Across all modes of freight transportation, companies and governments are leveraging innovative technologies to better manage the growing demands of the freight industry. One of these technologies being implemented by Transport for NSW (TfNSW) is a world-first, Australian-designed, solar-powered camera. Over the past year, the number of these devices installed across the state have only grown.
While challenges may present themselves at this initial stage, Road Freight NSW CEO Simon O’Hara is hopeful the device can pay dividends in due course.
“We understand this is a trial process. However, we hope that the machine-learning camera technology will focus on providing data that will deliver productivity dividends for the NSW freight industry and supply chain,” O’Hara told ATN.
More than 20 of these solar-powered cameras have been installed across NSW, spanning from mid-north coast town Raleigh to border town Tocumwal. Across these locations, the cameras take high-definition pictures of heavy vehicles. Using artificial intelligence (AI), the camera can identify different types of heavy vehicles along with load changes.
In a fact sheet released last year, TfNSW says the project was developed to understand if AI and machine-learning can adapt together to extract heavy vehicle information. Its other objective was to help TfNSW better understand heavy vehicle and freight movements in the state. Along with these objectives, TfNSW says the cameras will enable it to plan how NSW’s growing freight load can be managed.
UNSW senior lecturer Dr Elnaz Irannezhad says that challenges could begin presenting themselves to TfNSW during the camera’s trial phase, with the technology currently only being operational during the day. Even then, Irannezhad says the camera may find it hard to identify the types and sizes of trailers at its current angle.
“When you’re on the roadside, the only lane visible to the camera is the one next to it,” Irannezhad told ATN.
“In a multi-way highway, there will be heavy vehicles in other lanes, so there could be some confusion for the machine-learning technology.”
Irannezhad says graphics processing unit (GPU) prices could also see the technology prove to be less cost efficient than other freight camera technologies currently available in NSW. One of these technologies being utilised at heavy vehicle stations is the Infra-Red Traffic Logger (TIRTL).
The device was first developed in 1997 by Australian company CEOS alongside TfNSW. With its infrared and radar settings, Irannezhad says that it can cover all types of traffic flow and precisely identify different types of heavy vehicles.
The first TIRTL device was deployed in NSW a year after they first began developing it. Since then, the device has been used across many major countries, including the United States, New Zealand, Germany and the Netherlands. Irannezhad says that the only challenge that these cameras face is that they have been caught in the wake of hedge trimmers.
TIRTLs aren’t the only device being used across the state. Irannezhad says TfNSW has invested heavily in installing Automatic Number Plate Recognition (ANPR) cameras at these heavy vehicle rest sites.
“With these cameras you can determine the number plate, the vehicle’s weight and the number of axles, you’ll have a full configuration of the truck,” Irannezhad says.
“When you log it, then you’ll have the very unique footprint of this vehicle, which means you can distinguish the origin, locations or its movement patterns.”
Irannezhad says there are still challenges aplenty for the new camera as it becomes as well-known as the TIRTL. It could still be beneficial to the industry now with its ability to distinguish between heavy and passenger vehicles easier than before.
More benefits could present themselves as TfNSW becomes more acquainted with the device’s workings. Among those learnings include TfNSW teaching its users the many algorithms and images that form the camera technology. Irannezhad says that, with the number of cameras, users will have to be trained to identify 1,000 images of heavy vehicle variations.
“If you employ it in one location, then it’s feasible because you’d only need 100 or 200 images,” Irannezhad says.
“But once you add it to a different highway and camera angle, then you basically need to finetune that training.”
While the camera may take time to have a major impact on the freight industry when its technology is in its initial stages of development, it could still help shape NSW’s freight future. As more research and development takes place on it, O’Hara says the camera will help reduce the risk of issues that present themself to the industry
“The new counting and classifying cameras will assist in understanding freight movements to help reduce congestion, improve road safety outcomes, and encourage more efficient deliveries,” O’Hara says.