Comparative study of three robust observers for automotive damper force estimation

Abstract

This paper aims at comparing three robust observers used to estimate the damping force of electrorheological (ER) dampers in a vehicle suspension system. Firstly, a nonlinear quarter-car model, augmented with a first-order dynamical nonlinear damper model, is developed. The first two methods are designed considering the nonlinearity as an unknown input and minimizing the effect of the unknown input disturbances (including a nonlinearity term, the measurement noise and the unknown road profile) on the estimation errors, by using an H2 and H∞ criterion, respectively. The latter method aims at minimizing only the effects of measurement noises and road profiles on the state variable estimation errors by using a H∞ criterion, while the nonlinearity is bounded through a Lipschitz condition. For implementation issue, two low-cost sensors signals (two accelerometers data from the sprung mass and the unsprung mass) are considered as inputs for the observer designs. Then, the observers are implemented in real-time on the INOVE test bench from GIPSA-lab (1/5-scaled real vehicle) to assess and compare experimentally the performances of the approaches. Both simulations and experimental results demonstrate a better effectiveness of the latter observer in terms of the ability of estimating the damper force in real-time despite the nonlinearity, the measurement noises and the road disturbances.