07 May 2025

The ACUMEN project is developing a governance and regulatory framework, as well as a decision support tool, to guide policy makers in integrating artificial intelligence (AI) and machine learning (ML) into transport management systems. To ensure practical relevance, ACUMEN conducted four workshops with its pilot cities – Luxembourg, Athens, Helsinki and Amsterdam – bringing together national and local traffic management stakeholders.

These workshops not only provided insight into current traffic management systems but also served as a platform for the exchange of best practices. Feedback from participants will be used to refine both the governance framework and the decision support tool to ensure that they are well aligned with real-world applications. The final tool will be made available to stakeholders after the workshops to highlight the benefits of the ACUMEN approach and encourage cities to adopt AI-driven solutions for long-term improvements to their transport networks.

The summary of the ACUMEN pilot project workshop in Luxembourg can be found below:

The Luxembourg pilot of the ACUMEN project focuses on improving door-to-door travel by integrating public transport with on-demand mobility services. Taking place between Campus Contern and Sandweiler-Contern station, as well as along the main pedestrian street in Esch-sur-Alzette, the pilot is aimed at transport service providers such as Sales-Lentz. It aims to demonstrate the ability of the ACUMEN Digital Twin to detect unexpected mobility patterns, such as delays, and to generate stress test scenarios. These tools help to evaluate and optimise the responsiveness and flexibility of on-demand services, ultimately supporting more seamless traffic management.

In order to get the broadest possible picture, stakeholders from HITEC, Luxinnovation, LISER, the E-Bus Competence Centre, Amazon, the University of Luxembourg and government authorities were invited to discuss the integration of Artificial Intelligence (AI) in traffic management. In particular, the challenges, opportunities and framework conditions for the introduction of AI in mobility systems were discussed.

Luxembourg’s traffic management system involves public authorities, private mobility providers and technology providers working to maintain mobility and safety. However, this also poses challenges, such as many traffic management tools being used in isolation with limited coordination across networks. Regulatory frameworks have not kept pace with technological advances, leading to uncertainty about compliance. Existing systems also rely on manual intervention, making it difficult to adapt to congestion and disruption in real time.

The use of AI could bring potential improvements, for example, in predicting traffic flow, dynamically optimising traffic signals, automatically detecting incidents, and planning fleets for on-demand transport. However, concerns remain about the transparency of AI decisions, the harmonisation of regulations, and the risk of over-reliance on automation. AI must complement human oversight, not replace it, to ensure adaptability and accountability.

It is also important to emphasise that the effectiveness of AI-based traffic management depends on real-time and historical data. The workshop highlighted the importance of traffic conditions, mobility patterns, passenger demand trends and the status of multimodal services. The implementation of AI in this context also raises issues of privacy, security and interoperability. Standardised data exchange protocols and reliable ICT infrastructures, such as 5G networks and cloud computing, are needed to support AI decisions.

Another issue is the regulation of AI. Many existing laws were not designed with AI in mind, leading to uncertainty in areas such as intelligent transport systems, fleet management and environmental policy. As AI technologies are evolving faster than policy discussions, greater collaboration between policy makers, AI developers and mobility providers is needed to ensure the relevance of regulations.

This could lead to the following recommendations for the development of a governance and regulatory framework:

• Clarify stakeholder responsibilities in AI-driven transport systems.
• Define the roles of public authorities, private companies, and AI developers.
• Establish legal accountability for decisions made by AI systems.
• Increase the transparency of AI through explainable and interpretable models.
• Strengthen data management practices, focusing on security and interoperability.
• Update the legal framework to align with EU digital transport policies.
• Integrate AI into multimodal transport management to enhance coordination between public transport, ride-sharing services, and private vehicles.