By Kody Cook, Journalist, Infrastructure Magazine
With maintenance costs at record highs across the country, the use of AI is increasing the efficiency of local road upkeep, by predicting where and when works will be needed ahead of time. Here, we look at how the City of Canterbury Bankstown is using new technology to improve its road maintenance services.
The City of Canterbury Bankstown has undergone a trial of a program called Asset AI, in partnership with Transport for NSW and the Institute of Public Works Engineering Australasia (IPWEA), which can highlight – and will eventually predict – road safety issues like damaged signage, faded line markings, potholes and rutting, and can prioritise them for council maintenance workers based on severity and safety risk.
The program is designed to improve over time and is expected to eventually be capable of drawing on weather data to predict potholes and cracks before they even begin to form, and long before they become dangerous and costly to repair. Asset AI is being supported by the New South Wales Government, as preventative road maintenance is expected to majorly reduce the costs for councils by reducing their reliance on time-consuming road audits and increasing the life-span of roads through timely repairs.
Asset AI received $2.9 million through the State Government’s Smart Places Acceleration Program, a special reservation under the Digital Restart Fund, which was established in 2020 to support partnerships between the New South Wales Government and councils for the development of innovative projects.
Canterbury bankstown trial
In an interview with Infrastructure Magazine, the City of Canterbury Bankstown Mayor, Bilal El-Hayek, said that Canterbury Bankstown is a leader in using new technology to improve the way it provides community services.
“That’s why we put our hands up to host this trial,” Mayor El-Hayek said. “This is a fantastic opportunity to see if artificial intelligence (AI) can better manage the conditions of our roads. “Early results of the trial have shown us that AI can successfully detect many different types of defects within our roads including potholes, pavement cracking, faded line marking, damaged signs and even graffiti.
“The trial is still in the early stages, but we are hoping the technology will better inform our program of works so we can manage the conditions of our road network more effectively. “It’s helping us identify faults in the road before they become a bigger issue. Getting onto issues proactively and improving our ability to fix them.”
The primary benefit of using Asset AI to assess the condition of local roads is its ability to enable more effective preemptive maintenance. It is always more expensive to undertake repairs once serious damage or failure has occurred, so by detecting faults before they become severe, the AI is working to improve road safety and prevent traffic accidents, reduce infrastructure maintenance costs, reduce work time and thereby increase traffic efficiency, and extend the lifetime of roads.
In order to make the accurate predictions necessary for preemptive maintenance, the program needs to regularly be supplied with data on the state of the road network, local traffic conditions, and even weather patterns. To help with this, Mayor El-Hayek said that the technology is installed on the region’s street sweepers which are constantly out and about keeping roads clean.
“The program analyses each section of the road that the street sweepers drive down, giving us much more data and information on the condition of our roads.” Where councils typically only undertake road audits once every three to five years due to their cost and length, when used this way Asset AI has the potential to provide data on the condition of the area’s local road network as often as every fortnight at a much lower cost.
Maintenance costs up, federal funding down
In early September 2023, the Grattan Institute revealed research which found that federal funding to local governments was failing to keep up with the increasing costs of construction and maintenance. The research showed that over the past 20 years federal Financial Assistance Grants failed to keep pace with the rising costs of constructing and maintaining roads.
It also found that regional, rural and remote councils faced additional and sometimes insurmountable difficulties, due to smaller ratepayer bases, larger geographical areas, and less staff. Councils collectively manage around 75 per cent of Australia’s road network by length, while collecting less than four per cent of national taxation.
In the face of these challenges Asset AI has the potential to cut costs for councils by increasing the efficiency of roadworks, helping to ensure that local governments can afford to maintain their road networks with money to spare for other critical community services.
Early trials
Canterbury Bankstown conducted a pre-trial of the technology in 2021, which has since enabled initial camera and sensor trials to be undertaken across the Greater Sydney region. In mid-2022, a trial was undertaken across the city in which 32 public transport buses were outfitted with cameras and sensors to collect data on the state of the city’s road network to inform the program’s machine learning.
By mounting sensor equipment onto vehicles with regular routes, like garbage trucks, public transport buses and street sweepers, councils can make good use of resources that are already on their roads and defects can be detected passively, including those un-reported by residents.
An additional trial was undertaken outside the city in which a ute was mounted with the road-scanning equipment and used to scout 100km of regional roads to create a report on rural road conditions and to diversify the AI’s dataset.
The future of asset AI
Mayor El-Hayek said that council is planning to continue using the program into the future and working on ways to best utilise the technology to benefit its maintenance programs.
“We are going to continue to develop the project and have ambitions to expand it. “There are potential benefits in looking at ways of using machine learning with AI technology to predict where road defects are going to occur before they appear.
“This would allow us to undertake much more targeted and effective road maintenance. “Partnering with Transport for NSW and IPWEA in this trial has been very successful. Each organisation has brought their own unique skills and experience, which has helped us achieve great outcomes. “We are proud to be the council leading this technology for the state of New South Wales.”
