An Approach for Combined Vertical Vehicle Model and Road Profile Identification from Heterogeneous Fleet Data

In this work a novel approach for concurrent vehicle model parameter and road profile identification is proposed which exploits the availability of fleet data. By combining measurement data, obtained from low-cost smartphone sensors, for multiple vehicles, typical identifiability issues present in single-vehicle measurements can be circumvented. Moreover, as the presented approach exploits a low order model where the parameters are identified for a specific asset (i.e. a digital twin), the shared road profile can be identified, rather than just the resulting forces. A computational framework is presented and a first experimental validation is performed where a speed-bump is identified using data from three different vehicles. This approach has the potential to improve the availability of accurate data for a.o. customer correlation durability design, road condition monitoring, and active suspension systems.