Optimal control of semi-active automobile suspension including preview information

In this dissertation synthesis of controllers for semi-active vehicle suspension systems is presented using an approach based on optimal control theory. For the purpose of this investigation, a vehicle is modeled as a system consisting of rigid bodies representing the body structure and the wheels with linear springs, and semi-active dampers that can vary the rate of energy dissipation across the suspension. Since the damping ratios of the semi-active dampers are control variables, the system is described by bilinear state equations involving products of the state vector and the control inputs. Using optimal control theory, the suspension system is optimized with respect to the performance index that trades off a measure of ride quality against an index of road holding while preserving constraints on suspension rattle space and on the amount of damping available in the control system. Particular emphasis is placed on the optimal use of preview information about the road/terrain input in front of the vehicle. Such information can be obtained either from sonar sensors located in front of vehicle or by monitoring the motion of the front wheels and using this information to control rear suspension. Additional aspects of design of suspension controls, such as addition of integral action to eliminate suspension deflection offset on sloping road, improved attitude control by reducing roll, dive or squat in transient maneuvers and body lean due to aerodynamic forces, are also considered. The theory developed here is presented for both continuous-time and discrete-time formulations to facilitate digital implementation. The results of numerical simulations for 2-degree-of-freedom and 4-degree-of-freedom vehicle models are presented showing potential for substantial improvement of performance over existing designs. Potential benefits brought about by having springs with variable stiffness (as is the case with some air spring designs), are also evaluated.