On 2 October 2025, the AIthena Project showcased its three-year journey with a Final Event in Brussels, marking the culmination of its work in advancing Connected, Cooperative, and Automated Mobility (CCAM) through Artificial Intelligence (AI).
Hosted at Comet Louise, the event brought together researchers, industry professionals, policymakers and leaders to showcase AIthena’s achievements, explore the latest innovations and discuss the role of AI in CCAM.
Opening remarks and keynote
The event began with a welcome address by Dr. Oihana Otaegui, the project coordinator from Vicomtech. She reflected on the consortium’s collaborative efforts and the importance of trust in AI for mobility systems.
This was followed by a keynote address from Suzanna Kraak (DG RTD) representing the European Commission. The speech emphasised the Commission’s ambition to restore the EU’s leadership in autonomous systems and digital innovation. The strategic value of projects such as AIthena was emphasised, as was their role in advancing Europe’s digital transition and sustainable mobility goals: “We are making significant progress in making AI trustworthy”.
Technical sessions and use cases
A series of technical presentations showcased AIthena’s research and its real-world applications.
Methodology overview
Daniel Franco (Rupprecht Consult) gave a detailed presentation on the project’s multi-layered, human-centric methodology for ensuring the trustworthiness of AI in CCAM environments. His presentation focused on the key principles of the AIthena methodology, such as fairness, transparency, and accountability. AIthena’s methodology ensures the development of trustworthy, ethical, and human-centric AI for automated mobility, operationalising fairness, transparency, accountability, and privacy through practical checklists, guidelines, and tools. The methodology supports continuous improvement, stakeholder guidance, and compliance with regulations such as the EU AI Act and GDPR, thereby bridging the gap between technical development and societal expectations.
Use Case 1: Trustworthy Perception Systems for CCAM
Till Beemelmanns (ika – RWTH Aachen University) demonstrated advancements in perception systems that can reliably detect and interpret environmental data. He highlighted a growing trend in AI development, noting that although components have become larger and more complex, they have also become less transparent and more difficult to interpret. The AIthena project’s Use Case 1 has developed a trustworthy perception system for connected and cooperative automated mobility, with a focus on reliable pedestrian detection in urban environments. The system uses multi-modal sensor fusion (radar, camera and LiDAR), explainable AI layers and visualisation interfaces to enhance robustness, transparency and user understanding. The approach emphasises documentation, regulatory compliance and open-source benchmarking in order to support safer automated driving and to build public trust in AI-driven mobility solutions.
Use Case 2.1: Collision prediction with Hybrid AI data fusion models
In a joint presentation, Jos den Ouden and Svetlana Orlova from Eindhoven University of Technology (TU/e) and Rubén Naranjo de las Heras from Vicomtech presented novel data fusion models designed to improve collision prediction in complex traffic scenarios. The presentation also emphasised the importance of adopting a more data-driven approach. Use Case 2.1 aimed to enhance the prediction of collision risk in automated driving by combining multiple AI models and data sources, particularly in complex or edge-case scenarios where traditional methods are ineffective. The approach uses ontology-based scene understanding and video foundation models to interpret road scenes, assess risk and detect anomalies. This results in improved data quality, explainability and reliability in rare or challenging situations. This hybrid approach enables automated vehicles to make safer, more transparent, and more robust decisions in real-world
traffic.
Use Case 2.2: Robust prediction modules for Robo-taxis in urban environments
In his presentation, Bernhard Hillbrand (Virtual Vehicle) explored how AI could ensure the safety and reliability of self-driving city taxis. Use Case 2.2 has focused on developing robust AI modules that can predict the movement of road users, enabling robo-taxis to operate safely and reliably in complex urban scenarios. This approach leverages real-world sensor data, automated annotation and deep learning to enhance situational awareness and ensure high accuracy and adaptability, even in challenging conditions. This approach facilitates the integration of scalable systems, ensures regulatory compliance and enhances the safety of automated mobility systems
Use Case 3: Trustworthy and human-understandable decision-making
Guido Linden (ika – RWTH Aachen University) addressed the crucial challenge of interpretability in AI, emphasising the importance of ensuring that AI-driven decisions are transparent and consistent with human reasoning. One promising approach to achieving this is to combine prediction and planning. Use Case 3 has focused on developing explainable and robust decision-making systems for automated vehicles, so users can understand and trust AI-based decisions. The approach combines deep learning-based behaviour planning with rule-based tactical planning and visualises decisions for occupants. It also measures system competence in real time to trigger safer strategies when needed. This approach promotes regulatory compliance and public trust by ensuring transparency, providing documentation and enabling deployment in real-world and simulated environments
Use Case 4: AI-based traffic management
Anton Wijbenga (MAPtm) presented a methodology using a simulation framework that road operators and transport authorities can apply to assess the impact of introducing automated vehicles on the network in terms of efficiency, sustainability, and safety. Evaluating the results of different simulation parameters and scenarios in relation to their effects on network KPIs improves the transparency of how the AI in automated vehicles influences the traffic network. Use Case 4 analysed and optimised the impact of automated vehicles (AVs) on traffic networks by applying AI-based traffic management strategies that prioritise efficiency, safety and sustainability. Microsimulation tools were used to model network-level effects, evaluate AV behaviours under different market penetration scenarios, and track key performance indicators such as capacity, safety, and emissions. This approach provides policymakers with actionable insights and highlights the importance of cooperative AV behaviours and trust in traffic management systems for maximising benefits.
Throughout the morning and at lunchtime, attendees visited the exhibition rooms, where use cases were demonstrated.
This was followed by a panel discussion featuring Silvia Barbaro, Hamza Guirrou, Ted Zotos (International Road Transport Union, IRU) and Nathalie Poissonnier (Federation of Belgian Bus and Coach Operators, FBAA). The discussion focused on regulatory frameworks, deployment challenges and the strategic potential of integrating trustworthy AI into mobility systems.
Nathalie Poissonnier pointed out that some opportunities are still “under the radar”, particularly with regard to reducing drivers’ workload: “More development is needed in planning, particularly driver planning, which takes up a lot of their time. But we also need to help predict the best routes to take,” she added, stressing that AI can play an important role in reducing traffic jams and improving efficiency.
Hamza Guirrou stressed the importance of building trust in the role of AI in mobility, stating: “We need to work on accepting that AI can support the driver and make his life easier without taking his place.”
Closing remarks
Pedro Alfonso Pérez Losa, representing the European Climate, Infrastructure and Environment Executive Agency (CINEA), delivered a summary and congratulated the AIthena project on its strong alignment with EU strategic priorities. ‘The basis is to introduce safety and security in autonomous mobility, making trustworthy AI a key element in CCAM’. He highlighted significant progress, including extensive collaboration with other countries that are actively investing in this area.
The event concluded with a speech by Dr. Oihana Otaegui, who asked the audience for feedback and celebrated the community built through AIthena’s work. This was followed by a networking lunch, which provided an opportunity for participants to connect with each other, discuss potential future collaborations, and share insights.
Watch the recap video of the AITHENA final event below (also available on the FIA Region I Vimeo channel: https://vimeo.com/1127437856)