Observer design based on nonlinear suspension model with unscented kalman filter

This paper presents a new approach to estimating suspension state information and parameter in real-time. An observer with unscented Kalman filter is designed based on a nonlinear quarter car model. The proposed observer could estimate the sprung mass, vertical velocity of sprung and unsprung mass for the nonlinear suspension systems with vehicle load variation. The designed observer has low sensitivity and robust to unknown road surfaces. The efficiency of the estimator is validated through the simulations with two different types of road excitation and payload variations. The simulation results clearly indicate that compared with the extended Kalman filter estimator, the unscented Kalman filter is more accurate and robust. The estimated state information and parameters could be used in the design of suspension control systems.