Vehicle-Independent Interpretation of Sensor SignalsWithout a-priori Knowledge of Their Semantics

The vehicle-independent analysis of fused sensor data is not only the basis for modern Advanced Driver Assistance Systems (ADAS), the detection of critical events in road traffic, but also for fault diagnosis in complex sensor systems. Within the LEICAR project, we follow the axis of the semantic interpretation in order to analyze fused vehicle data in such a way that they can be classified and then used in a semantically higher level for event recognition. In this paper, we show a method for automatically interpret the semantics of the recorded raw sensor data, transmitted on the vehicle bus system. The sensors messages can be understood as bit patterns and the data from all sensors can be understood as time series with high resolution. They are interpreted in a multi-staged process in which both, manual and automatic classification and correlation of the sensor data is performed to determine the semantics of the signals.