ExpectedOutcome:
Project results are expected to contribute to all of the following expected outcomes:
- Concepts, techniques and models based on Artificial Intelligence (AI) used for situational awareness, prediction, decision making and triggering of actions for time critical and safety relevant CCAM applications as well as for cyber threat detection and mitigation.
- A clear understanding of the capabilities, limitations and potential conflicts of AI based systems for CCAM.
- Increased user acceptance from an early stage, based on explainable, trustworthy and human-centric AI. Interactions with vehicles using AI should be understandable, human-like and reflect human psychological capabilities, and free of gender, ethnic or other biases.
- Accelerated AI development and training for CCAM enabled by a relevant set of real and synthetic traffic events and scenarios.
- AI based CCAM solutions will evolve from reactive and/or adaptive system support into predictive system state awareness (including driver state and user diversity), decision-making and actuation, enhancing road safety especially in near-critical situations.
Scope:
The deterministic understanding and consequential design of assistance systems are mostly reactive or to some extent adaptive. In the transition from driver assistance systems towards fully automated systems, a critical aspect is the decision making (i.e. planning and acting), based on robust and reliable detection and perception. AI has a huge potential to advance this process.
Specifically, in more complex and dense traffic environments, highly automated driving functions will benefit from the system state prediction enabled by AI. Yet, the current state of technology using AI for CCAM has limitations regarding human-like actions, more specifically the intuitive, split-second (predictive) assessments and ‘reflex decision making’. As such, any AI requires good integration into the overall system with close interaction and compatibility with the active safety systems (e.g. automated emergency braking).
For the development process, training is essential for the performance of unbiased AI. It requires sufficient traffic and event data under varying conditions from all over Europe, avoiding limited data sets. The current, mainly deterministic approaches for validation in automotive development will not be sufficient for future training and validation of AI-based or AI-supported functions, which will also need to be able to deal with complex issues as (un)intended miscommunication.
Proposed R&I actions therefore are expected to address all the following aspects
- Support the development and integration of AI in CCAM with explainable, trustworthy and human-centric and unbiased concepts, techniques and models; this can be on vehicle level and on transport system level, where tactical and strategic links to traffic management and traffic conditions need to be established.
- Address the knowledge gap on AI training and validation approaches as well as efficient and ethical approaches for data handling of increasing amounts of data.
- Build upon existing and generated data for training and verification of AI supporting situational awareness in CCAM in more complex traffic scenarios (e.g. digital twins).
Specific automotive requirements on functional safety and security need to be considered in the development process of an automotive-grade AI ensuring consistency with existing validation procedures.
This topic requires the effective contribution of SSH disciplines and the involvement of SSH experts, institutions, as well as the inclusion of relevant SSH expertise, in order to produce meaningful and significant effects enhancing the societal impact of the related research activities.
In order to achieve the expected outcomes, international cooperation is advised, in particular with projects or partners from the US, Japan, Canada, South Korea, Singapore, Australia.
This topic implements the co-programmed European Partnership on ‘Connected, Cooperative and Automated Mobility’ (CCAM).
Specific Topic Conditions:
Activities are expected to achieve TRL 5 by the end of the project – see General Annex B.
Cross-cutting Priorities:
Digital AgendaCo-programmed European PartnershipsInternational CooperationSocial sciences and humanitiesArtificial Intelligence
