Nonlinear observer/controller design and its application to friction compensation

The control design for output-feedback nonlinear systems where only system output is measured attracts a great deal of attention in the field of control because of their ubiquity in realistic applications. On the other hand, the presence of unknown parameters is inevitable in practical control design problems because of uncertainties in system modeling and varying application environment. Therefore, it is of importance to identify practically interesting nonlinear systems in the absence of full-state measurement and complete knowledge of model parameters.; In this dissertation, we first develop control design methodology for global adaptive output-feedback stabilization and tracking of a new class of nonlinear systems that can be transformed via a global diffeomorphism into systems that are linear up to output injection and linear with respect to the unknown parameters. As a special case of the extended class of nonlinear systems, we further consider nonlinear systems in which multiplicative nonlinearities of the system output are absent but affine functions of the derivative of the measured output with coefficients that are smooth nonlinear functions of the measured output appear.; Then we apply adaptive nonlinear observer and controller design techniques to friction compensation. We present two adaptive nonlinear friction compensation schemes that can be utilized to handle two types of parametric uncertainties in the LuGre dynamic friction model: non-uniform friction force variations and normal force variations. Several velocity tracking control examples are introduced to explain basic design methodologies. Our design features a dual-observer structure for estimating different nonlinear effects of the unmeasurable friction state in the LuGre friction model. Incorporating the friction parameter estimates and the dual observers for the unmeasurable friction state, adaptive nonlinear controllers are proposed to achieve globally asymptotic tracking of the given velocity reference signal.; Furthermore, we extend our friction compensation scheme for non-uniform friction force variations to the small signal model of induction motor drives. An adaptive friction compensation scheme is proposed not only to compensate existing dynamic friction in the induction motor drives but also to handle mechanical uncertainties such as inertia variation and load disturbance.