Sydney, July 13: The Australian state of New South Wales (NSW) has announced a new AI (artificial intelligence) technology that is set to automate and revolutionise the way the state’s roads are maintained and repaired.
The project announced on Tuesday would fund a 2.9-million-Australian dollar ($1.96-million) trial from AI company, Asset AI, which would install sensors on 32 public buses with routes across greater Sydney.
The sensors use AI to combine visual data with local weather conditions to predict the rate of deterioration in the city’s roads — meaning it would in theory be able to alert maintenance teams before potholes or other road damages pose a risk to traffic.
“There will always be cracks in the road and there will always be potholes but with smart tech like this we can predict deterioration, streamline maintenance and get to better outcomes much faster,” said NSW Minister for Customer Service and Digital Government, Victor Dominello.
At present, road damages and defects rely on reports from residents. However, in months of heavy rainfall, as has been experienced in Sydney, roads often deteriorate faster than can be adequately monitored.
The technology would allow pothole “heatmaps” to be generated across the city, indicating which areas are in the most need of repair.
Khal Asfour, Mayor of Canterbury Bankstown, an area in southwest Sydney which underwent a pre-trial of the technology, said the AI technology has already helped improve road safety in the area.
We do an audit of our roads once every four years and it is very expensive. This new technology will allow us to do it on a weekly basis instead,” Asfour added.
“Asset AI uses predictive analysis to improve road maintenance by predicting the risk to the community rather than just reporting the condition of the road assets, and that’s great news for our residents.”
Beyond the initial trials in Sydney, the project would also look to gather road data across 100 km of rural roads in the state — areas which are particularly costly and difficult to maintain.