JULIA investigates AI use in the future of mobility

The JULIA project, a Horizon Europe initiative coordinated by Factual that sees POLIS as part of the partnership, aims to enhance public transport's reliability, safety, and sustainability by integrating cutting-edge AI and EU Space Data services.

In a recent post on the project website, which was inspired by an article drafted on LinkedIn by Joseph Laborda, the Project Coordinator, the partnership looked into the role of AI in innovating mobility, which is at the core of the project itself.

Revolutionising DRT in rural areas

Demand-Responsive Transport (DRT) offers a flexible and cost-effective alternative to traditional public transportation, particularly benefiting sparsely populated regions. As highlighted in the project article, NEMI, a leader in DRT solutions in Spain, employs an AI-driven optimisation tool within the JULIA project to simulate and improve shared mobility services. By using Galileo-enabled devices, JULIA’s AI algorithms process Origin-Destination data to create personalized bus travel solutions. This results in broader route coverage, better accessibility, accurate passenger information, reduced travel times, and improved digitalization of public transport services. Additionally, AI optimises vehicle map-matching, reducing geotagging errors and ensuring precise location of dynamic bus stops.

Galileo services, the European Global Navigation Satellite System, provide crucial accurate planning and real-time route adjustments, minimizing user friction and enhancing the efficiency of shared mobility services.

Automating cycling infrastructure safety assessments with AI

The JULIA project also focuses on enhancing cycling infrastructure safety through AI. Lane Patrol, a pioneering software tool, utilizes the CycleRAP methodology to conduct evidence-based risk assessments from cycling infrastructure images or video footage, as detailed in the original article. Traditionally a manual process, Lane Patrol now employs deep learning models to automate tasks such as image classification, object detection, and semantic segmentation. This automation significantly speeds up and improves the accuracy of CycleRAP evaluations, enhancing the safety assessments of cycling lanes.

Privacy protection is ensured through AWS AI algorithms that blur faces and license plates before image processing. The precise calculation of distances for CycleRAP assessments relies on multi-constellation GNSS receivers, including Galileo, which is especially effective in urban environments.

Read more in the original post on the project website!