An Approach for Autonomous Damage Detection by Means of Modal Fingerprints

Abstract

This article discusses whether elastic and plastic impacts and scratches caused by applied forces (e.g., car bump, key scratch etc.) to a parked vehicle can be distinguished by an autonomous system. The authors assume that a force impact to a vehicle stimulates characteristic frequencies in the vehicle body shell components that can be used to distinguish the given damages. Eight piezoelectric film sensors made of polyvinylidene fluoride mounted at the inside of the vehicle body shell components doors, bumpers, bonnet, and tailgate are used as the sensing element. An experimental modal analysis of several vehicle components is performed to identify characteristic frequencies. The results are used to compute the best sensor position in order to detect all considered eigenfrequencies of the component. A constructed model set up for force generation was used to record reference data. Derived features of the reference signals are computed and assessed to their relevance of distinguishing the damages on the vehicle.