Distributed Diagnosis for Embedded Systems in Automotive Vehicles

Fault diagnosis is important for automotive vehicles, due to economic reasonssuch as efficient repair and fault prevention, and legislations which mainlydeals with safety and pollution. In embedded systems with dozens of electroniccontrol units that states local diagnoses, it can be arbitrary difficult to findwhich combination of local diagnoses that points at the correct faultycomponents. An algorithm is presented that both finds the diagnoses that inthemselves are complete and chooses only those diagnoses that are more likely tobe correct, this restriction is wanted due to the limitations in processingpower, memory, and network capacity. The embedded system in a Scania heavy dutyvehicle has been used as a case study to find limitations in the embeddedsystem, and realistic requirements on the algorithm.