Optimization-Based Unknown Input Observer for Road Profile Estimation with Experimental Validation on a Suspension Station

Knowledge of road profile information can be used to enhance vehicle suspension control and detect road anomalies such as potholes. As such, numerous studies have been proposed to estimate road profile. However, it is very difficult, if not impossible, to validate the estimation effectiveness since there is no true road profile to compare with; the estimation performance of these methodologies are either assessed in simulations or evaluated qualitatively. In this paper, we develop a new optimization-based unknown input observer to estimate road profile and validate it experimentally on a suspension station. As compared to approaches that require both acceleration and suspension deflection measurements, the algorithm only needs the suspension deflection measurement. This is important because the accelerometers are typically installed at the center of gravity of the vehicle and are therefore inaccurate to be used in quarter-car models, especially driving on uneven roads. We evaluate the estimation performance in various cases on a lab suspension station, in which road profile can be explicitly specified and compared against. We perform system identification for the suspension workstation to obtain an accurate model that is used in the optimization-based UIO design. We show promising estimation performance in experimental validations.