Connected vehicle data could help local authorities identify emerging road safety risks before serious collisions occur, according to the findings of a pilot carried out thanks to funding from The Road Safety Trust.
The ‘From Prediction to Prevention’ project brought together Leeds City Council, AISIN RoadTrace, Citisense and Metis Consultants to explore how connected vehicle data, AI-powered video analysis and road safety engineering could be combined to support earlier decision-making.
The pilot examined whether connected vehicle data could identify locations where drivers were repeatedly experiencing difficulty. The approach was then independently tested using temporary AI camera technology, with road safety engineers assessing the findings and considering potential interventions.
In an interview with The Road Safety Trust, Wesley Bateson, UK and Ireland lead for AISIN RoadTrace, said the project was prompted by a fundamental question: “Why should we have to wait?” for people to be killed or seriously injured before action is taken at a location.
Historic STATS19 collision data remains an essential part of understanding road safety risk, but connected vehicles generate large amounts of information about how drivers interact with the road network every day.
The pilot explored whether anonymised data could identify locations where drivers were repeatedly braking harshly, encountering unexpected hazards or behaving in ways that indicated elevated risk.
“Today’s vehicles generate an enormous amount of information about how they are being driven and how drivers interact with the road network every single day,” Bateson said.
“If that anonymised connected vehicle data can tell us where drivers are repeatedly braking harshly, encountering unexpected hazards or behaving in ways that indicate elevated risk, shouldn’t we be exploring whether it can help us intervene before a serious collision occurs rather than afterwards?”
Combining different sources of evidence
The Leeds pilot did not seek to replace traditional collision analysis, but to investigate whether predictive data could provide an additional layer of intelligence.
Connected vehicle data was used to identify locations that warranted further investigation, with temporary AI camera technology then used to validate what was happening on the ground. Road safety engineers subsequently assessed the findings and considered potential interventions.
Bateson said the project demonstrated the value of combining different forms of evidence rather than relying on a single dataset.
“Connected vehicle data highlighted potential concern. AI video analysis helped explain what road users were actually doing. Engineering expertise translated those observations into practical recommendations that could be considered by the highway authority,” he said.
The approach identified locations where drivers were consistently experiencing difficulties, including some that had not yet become high-profile collision sites.
According to Bateson, the pilot demonstrated a repeatable methodology for understanding road safety risk and showed how emerging technology can enhance established road safety practice.
Wider highways applications
The project also highlighted potential applications for connected vehicle data beyond road safety.
Bateson said the information could contribute to decisions relating to maintenance programmes, active travel schemes, traffic management, network resilience, asset management and investment priorities.
“Road safety shouldn’t exist in isolation from the wider highways function,” he said.
“Every investment made on the network has the potential to influence safety outcomes, whether that’s a junction improvement, a walking and cycling scheme, resurfacing programme or changes to traffic management.”
He believes greater use of shared data could help break down traditional divisions between road safety and other areas of highway management.
Comment on this story