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

Traditional diagnostic methods activate when malfunctions become apparent, often signalled through physical symptoms or fault codes, making the process reactive rather than preventive. While effective to a degree, these methods are challenged by the complexity and interconnectivity of modern vehicle systems, leading to time-consuming and expensive diagnostics.

Pre-diagnostic strategies we propose aim to identify potential issues before they manifest into significant problems, employing subtle indications of component degradation that might not trigger standard fault codes. Recent progress in AI and ML technologies is the key enabler for such radical development. By analysing vast amounts of vehicle data, AI-driven systems can identify patterns indicative of impending failures, streamlining the diagnostic process, reducing repair times, and cutting costs.

This pre-emptive strategy is particularly critical for vital components such as engines, where early detection of issues can prevent severe breakdowns and the consequential logistical and financial ramifications. We will leverage “ranking data”, a rich source of information typically used for fault code generation but underutilised for predictive diagnostics. The use of advanced ML algorithms, including recurrent neural networks and transformers, promises to enhance diagnostic accuracy and efficiency, extending the lifespan of components and the vehicles themselves by pre-emptively identifying potential issues.