TORRES Project Overview

The TORRES project (Traffic prOcessing foR uRban EnvironmentS) is an applied research initiative funded by Innoviris, running from 2023 to 2025. Its purpose has been to design and demonstrate advanced tools for understanding, monitoring, and analysing urban traffic using state-of-the-art Artificial Intelligence (AI) methods. The transport network of the Brussels region served as the main real-world pilot for data collection, modelling, and demonstration of tools.

The research team at ULB has been led by Prof. Gianluca Bontempi from the Machine Learning Group and included the researchers Eladio Montero Porras, Alejandro Morales Hernández, Ali Enes Dingil, and Davide Andrea Guastella (later as an external collaborator from Aix-Marseille University). The partnership also involved VUB’s Electronics and Informatics Department and Macq Mobility, combining academic and practical expertise in AI and transport.

Methodology and Technical Approach

TORRES developed a suite of methods and tools targeted at capturing, integrating, and analysing heterogeneous traffic data. The project combined real traffic information from existing monitoring infrastructures and anonymised IoT data with synthetic inputs created through data augmentation. These rich datasets have been utilized to develop AI-based models that interpolate missing information, forecast traffic dynamics, and facilitate high-level analysis across the transportation system. Additionally, interactive dashboards and analytical frameworks were created to visualise traffic conditions and evolution at the city scale. A key focus has been on enabling “what-if” scenario analysis. This includes evaluating potential outcomes of changes on the ground, such as adjustments to speed limits, closure of key road segments, or rerouting of traffic flows. These analytical capabilities provide authorities with a means to quantify likely effects before implementing policy or infrastructure changes.

Impact on Urban Mobility and Decision-Making

Urban mobility challenges such as congestion, air pollution, and road safety have direct social and economic costs. By delivering tools that integrate diverse data sources and leverage AI for prediction and simulation, TORRES supports evidence-based decision making. The ability to model different scenarios helps city planners and transport authorities to assess potential interventions, communicate expected benefits to the public, and develop more resilient and adaptive transport strategies.

Scientific Contribution and Dissemination

Beyond the concrete deliverables, TORRES aligns with MLG’s ongoing research in digital twins and intelligent transport systems, demonstrating methodological leadership and practical applications. The project outcomes have been disseminated through scientific events, publications, and collaborations with FARI and PARADIGM for TULIPE’s project, including integration into the CAVE demonstration, which has raised the profile of the team within the academic and urban innovation communities.

Learn more here