Observer based half-car model identification of magnetorheological vehicle suspension

The paper presents an identification procedure for parameter estimation of vehicle suspension system, which includes MR (magnetorheological) damper. The vehicle is assumed to be equipped with limited number of sensors, i.e.: accelerometers, suspension deflection sensors and a gyroscope. Full models of vehicle dynamics include vehicle suspension model, wheels and tires models as well as map behavior of other vehicle elements. However, due to the base-excitation approach the model of the vehicle dynamics is limited to the vehicle suspension model, which exhibits 2 DoFs (degrees of freedom). The actual suspension system is known to be nonlinear because of its structure and strongly nonlinear suspension elements such as MR dampers. In the paper the assumption is made, that some kinematic quantities, which are required by the identification algorithm and describe motion of the suspension system, are unmeasurable and need to be estimated. The state observer theory is utilized to support the least-squares method included in the suspension identification procedure. The Kalman filter based method is compared with the straight-forward method, which is based on numerical integration and differencing. Moreover, validation is performed and resistance to measurement noise of both identification algorithms is examined based on simulation.