Design of Practical Observer for Semiactive Suspensions

Abstract

This paper focuses on the development of practical observers for semi-active suspensions. Many sensors are necessary to measure all the states used for a control system. Therefore, the estimation of states is valid for developing simple and low cost semi-active suspension systems. The nonlinearity, which is included in the absorber of semiactive suspensions, is a difficult problem in estimating accurate states using a usual linear observer theory such as the Kalman filter theory. To solve this problem, this paper presents a new practical observer which has two improvements. The first point is that the observer model includes a data map which reveals the nonlinear characteristic of the absorber. The second is that the observer feedback gain is scheduled. We design the feedback gain L₀ for a linear quarter car model whose damping ratio is minimum, and L₁, for that whose damping ratio is maximum. The feedback gain is determined using linear interpolation of L₀ and L₁ according to the control signal of the absorber valve. Applying this observer for actual systems, however, the problem that this observer needs more quantity of computation than the usual linear observer must be considered. Therefore, we reveal a method which reduces the quantity of the computation. To confirm the validity of this practical observer, simulations and experiments are carried out. These results show that the observer developed in this paper can estimate the suspension stroke velocity accurately using the vertical acceleration sensor on the unsprung mass.