Algorithm Development for Minor Damage Identification in Vehicle Bodies Using Adaptive Sensor Data Processing

Abstract

Abstract The paper presents an approach for fully-automatic detection and recognition of minor vehicle body damages in a scenario of frequently changing drivers, such as in the car rental or car sharing businesses. It utilizes a sensor network integrated into the vehicle body. The algorithmically contribution is the multi sensor-data fusion of the signals from these sensors and the subsequent reasoning framework stage. The present research assignment is the inclusion of additional information from the network of the vehicle's electronic control units. This additional information about the vehicle state (e.g. closed or opened door status, speed, changing fuel level) is used to verify the calculated damage events. The efficiency of a structural damage detection directly depends on the acoustical vehicle noise level. Additional structure-borne sound sources also introduce oscillations into the vehicle body (e.g. road and engine noise, opening and closing doors). Some of this information uses a central electronic unit for adaptive sensor data processing of the damage detection system. The higher confidence of correctly identified damages increases the credibility of the system.