On Sliding-Mode Stable Adaptive Control For Uncertain Nonlinear Systems Based On Neural Networks

Recently, intelligent control technology for unknown nonlinear systems has gained more and more attention. Many important results have been attained in this field.There are mainly two parts in this thesis: firstly, by using the features and capability of MNNs1 (Multi-layer neural networks) approximating nonlinear functions, based on the principle of sliding mode control and the theory of Lyapunov stability, two new kinds of control schemes are proposed and analyzed. Furthermore, the schemes designed in this part guarantee global stability of the closed-loop systems while the tracking error converges to zero. Simulation studies illustrate the effectiveness and robustness of the schemes. Secondly, according to the results gained in the first part, an adaptive sliding mode controller for a class of large-scale decentralized systems is designed. By theoretical analysis and the simulations, the closed-loop decentralized adaptive control system is proved to be globally stable with tracking error converging to a neighborhood of zero. At the same time, the simulations show the effectiveness and robustness of this scheme.It is necessary to point out that the schemes in this thesis are concentrated on using the MNNs. Further studies are needed if other kinds of neural networks are used to deal with the corresponding control problems.