Multiobjective output feedback control via LMI optimization and quarter-car active suspension system

In this thesis, the linear matrix inequality approach to multiobjective synthesis of linear output-feedback controllers is investigated. Design objectives are a mixture of the H_? performance, H_2 performance, the time-domain constraints, and the constraints on the closed-loop pole locations. As an application of the LMI approach, the quarter-car model active suspension system design with vertical acceleration measurement is studied. Two degree of freedom quarter-car model is defined and parametrized to the white noise velocity road inputs. Multiobjective output feedback controller is designed in order to minimize the rms vertical acceleration for achieving better ride comfort while keeping the tire deflection rms gain and the suspension travel rms response below some given bounds. MATLAB simulations of active and passive suspension systems are compared. The proposed approach achieves best possible ride comfort while keeping the constrained variables within the given bounds.