Adaptive Output Feedback Control For A Class Of Non-affine Nonlinear Systems

In recent years, the control problem of nonlinear systems always is one of the hottest researches in the automation control fields. In the study of modern control theory, differential geometry and Lyapunov stability theory are the important way to explore such problems. But the considered objects of many studies mostly are simple or common nonlinear systems. We have obtained significant achievements in feedback linearization, adaptive refutations, neural networks, fuzzy logical control and so on. The research into complicated nonlinear system, (such as no-affine nonlinear systems) is few, based on output feedback adaptive control of non-affine nonlinear system problems are less.This dissertation explores a class of implicit control input for nonlinear systems (non-affine nonlinear systems) adaptive control problems. Based on the previous research results, a high-gain observer for this complicated nonlinear systems is constructed, which looks at Lyapunov method as a basic tool, combines with adaptive neural network approximation theory, utilizes differential algebra to design a novel scheme, and completes an output feedback control problem of a class of non-affine nonlinear system. The main results of this study are as follows:1. For a class of general single-input single-output (SISO) non-affine nonlinear systems, which the states of systems are not fully measurable. For this problem, we make use of some mathematical theories, and design a high-gain state observer to estimate the unknown state in systems. We use the neural network structure to approximate the non-affine portion of system, utilize the Lyapunov theory to design an adaptive law, and resolve a class of non-affine nonlinear systems with dynamic output feedback control problem, which renders the tracking error arbitrary small and all signals of the closed-loop systems are bounded through the proposed scheme.2. We present the existing adaptive neural network control theory to extend the general multiple-input multiple-output (MIMO) non-affine nonlinear systems. The systems require for the system relative degree being known, but not doing the dimensions of the system requirements, so that some assumptions of previous are weakened. About the more complex system, we introduce the concept of mapping, reconstruct the system, and change the equation of state of order in standard form. Then, using neural network structure and adaptive algorithm, a control scheme based on adaptive output feedback is designed, which enables the tracking error tends to zero, and all signals in the closed-loop system are bounded.