Nonlinear Control Applications for Nonholonomic Autonomous Systems Nonlinear Control Applications for Nonholonomic Autonomous Systems Nonlinear Control Applications for Nonholonomic Autonomous Systems

NONLINEAR CONTROL APPLICATIONS FOR NONHOLONOMIC AUTONOMOUS SYSTEMS by Lucas C. McNinch Villanova University, 2009 THESIS ADVISORS: Kenneth R. Muske, Ph.D. and Hashem Ashrafiuon, Ph.D. This thesis presents several nonlinear control applications for nonholonomic autonomous wheeled mobile robot and unmanned surface vessel (USV) systems. First, an overview of nonlinear control and nonholonomic systems is given and sliding mode and model predictive control methods are introduced. Next, a kinematic controller is applied to an experimental wheeled mobile robot for validation of rapid prototyping techniques and new trajectory planning algorithms. Then, sliding mode controllers are developed for setpoint and trajectory tracking of USVs. Sliding mode control provides a robust control law with modest computational requirements that can be implemented on small-scale model USV systems. However, tuning the control parameters can be very non-intuitive and one set of parameters is rarely suitable for different initial conditions. A receding horizon model predictive controller is then developed for setpoint and tracking control. This controller yields open-loop optimal performance regardless of initial conditions, constraints or disturbances but the high computational demands make it challenging to apply this method to small-scale USV systems with fast dynamics. Finally, a predictive sliding mode cascade controller is developed combining the strengths of the sliding mode and model predictive control methods. Simulation and/or experimental results are presented for each application.