| This dissertation addresses the problem of steering control of vehicles in two ways: it proposes a new steering strategy to improve vehicle handling, and designs steering controllers for automated handling of vehicles.; The first part of the dissertation, Non-parallel Steering, proposes and analyzes a new, non-parallel steering technique which can improve the handling characteristics of road vehicles. Non-parallel steering exploits the lateral load transfer that occurs in road vehicles travelling on curves. By giving different steering inputs to the two front wheels, the effective cornering stiffness of the front tires can be changed, causing the vehicle to become more maneuverable or directionally stable, as desired. Simulation results are presented, as well as a study on the issues that will arise in actual implementation.; The second part, Automated Steering, develops three steering controllers, as part of the PATH (Partners for Advanced Transit and Highways) project underway at the University of California, Berkeley. A sliding mode controller is designed to steer the vehicle in the center of a pre-defined lane. The controller is guaranteed to provide bounded tracking errors in the presence of bounded variations in road-tire interactions and vehicle parameters. Parameter adaptation is then used in the controller to guarantee tracking error convergence to zero. Finally, a non-linear sliding surface is introduced into the controller to check against tire/actuator saturation, and to improve ride comfort. Simulation and experimental results are presented to verify the controllers. |
