Ranking Data for AI-driven Precision Diagnostics in Heavy-Duty Trucks (PRIDE)

Purpose and goal

Traditional methods of diagnostics are only utilised when faults become apparent, usually through physical symptoms or fault codes, making the process reactive rather than preventive. Although effective to a certain extent, these traditional methods are challenged by the complexity and interconnectedness of today´s modern vehicle systems. This results in a diagnostic process that is both time-consuming and costly. The pre-diagnostic strategies we propose in the PRIDE project aim to automatically identify potential problems before they develop into more serious faults.

Expected effects and result

State-of-the-art AI systems can identify complex patterns that indicate upcoming failures, streamlining the diagnostic process, reducing repair times and lowering costs. Such a preventive approach is particularly critical and important for the most vital components, including the engine, where early detection of failures can prevent serious engine breakdowns and thereby also reduce the subsequent logistical and economic consequences. Our solution will complement the established workshop routines and support faulty vehicle inspections to find the cause of the problem.

Planned approach and implementation

Recent advances in AI and Machine Learning technologies enable such radical developments by providing tools for analysing the information stored in large amounts of vehicle data, and exploiting subtle signs of component degradation that may not necessarily result in fault codes. We will utilise "ranking data", a rich source of information that is commonly used to generate fault codes but underused in predictive diagnostics. The PRIDE project is coordinated by CAISR (Centre for Applied Intelligent Systems Research) at Halmstad University, with Volvo GTT as the key partner.