A Study Of Uncertain Nonlinear Control Systems

The main results of this research report include:studying the global stabilization problem for a class of nonlinear systems with upper triangular structure. For a subclass of these systems, we prove the feasibility of global smooth stabilization. For a general uncertain system with upper triangular structure, we propose a systematic design procedure for global continuous (but not smooth) stabilizer, studying universally the control problem for systems containing input nonlinearities, such as deadzone, hysteresis, quantization (with unknown width) , etc. By modifying the tracking controller proposed by E. P. Ryan, we achieve asymptotic tracking. We also systematically design λ -tracking controllers, for a class of uncertain nonlinear systems with down triangular structure and actuator nonlinearities. studying nonlinear adaptive control problem. By using diffeomorphism containing estimated parameters, we propose a new nonlinear adaptive control strategy for non-parameter-pure-feedback systems. We also make some improvement for the nonlinear adaptive strategy proposed by I. Kanellakopoulous, et al for parameter-pure-feedback systems, and the refined one can resist bounded matched disturbance, study nonlinear robust control problem. We give a new robust control strategy weakening the "equivalently matched" condition proposed by Z. Qu. The regulated state is guaranteed to enter and remain in a given arbitrarily small neighborhood of origin within a finite time interval, proposing two new tracking control strategies for robot. The first one extends the method of feedback linearization proposed by T. J. Tarn, et al for robot to uncertain case. By exploiting the important properties of robot dynamics, the second strategy removes the'requirement of a priori bound on parametrer uncertainty and weakens the restriction on external disturbance.