Stochastic optimal preview control of an active vehicle suspension

Abstract

Preview control with an estimation scheme is investigated for an active vehicle suspension system with look-ahead sensors. Design of a preview compensator that may be called stochastic optimal, output feedback, preview regulator problem is reduced to the classical linear quadratic Gaussian problem by augmenting dynamics of the original system and previewed road inputs. The resulting solution is a combination of deterministic optimal preview controller and stochastic optimal estimator. The optimal estimator takes the form of a Kalman filter with an additional term of the estimate for the road input, which is given as the weighted preview sensor signal. The Kalman filter gain and the weight used for estimating state and road input, respectively, are designed so that performance degradation by measurement noise is minimized. Numerical examples of a quarter car model are given to verify the performance improvement achievable with the proposed preview control when the estimation from noisy measurement is considered. (C) 1999 Academic Press.