Robust Tracking Control for Nonlinear Systems
|Course||Operational Research and Cybernetics|
|Keywords||Nonlinearly parameterized systems Adaptive robust control Output tracking control Asymptotica stabilty Asymptotica regulation|
In this thesis, we mainly investigate the robust tracking control for nonlinear system models by choosing a proper Lyapunov function and designing a perfect controller recur-sively:the adaptive robust H∞control of nonlinear triangular systems, the robust control of nonlinearly nonholonomic systems and the output tracking control of nonlinear feed-forward systems. The uncertainties for the three models stem from unknown continuous parameters, time-variable perturbation and unexpected nonlinearies.Firstly, we introduce the problem about the robust H∞disturbance attenuation with internal stability for nonlinearly parameterized systems with uncontrollable linearization. By adding a power integrator and using the parameter separation technique, an explicit robust dynamic feedback law that solves the problem of global stabilization is designed. The systematic model is also developed and the results are more general.Next, we consider the problem of robust adaptive control for a class of nonholo-nomic systems in chained form. Combining the parameter separation technique with the feedback domination design, a solution to the problem of global-adaptive control for the uncertain nonholonomic systems is obtained and a switching strategy is developed to elim-inate the terms of uncontrollability. The presented adaptive control algorithm guarantees that all the states converge to the origin and other variables are bounded.Finally, we study the problem about the practical tracking control of nonlinear feed-forward systems using the homogeneous systems theory and present the homogeneous domination approach to solve the tracking problem. The nonlinear systems in this paper are high-order and admit the high-order growth with respect to the unmeasurable states. The proposed controller that guarantees the properties such as robustness to disturbances does not depend on the reference signals and makes the error enough small.