Software-Defined Vehicles (SDV) are fostered through initiatives like SOAFEE and Eclipse SDV promoting the use of cloud-native approaches, distributed workloads and service-oriented architectures (SOA). Vehicles are connected to the cloud and functions are partially executed in the cloud and the vehicle. However, these approaches do not provide solutions for monitoring and diagnosing SDVs neither on-board nor off-board.
Given the cost-sensitive nature of vehicles, OEMs are unlikely to allocate significant resources to diagnostics. Consequently, conventional data centre monitoring strategies, which rely on transferring large datasets to dedicated servers, are not directly applicable to vehicles.
The paper describes a SOA used in research for several years where each vehicle functionality is implemented by independent services with an orchestrator for managing them. ASAM SOVD (ISO 17978) specifically supports diagnosing SDVs by providing functionality beyond UDS, such as access to log data. Though perfectly prepared for SDV diagnostics, functionality such as validating service quality, chain-of-effects, or dynamic resource usage are not properly covered, yet. As services will be distributed between the cloud and the vehicle in the future, the diagnostic functionality for identifying the root cause of a problem must not only consider the vehicle but also services running in the cloud.
By transferring established solutions for monitoring and diagnostics to vehicles and extending the SOVD standard, the paper proposes a solution that fills current gaps: on-board monitoring of services including their chain-of-effects, fault generation for erroneous conditions, analysis of historical data, etc. With the core diagnostic logic running on-board, functionality is even supported in regions with intermittent connectivity whilst still meeting the requirements for cost-sensitive vehicles.
With commercial vehicles and construction machinery moving towards Ethernet-based communication with HighSpeed InterConnect – and thus SOVD – the approach fits also there. The paper will also give a brief outlook on how SOVD affects production.