Key points are not available for this paper at this time.
Software will lead the development and life cycle of vehicles in the future. Nowadays, more and more software is being integrated into a vehicle, evolving it into a Software-Defined Vehicle (SDV). Automotive High Performance Computers (HPCs) serve as enablers by providing more computing infrastructure which can be flexibly used inside a vehicle. However, this leads to a complex vehicle system that needs to function today and in the future. Detecting and rectifying failures as quickly as possible is essential, but existing diagnostic approaches based on Diagnostic Trouble Codes (DTCs) are not designed for such complex systems and lack of flexibility. DTCs are predefined during vehicle development and changes to vehicle diagnostics require a large amount of modification work. Moreover, diagnostics are not intended to handle dynamically changing software systems and have shortcomings when applied to in-vehicle software systems. In the Cloud, there are already established approaches to observe and diagnose software systems. However, these approaches are too comprehensive and cannot simply be applied to the whole vehicle. Anyway, they are a helpful addition to adapting vehicle diagnostics. Therefore, their vehicle applicability needs to be investigated. In this paper, we discuss the challenges of transferring and adapting the DTC approach to in-vehicle software systems, as well as monitoring and observability approaches to vehicles. Based on this, we introduce a concept for future vehicle diagnostics that addresses existing diagnostic approaches based on DTCs in combination with established approaches for monitoring and observability. Our presented concept provides a basis for further future work in the context of vehicle diagnostics for SDVs.
Bickelhaupt et al. (Tue,) studied this question.