Intelligent Techniques Based Stable Adaptive Control For Nonlinear Systems

Base on the combination of artificial neural network and fuzzy system theory with the techniques of feedback linearization and sliding mode control theory, the design and analysis approaches for the adaptive control of nonlinear systems are proposed in this dissertation. The thesis consists of the following five chapters.In chapter one, the state of arts of the nonlinear adaptive control theory are stated, and the drawbacks of conventional control theories based nonlinear system control are pointed out. Some recent developments of fuzzy adaptive control and neural network control are also surveyed.In chapter two, some basic knowledge about norms and the definition of stability as well as the general description of nonlinear system control problems are presented.In chapter three, adaptive control of general nonlinear systems based on neural networks is studied. By analyzing the equivalence of fuzzy system and single-layered neural network, stable single-layered neural network adaptive control approach for nonlinear systems is proposed. Firstly, a class of affine nonlinear systems are considered. The single-layered neural networks are used to approximate the nonlinear functions of the systems. Stable adaptive control algorithms are presented based on system states feedback and system output feedback respectively. The boundary of the tracking error is guaranteed. Then, by using the technique of input-output linearization, the stable adaptive control method for general nonlinear systems is proposed. Finally, the effectiveness of the proposed adaptive control algorithms are verified through the simulation researches for the control of robot manipulator.In chapter four, indirect fuzzy sliding mode control for a class of nonlinear systems is discussed. By using the concept of sliding mode control design and Lyapunov synthesis approach, we propose an indirect adaptive fuzzy sliding mode control scheme (IAFSMC) for a class of nonlinear systems. In contrast to the existing sliding mode based fuzzy control system design, where sliding mode control law is directly substituted by a fuzzy controller, in our approach both the equivalent control term and switching-type control term in the sliding mode control law are approximated by fuzzy systems respectively. An on-line adaptive tuning algorithm for the consequent parameters in the fuzzy rules is also designed. The sliding mode based design procedure ensures the stability and robustness of the proposed IAFSMC. We prove that the adaptive scheme can achieve asymptotically stable tracking of a reference input with the guarantee of the bounded system signals. The chattering phenomena in the sliding mode control is also alleviated in our approach. Extensive simulation studies have shown that the presented adaptive design of fuzzy sliding mode controller performs very well without the need for an a priori heuristic knowledge about the controller as well as the system model, and even in the presence of unknown disturbance.In the last chapter, some conclusions are drawn and the future research objectives are given.