Transforming Mobility with AI, IoT, and Digital Twins
The METROPOLIS project, launched in 2021 and developed by a consortium of polytechnic universities (UPC, UPM, UPV, UPCT) and Spanish companies (ALSA, AMB Informació), has reached its final phase, consolidating advancements in urban and metropolitan mobility. Throughout its implementation, it has integrated new technologies and strategies to enhance the efficiency, sustainability, and safety of transportation in the cities of the future.
In the Barcelona case study, METROPOLIS has developed a Mobility Digital Twin based on urban traffic simulation software, calibrated with data from IoT sensors, traffic cameras, and loading/unloading parking records. This tool has facilitated the analysis of innovative strategies for urban freight distribution, including the implementation of urban consolidation centers (UCCs), demand grouping, the use of smart lockers, and nighttime or off-peak distribution.
The results have shown that consolidation in UCCs and demand grouping improve operational efficiency and last-mile logistics profitability, significantly reducing vehicle travel distance and environmental impact. Additionally, the model has allowed for the evaluation of various cooperative transport scenarios, showing that the use of smart lockers can reduce the average delivery cost in cities by 69.3% and that nighttime distribution increases the average delivery speed by 12.1%, while also reducing CO₂ emissions by 6%.
Combining these strategies with the use of electric and autonomous vehicles has demonstrated great potential for transforming urban freight mobility into a more sustainable and efficient model. To complement this approach, IoT sensors and artificial vision systems with neural networks such as YOLOv8 were deployed to monitor the occupancy of loading and unloading zones. Models like CNN and ViT were used to recognize human activities, achieving accuracies of 94.90% and 95.77%, respectively.
As part of the project, a case study was also carried out at the Moncloa interchange and the A-6 corridor, one of the main access routes to Madrid from the northwest of the region. To improve passenger transport efficiency in this area, IoT sensors, simulations, and artificial intelligence models were implemented to analyze passenger and vehicle flow in real time. A cooperative management model for the corridor and interchange was developed, integrating data from pedestrian counting and traffic from urban buses, intercity buses, and private vehicles.
Edge Computing has enabled real-time processing of data captured by OAK-1 PoE cameras, successfully detecting passenger movements with accuracies between 91% and 98%, even in complex scenarios with up to 122 people per minute. Additionally, autonomous transport models have been developed using platforms based on Jetson Nano and virtualization software such as XtratuM, ensuring safe operations in urban environments. The designed autonomous vehicle prototype not only performs autonomous navigation and traffic sign recognition but also carries out additional operations such as onboard passenger counting. The project has also driven the creation of an Intelligent Mobility Data Center, consolidating key information for strategic decision-making.
METROPOLIS represents a milestone in applied research for smart mobility, laying the groundwork for a more efficient and sustainable transportation model. Its solutions can be adopted by public administrations and private operators, contributing to the development of safer cities adapted to future challenges.
This publication is part of the R&D project PLEC2021-007609, funded by MCIN/AEI/10.13039/501100011033 and the European Union NextGenerationEU/PRTR.
