The vehicle information processing architectures are generally evolving towards domain-centric and zone/vehicle-centric archetypes, redefining how systems and software functions should be built and integrated. On the software side, enabling over-the-air updates – both from a regulatory point of view and as a means of meeting consumer expectations – present specific challenges in terms of version management and security. More generally, complexity is increasing exponentially due to the introduction of more complex and distributed functionality (e.g. autonomous driving and infotainment) and the need to integrate multiple functions within and across on-board and off-board domains. As a result, the cost of producing, deploying and maintaining software for the low-volume automotive market (compared to, say, smartphones) is becoming increasingly difficult for OEMs. For Tier-1 suppliers, the cost of adapting standardised modules (if they have any) to the target architectures of OEMs and their individual car lines is also becoming unmanageable. Furthermore, the increase in software complexity is far outstripping the increase in software productivity.
HAL4SDV will focus on methods, technologies and processes to be used in series development of vehicles ready for market in 2030 onwards, extrapolating the expected dramatic developments in the constitutive areas of microelectronics, communication technologies, software engineering and AI for the period after 2030. The following assumptions about abstract requirements as basis for this period are: 1) Absolute data-centricity and complete code portability through appropriate hardware abstraction ensure maximum functional flexibility, reusability, and minimum time-to-market; 2) Ensure efficient fusion of decentralized, heterogeneous data streams for efficient high-performance processing using specialized, reliable hardware architectures; 3) Inside-out continuum, i.e., there is no longer a distinction between inside and outside the vehicle, all resources in the vehicle are accessible from the outside, and the vehicle has full access to all resources on the outside; 4) Unlimited horizontal scalability (between moving road users/vehicles/infrastructure) and vertical scalability (e.g. vehicle sizes); 5)Real-time capability is guaranteed, adequate computational models and suitable hardware/software or E/E architectures are used; 6) Vehicles are fully “liable”, and they are defensible against cyber-attacks.
With this overall and long-term vision in mind, HAL4SDV will focus on the short-term objectives of specifying the hardware abstraction layer, its standardisation, the development of appropriate tool chains and methodological aspects to achieve or regain technological leadership in core areas. Needless to say, the latter may also have an impact on subsequent projects (dedicated to the layers above the hardware abstraction) as well as on parallel development work in Europe, e.g. on chip design and fabrication, mobility services and software ecosystems in general.