Velocity Estimation Algorithms for Suspensions

Semi-active suspension systems aim to improve stability and comfort of vehicles. The most well-known semi-active suspension control strategies assume knowledge of the vehicle vertical-velocities. However, the vehicle normally does not includes sensors for these variables. In this paper various vertical-velocities estimation algorithms are compared: the traditional filtered integration algorithm, the classical Luenberger observer, the Unknown Input Observer and the Kalman Filter. A comprehensive comparison between the estimation methods in several perturbation scenarios is performed. The results show that it is possible to archive good estimation performance with a low level of complexity using a combination of estimators.