Road Profile Inversion from In-Vehicle Accelerometers

Abstract

This study was motivated by the need for a road profile monitoring concept that can provide frequent measurements over large areas across all weather conditions. A profile inversion method was proposed in this context, based on measured in-vehicle accelerations alongside speed and location information. The method combined a quarter-car response model, a proportional-integral-derivative controller, and a nonlinear error minimization algorithm. In general terms, the underlying idea for profile inversion was centred around best-matching measured accelerations with accelerations calculated in a precalibrated quarter-car response model. The new method was first verified with synthetic/manufactured acceleration traces. Next, it was applied to real data obtained from 10 nominally identical electric cars that were driven over highways and urban roads. Inverted road profiles were compared to those measured by a standard laser profilometer. Very good match and reproducibility metrics were obtained, especially for highways, indicating that the new method is potentially useful as a large-scale road profile monitoring concept.