A new EU-funded research project is testing whether artificial intelligence can warn road operators about dangerous situations before they develop into crashes.
PROSAFE-TWIN has received nearly €5 million from Horizon Europe and brings together 17 organisations from nine European countries and the United States. Coordinated by the Mobility and Logistics Cluster of Euskadi (MLC), the project began with a consortium meeting in Bilbao on 20 and 21 October and will run for 42 months until March 2030.
Predicting risk in real time
The project will create hybrid digital twins of real road networks. These virtual models will combine traffic-flow physics with machine learning based on data from roadside sensors, connected vehicles and weather systems.
The intended result is a live risk map that can identify where hazards such as concealed queues or conflicts involving pedestrians and cyclists are likely to emerge. Predictions will include confidence levels, allowing traffic operators to distinguish credible warnings from statistical noise and reduce the risk of alert fatigue.
“PROSAFE-TWIN is about moving road safety from reaction to prevention, giving road authorities the intelligence to identify risks and intervene before critical situations become crashes,” said Sofía López de Aguileta of MLC, which coordinates the project and leads its Bizkaia pilot.
She said the project combined specialist expertise with real-world testing environments.
Four pilot locations
Trials will take place in four different settings. In Bizkaia, Spain, the project will cover about 600km of rural and interurban roads, building on the Basque Country Corridor living lab and its connected-vehicle infrastructure.
Rotterdam in the Netherlands will provide a test environment of urban roads with speed limits of 30–50km/h, used by cyclists, light electric vehicles, lorries and cars. The city has operated an AI-based traffic safety prediction model since 2019, which PROSAFE-TWIN aims to develop further.
In Portugal, the trial will cover a 4.2km section of the A4 Trasmontana motorway, including the 2.8km Corgo Viaduct. It will examine high-risk locations and near-misses in fog, rain and congestion, as well as ways to help operators deal with incidents more quickly.
The fourth pilot will run along the Celje–Velenje–Slovenj Gradec corridor in Slovenia. With relatively limited sensor coverage and the Third Development Axis under construction, it will test whether predictive safety tools can operate effectively when data is scarce.
Human oversight and connectivity
PROSAFE-TWIN is intended to support, rather than replace, the decisions of traffic operators. Social scientists involved in the project will work with control-room staff on how warnings should be displayed and assess whether AI-supported decisions treat different road users fairly, particularly pedestrians and cyclists.
LMT Group will develop and demonstrate low-latency 5G Standalone connectivity and edge-computing systems, enabling traffic data to be processed closer to where it is generated.
In Portugal, LMT will provide a 5G SA network backbone for the motorway pilot to support real-time trajectory analysis, with latency below two seconds. In Slovenia, it will work with project partners to integrate live data from road-weather stations reliably into the regional corridor pilot.
The work could also expand LMT’s traffic-monitoring platform beyond detecting violations, towards identifying hazardous situations and issuing safety warnings.

