barrier Lyapunov function is introduced in each step of the controller design.At the same time,combined with adaptive backstepping control technology,a decentralized adaptive neural network finite-time controller is constructed.Through Lyapunov stability analysis,all closed-loop signals are semi-global practically finite-time stable and the tracking error converges to a small neighborhood near the origin in finite time.Finally,the feasibility of the method is verified by simulation examples.An effective neural-based decentralized adaptive control strategy is proposed for nonlinear interconnected systems with unknown dead zone outputs in Chapter 4.By introducing the Nussbaum function,the control design obstacles caused by the output nonlinearity are solved.In order to further achieve tracking error constraints,the prescribed performance technique is introduced.According to Lyapunov stability theory,it is verified that the proposed control scheme can ensure that all closed-loop signals are bounded,and the tracking errors remain within a small compact set with the prescribed performance bounds.In addition,the simulation results are given and the correctness of the designed control method is demonstrated.