Centralized/local optimal output feedback control and robustness with application to vehicle active suspension

In many control applications, the controller structure can be pre-specified based on the feedback of centralized/local measurements. The optimization of centralized/local output feedback control, along with a numerical solution, is presented in this dissertation.; An industrial application of low-order centralized/local control, requiring no state estimation, is developed for vehicle active suspension. Manufactured vehicle parameters and a seven-degree-of-freedom full-car rigid body model are used to study the performance of this control. The vehicle model is simplified by eliminating lightly coupled modes. A rigid/flexible full-car model including an arbitrary number of vibratory and flexible modes is then developed.; A fourth-order active suspension control law is developed. This control law is then extended to the variable damper based semi-active suspension control. It is shown that only minor differences exist between the active and semi-active control for the vehicle ride control considered in this work.; A robustness and performance study of the centralized/local active suspension control is performed. The study is conducted using the matrix singular values approach and the multivariable sensitivity and complementary sensitivity functions. Both rigid and flexible full-car models are studied. It is shown that the fourth-order centralized/local active suspension control has satisfactory robustness characteristics. The matrix singular values approach proves to be a convenient and useful measure of performance and robustness.