Diagnosis of Automotive Systems Based on Logfile Similarity
Mohammed Nabil Kabbaj, Andrei Doncescu · 2011
In this paper, a methodology of failure diagnosis in automotive systems based on log file similarity is presented. It supposes that we have a set of classes representing different failures of a vehicle, each class is composed of similar log files. Then, we define a similarity metric based on Kolmogorov complexity that measures the amount of shared information between the log file of a broken vehicle and the prototype of the different referenced failure classes.