News

More data does not automatically mean safer vehicles. In CoREDaC, Volvo Cars, the University of Gothenburg and Zenseact are working together to develop smarter data collection for advanced driver assistance and automated driving. The aim is to focus resources where they matter most: capturing relevant traffic situations and building stronger evidence for safety. 

The goal of developing new vehicle technology should be clear: preventing crashes, reducing injuries and making the road transport system safer for everyone. Data collection needs to serve that goal. The important question is not simply how much data we can gather, but what we need to learn.

Today, vehicle fleets collect large volumes of image data. Much of it can be repetitive, while important safety-critical situations remain insufficiently represented. This creates costs for transferring, storing and processing data, without necessarily providing the evidence needed to assess safety confidently. It can also contribute to an overly positive picture of how well a system performs. CoREDaC aims to address both challenges. 

From collecting more to collecting what matters
CoREDaC starts with a practical question: 
which traffic situations need to be better represented, and what new data would help fill those gaps?

The project will develop methods that help determine, during collection, whether image data adds useful information or mainly repeats what is already available. It builds on findings from the earlier DeVeLop project, moving the focus from cleaning datasets afterwards to making better decisions at the point of collection. The aim is to reduce unnecessary repetition while preserving coverage of safety-critical situations. 

The researchers will bring these methods together in a prototype and evaluate it using recorded driving data, experiments and industrial validation. They will also identify what is needed to integrate the approach into industrial development. The ambition is to make smarter data collection practically useful, supporting more reliable safety evaluation, lower costs and more sustainable use of resources. 

Connecting with priorities across the SAFER community
CoREDaC addresses questions that have also brought the SAFER community together: what data do we need, where are the gaps, and how can better evidence support safer technology?

At the SAFER Research Day in April 2024 partners explored data use, accessibility, bias and ethics. The program included experiences from planning and collecting datasets at Revere, opportunities with streaming data, and a joint workshop on partners’ data needs. 

Nearly 30 ideas and needs were gathered during the day. These informed the follow-up workshop “Data for traffic safety research” in August 2024, which aimed to turn identified needs into collaborative research proposals. Among the topics were better data on near misses, critical situations and vulnerable road users. 

A complementary perspective was explored later on at the workshop “Deep dive into the future of ADAS – Advanced Testing and Verification”. Here, SAFER partners focused on evaluating safety systems across varied traffic and weather conditions, alongside a workshop to identify shared research priorities. 

CoREDaC connects with these discussions through a common ambition: letting safety needs guide the data we collect and the evidence we build.

Helping the project and the knowledge to reach further
Connecting CoREDaC with the SAFER community creates opportunities for exchange beyond the project partnership. Researchers and practitioners could help challenge assumptions, discuss which situations need greater attention and explore how the methods might benefit other areas of traffic safety research.

The wider dialogue continues at SAFER’s Research Day “AI in traffic safety – from understanding and data to real-world application and impact” on 28 September 2026, where the community will explore meaningful safety benefits, remaining challenges and priorities for future collaboration. 

For SAFER, the opportunity is not only to share results when the research is complete. It is to encourage learning along the way, connect relevant expertise and help promising knowledge move towards practical use.

The shared direction is clear: 

not more data for its own sake, but better knowledge that helps prevent crashes and injuries.

About CoREDaC

• Full title: Cost and Resource Efficient Data Collection
• Partners: Volvo Cars, University of Gothenburg and Zenseact
• Project period: 15 June 2026–15 December 2028
• Funding: Supported by Vinnova through FFI
• Total budget: Approximately SEK 4.3 million
• Contact: Darko Durisic at Volvo Cars