Adaptive control of nonlinear and large scale systems with applications to flexible manipulators and power systems

Systematic design procedures are developed for adaptive state and output feedback control on nonlinear systems with both parametric & dynamic uncertainties and unmeasurable input disturbances. The central element of this approach is the combination of the integrator backstepping result with the Robust Adaptive Nonlinear Damping Matched (RANDM) Lemma developed here. The proposed design offers a major advancement in the area of adaptive nonlinear control where traditionally only parametric uncertainties have been considered.; In the context of large-scale nonlinear systems, decentralized adaptive state and output feedback controllers have been designed. These designs are extensions of the results developed in the centralized case. The distinguishing contribution in this case is the the extension to the class of systems with higher order interconnections which do not satisfy the strict matching restriction. The proposed controllers are global; maintain robustness to a wide class of perturbations in the dynamics; and do not require redesign if additional subsystems are appended to the original system.; Applications of adaptive nonlinear control to vibration damping in flexible manipulators and payloads are considered next. An adaptive design procedure is developed for manipulators with joint flexibility, where the Coriolis, centrifugal, joint flexibility and friction terms in the manipulator dynamics are completely unknown. Next, the problem of vibration damping in flexible-link manipulators is addressed using an adaptive nonlinear control design with input preshaping. Simulation and experimental results on a two link flexible arm at the Control/Robotic Research Laboratory (CRRL) are presented to illustrate the efficacy of this scheme. Finally, an approach utilizing a wrist mounted multi-axis force torque sensor is discussed to address the problem of vibration damping in payloads. Simulation and experimental results on the one-link CRRL manipulator with a flexible payload are presented.; Application of adaptive nonlinear control to guarantee robust connective stability to faults in power systems is considered next. First, an excitation control scheme based on adaptive feedback linearization utilizing full state feedback is presented. Based on the decentralized adaptive design developed in this thesis, decentralized state and output feedback schemes are presented to achieve robust stability and performance for large scale power systems with unknown interconnections.