AI successfully resolved hourly intensive behaviour task

Vestland fylkeskommune (Vestland County Council) wanted to know whether reducing pavement width at outdoor seating areas would affect accessibility and traffic safety for vulnerable road users. Norconsult solved the assignment using innovative tools based on big data and artificial intelligence.

"The analysis is very thorough and technically challenging. It has documented the proportion of vulnerable road users in Bergen city centre in a new and innovative way that has not previously been possible, and on a scale that we are not familiar with has been done in Norway, " says Mats Korneliussen, Head of Management in Infrastructure and Roads in Vestland County Council.

500 hours of video

In 2023, on behalf of Vestland County Council, Norconsult conducted behavioural analyses for vulnerable road users in Bergen city centre. The analyses were performed using innovative tools for collecting and analysing traffic data from video recordings using artificial intelligence.

A primary objective of the assignment was to assess whether reduced pavement width at outdoor seating areas could compromise accessibility and road safety for vulnerable road users.

The project gathered traffic data from approximately 500 hours of video recordings at 32 different locations in and around Bergen city centre. The tools used image recognition to detect all road users in the recordings and automatically classifying them based on their user group. Movement patterns were mapped and stored. A total of 680,000 road user movements were registered, and the data were used to perform behaviour analyses of vulnerable road users.

The findings revealed instances of congestion and reduced accessibility in connection with outdoor seating areas, with a possible increased risk for pedestrians and cyclists on days with very high pedestrian volumes.

– Engaging and innovative

As part of the delivery to Vestland County Council, a customised web-based transparency tool was developed, allowing them to study and analyse the data collected. Compared to traditional delivery formats, this provides increased insight and enhanced utility value of the data in the county's further work.

"The work carried out by Norconsult gives us a valuable knowledge base to build on to ensure accessibility and traffic safety for pedestrians and cyclists. Norconsult’s project team has shown engagement, innovation and problem-solving skills, delivering the best possible knowledge base to Vestland County Council," says Korneliussen.

New opportunities using AI

The use of artificial intelligence to collect and analyse traffic and mobility data has been a priority for Norconsult. Not long ago, manual observation of pedestrians and cyclists were standard practice. Today, most registrations are done using machine learning and artificial intelligence. It allows for automating most of the process and enabling analyses that were previously impossible.

"This has been a very exciting project to work on. The new technology enabled us to collect a far greater scope of data of vulnerable road users than was previously possible using traditional methods. It enabled us to respond quantitatively to issues that previously were assessed qualitatively through field observations, for example assessments of congestion and potential conflicts between different road users," says Kristoffer Åsen Røys, senior traffic planner at Norconsult.

Significant reduction in time usage

In addition to expanding the range of analysis possibilities, the new tools increase efficiency compared to traditional methods.

“Executing this project would have been challenging without digitization and the use of new technology. The extensive data collection, coupled with complexity of registering the data points using traditional methods, underlines the challenge. Conservative estimates indicate that digitalization and new technology have reduced time usage in this project by at least 90 percent,” says Røys.

  • Thomas Fløien Angeltveit

    Leader of Digital Transformation

  • Kristoffer Åsen Røys

    Head of machine learning and systems development

  • Contact us