Investigations and Comparison of Noise Signals to Useful Signals for the Detection of Dents in Vehicle Bodies by Sound Emission Analysis

Abstract

The context of this paper is the German research project KESS (Konfigurierbares Elektronisches Schadensidentifikationssystem) funded by the Federal Ministry of Education and Research. In this Project, an entire system from sensing structure-borne sound up to Car-2-Infrastructure communication for the minor damage detection of car sharing vehicles is under development. Therefore, the suppression of noise signals is essential for the further algorithms to classify minor damages. Minor damages in vehicle body parts create structure-borne sound which overlaps with additional oscillations of the vehicle. In the ordinary usage of vehicles, those additional oscillations are mainly caused by road and engine noise. Within this paper the differences of such noise signals and signals of structure damages will be described. Therefore, an analysis method to calculate spectrally decomposed envelope signals of structure-borne sound signals is given and the analysis results will be presented. A digital filter method to attenuate the majority of noise signals and to preserve the useful signals for the detection of minor damages will be given at the end of this paper. This digital filter method has been successfully designed and implemented into the algorithm of the sensor by using model based design techniques.