Machine Learning Based Root Cause Analysis of Monitor Performance Ratio

Vehicle software includes functions specific to the On-Board Diagnostic System (OBD) in order to monitor the performance of manufactured vehicles and detect in-vehicle faults and regulations use In-Use Monitor Performance Ratio (IUMPR) to monitor the performance of these systems. In the American and European markets, it is a requirement during vehicle operation that the IUMPR remains above certain limits. For this reason, engineers perform many driving tests and evaluate the results to observe how the calibrations affect vehicle performance. When the results are below the limits, root cause analysis is performed using software documentations. In this study, IUMPR root cause analysis is performed quickly and automatically by using machine learning techniques. Two different vehicle datasets were analyzed using supervised and unsupervised machine learning methods. The root causes obtained as a result of the analysis performed by engineers confirm the root causes obtained using machine learning.