Robust gain scheduling control with applications in smart materials and biomedical robotics

This dissertation addresses the problems of compensating for the hysteresis nonlinearity present in Shape Memory Alloy (SMA) wire actuators and control of a biomedical robotic system.;In the first part of the dissertation, we examine sonic important properties of the hysteresis nonlinearity and formulate the control methods implemented on the designed experimental systems. The first system is actuated with an SMA wire actuator. Due to the temperature dependent hysteretic nonlinearity, it is a challenging problem to model and control the response of the SMA system precisely. We formulate an Adaptive Neuro-Fuzzy Inference System (ANFIS) model and controller, robust Hinfinity and LPV gain scheduling controllers for this purpose.;In the second part, we experiment ally verify the designed modeling and control methods and compensate for the hysteretic nonlinearity in real-time. The robustness of the controllers plays an important role in hysteresis compensation, especially for systems actuated with SMA wires. We compare the performances of the controllers under high levels of plant disturbance (heat fluctuations).;The third part is dedicated to the modeling, prototyping and control of a novel biomedical robotic system. This system is designed to assist the surgeon during the delicate minimally invasive direct apical approach to aortic valve replacement operations. The designed robot has the capability of exerting haptic force feedback on the medical tool that the surgeon uses during the operation.;For a realistic real-time implementation, we use a dynamic virtual model of the left ventricle of the heart. The workspace of the robot is defined as the boundaries of the left ventricle. For haptic controller, we model these boundaries as virtual walls and provide the operator the interaction forces acting between the end effector of the robot and these unilateral constraints.;Finally all LPV gain scheduling controller is implemented on the haptic system for precise trajectory tracking applications. The designed controller is then tested on the experimental system in real-time for a pre-defined trajectory on the workspace of the robot.