New South Wales, News, Road upgrades

Artificial intelligence helping fix potholes and rutting in rural roads

NSW councils and the state’s transport agency are turning to artificial intelligence (AI) to help them detect, log – and eventually predict – critical road defect issues.

The $2.9 million NSW Government-backed Asset AI project uses a combination of dash-mounted cameras on vehicles and sensors feeding near-real time road condition updates into the platform.

A risk-based defect priority scoring system then helps maintenance crews ensure they address the most critical defects first, “reducing the overall risk to road users and the community,” said Transport for NSW (TfNSW).

It was first rolled out more than a year ago with project partner Canterbury-Bankstown and Griffith councils with cameras on vehicles including street sweepers and utes.

Shoalhaven City and Warren Shire councils have just joined the project and are now feeding data into the platform and receiving updates through the system.

The project uses a combination of dash-mounted cameras on vehicles and sensors. Image: TfNSW

Other councils ready to be onboarded in 2024 include Liverpool Plains Shire and Mid Coast. Another 48 local councils across NSW have also expressed an interest in joining the project, said TfNSW.

The TfNSW asset inspection team that carries out quality assurance monitoring of state roads across Sydney will also trial three vehicles fitted with dash-mounted cameras.

There are nearly 3000km of state roads around Sydney with 85,000 metres of line markings, 5000 metres of median strips and more than 800 bridges and 18 tunnels.

“It’s great to see our transport teams as early adopters using this new technology and its potential to enhance our current quality inspection capabilities and help shift contracted road maintenance programs towards preventative, and ideally predictive maintenance,” said Matthew Wilson, TfNSW executive director of road maintenance and motorway partnerships.

Leave a Reply

Your email address will not be published. Required fields are marked *

Send this to a friend