Adaptive control of multivariable systems with actuator nonlinearities

This dissertation is to address several important problems in multivariable adaptive nonlinear control, namely, adaptive backstepping design for multivariable systems, parameter estimation for coupled error models, multivariable adaptive compensation for multivariable actuator nonlinearities with known or partially known linear dynamics, state feedback control and output feedback control of multivariable systems with nonsmooth nonlinearities at the input of linear dynamics, and optimal backlash compensation for a multi-input tracking system. Related issues such as plant modeling, controller structure, error model and stability analysis are investigated in this dissertation.; Motivated by the multivariable nature of many practical systems and the common existence of actuator nonlinearities in such systems, we focus our research on the adaptive control of multivariable systems with unknown actuator nonlinearities.; Our main objective is to develop adaptive control designs for multivariable systems with actuator nonlinearities using an adaptive inverse approach to cancel the unknown actuator nonlinearity effects. Our results in achieving this objective include (1) a new parameter estimation algorithm for coupled error models; (2) an open-loop adaptive actuator compensation design; (3) two state feedback control schemes for linear dynamics known; (4) an output feedback control design for linear dynamics known; (5) an adaptive feedback control design for multivariable systems with actuator nonlinearities and unknown linear dynamics but known high frequency gain matrix; (6) adaptive actuator nonlinearity compensation of multivariable systems with only partially known high frequency gain matrix; (7) an adaptive feedback control design for multivariable systems with actuator nonlinearities and only partially known high frequency gain matrix.; As an extension of our research effort on multivariale adaptive control and a preparation for adaptive inverse control of nonlinear systems with unknown actuator nonlinearities, we also present an adaptive backstepping design for linear multivariable plants.; To further develop multivariable adaptive compensation techniques for practical multivariable systems with unknown actuator nonlinearities, we study a multi-input tracking system with actuator backlash. As our results, an optimal compensation scheme and its numerical solution are presented.; Extensive simulation results are also presented in this dissertation to illustrate the significant system performance improvement gaining from the adaptive inverse compensation.