Generative AI for smarter CCAM: enhancing perception, decision-making, and validation (CCAM Partnership)

What is this call about?

The topic funds projects that apply generative AI to perception, decision making and scenario-based validation for connected and automated vehicles, with a focus on safety for vulnerable road users.

What does this call fund?

Scope

Pilots and demonstrations of Level 3 and 4 vehicle services face challenges in perception and decision making, especially in complex urban environments. They need low-latency solutions and limits on latency, bandwidth and energy use, as well as better security, privacy and reliability. The topic seeks ultra-safe, trustworthy and energy-efficient real-time perception and decision-making for automated vehicles, using scalable solutions and generative AI (GenAI), and contributes to the AI Continent Action Plan. It implements the co-programmed European Partnership on Connected, Cooperative and Automated Mobility (CCAM).

Expected actions

Proposed actions are expected to address all of the following:

• Tools and approaches for robust environment perception and decision making at the edge, on board, in the infrastructure or in a back office, to improve reasoning, efficiency, cybersecurity and reliability, with path planning as the initial use case. This supports perception of vulnerable road users and prediction of their behaviour, and includes data sharing to widen the time window for action in near-accident situations. Large language, vision language and vision language action models may be used.
• Generation of scenarios for interactions between automated vehicles and other road users, extending existing datasets with variations such as cultural differences and infrastructure variability.
• Integration of GenAI into existing development, training and validation approaches, with guidelines on its limits and benefits, including gender bias and fairness, and an outlook on uptake across CCAM components and systemic uses such as traffic management and remote control.
• Encouraged: collaboration with the European Software-defined Vehicle initiative by adopting existing interfaces and building blocks and proposing new ones.

Expected outcomes

Advanced, trustworthy and energy-efficient perception systems using GenAI; better safety for vulnerable road users; more robust on-board and infrastructure systems through training and virtual testing in GenAI-generated scenarios; a clearer understanding of the relevance and limits of GenAI for CCAM; and tools and harmonised approaches for using GenAI in mobility technology development. Proposals should include measures for close coordination with the European Connected and Autonomous Vehicle Alliance, and projects are expected to report to the CCAM Partnership, liaise with the AI, Data and Robotics Association partnership, and apply the European Common Evaluation Methodology for CCAM. Activities are expected to reach technology readiness level 5-6 by the end of the project.

Further expectations

• Alignment with national projects and initiatives is encouraged.

When is the deadline?

8 October 2026