Adaptive Control of Mechatronic Vehicle Suspension Systems

The Thesis presents two new adaptive vehicle suspension control methods, which significantly improve the performance of mechatronic suspension systems by adjusting the controller parametrization to the current driving state. Thereby, ride comfort is enhanced while the dynamic wheel load and the suspension deflection remain uncritical. The first concept is an adaptive switching controller structure, which dynamically interpolates between differently tuned linear quadratic regulators. The required estimates of the state vector and the dynamic wheel load are provided by a new estimator concept based on parallel Kalman filters. The stability of the switching controller structure is analyzed employing a common Lyapunov function approach, that takes into account arbitrary fast controller parameter variations and the nonlinear damper characteristic. The performance of the concept is successfully validated in experiments on a quarter-vehicle test rig for a fully active suspension. To overcome the drawbacks of fully active systems, i.e. primarily their high power demand and complex actuators, a new suspension concept called hybrid suspension system is presented. It involves a continuously variable semi-active damper and a low bandwidth actuator integrated in series to the primary spring. The potential of the hybrid concept is shown in an optimization-based analysis. To experimentally validate its performance potential, a hybrid suspension strut is constructed based on stock components from production vehicles and it is integrated in an appropriately designed automotive quarter-car test rig. The second control approach (adaptive reference model based suspension control) emulates the dynamic behavior of a passive suspension system, which is optimally tuned to improve ride comfort for the current driving state while keeping constraints on the dynamic wheel load and the suspension deflection. Its stability is proven by a physically motivated Lyapunov function approach and a switching restriction for the adaptation of the spring stiffness. The adaptive controller structure employs the well-known tuning parameters of passive suspensions, natural frequency and damping ratio of the sprung mass, which fosters transparency and tuneability of the control approach. The new suspension control concept is implemented on the automotive quarter-car test rig for the hybrid suspension employing a new filter-based estimator. Simulations and experiments show that the hybrid suspension system in combination with the adaptive reference model based control achieves performance improvements that are similar to the adaptive switching controller structure for the fully active suspension, however, with a lower power demand.