On a continent where the tyranny of distance is a day-to-day reality, autonomous approaches underpin the operation of mining, agriculture, transport and more.
The International Space Station soars through skies at an average low-orbit height of 400 km. As human operations go, that’s impressively remote – but it’s nothing compared to a remote iron ore operation in the distant north of Western Australia.
There, resource extraction machines trundle along under the supervision of advisors sitting in an operations centre in Perth, more than 1000 km away.
“In Western Australia, we are the largest iron ore producer in the world,” said Michelle Keegan, who finished up her term as the director of the Australian Remote Operations for Space and Earth (AROSE) consortium in April 2025. “We mine four times more iron ore than 20 years ago at the beginning of the iron ore boom, and we weren’t able to get there without being able to automate the equipment.
“Moving billions of tonnes of iron ore today is done with automated trucks and automated drills – and Rio Tinto has automated trains.”
Some of the country’s biggest economic operations, then, are carried out with little human involvement at all – and many of the humans who are involved sit at a lonely remove from the earth-hauling and machine-scraping action.
For some industries in Australia, thanks to the hardships imposed by distance and harsh terrain, autonomous systems aren’t a hope for the space-age future; they’re a very real part of day-to-day life.
Country to city
Rio Tinto’s automated train, a system named AutoHaul, has shuttled iron ore from the Pilbara to coastal port facilities about 800 km away for the past seven years. But in a very different environment, on the opposite side of the country, the cargo comes in the form of the urban commuter.
Opening in the northwestern suburbs in 2019 and extending into the city from 2024, the Sydney Metro network demonstrated that driverless trains could operate equally well in a bustling metropolis as they could crossing the desert.
Among public transportation specialists, trains are rated according to their level of autonomy, beginning with GoA0, which contains no level of autonomy whatsoever. Sydney Metro trains are rated GoA4, the highest grade.
“GoA4 indicates that it is a fully automated driverless system,” said Kate Ford CPEng, the Trains, Systems, Operation and Maintenance Delivery Director for Southwest at Sydney Metro.
“With GoA3, you have a driver on board who would drive to a speed profile, and the system would automatically brake to stop at a station, but the driver could control doors opening and closing, or train departures. Whereas in a GoA4 system, everything is automated.”
The benefits of such an approach can be seen immediately, Ford said.
“One of them is about repeatability – human interactions with systems drive small, minute differences,” she said. “With an autonomous system, every train behaves in the same way, which means you can be more efficient on braking from an energy perspective. We’ve got regenerative braking in the city because it’s repeatable.”
The autonomous approach also increases the safety and reliability of the Sydney Metro, a particularly important attribute for a system that sees trains leaving a station every four minutes during peak hour.
Removing humans from the decision-making tasks surrounding arrival and departure – or opening and closing carriage doors – allows the system to operate identically each time, without delay or disruption.
“The flip side is obviously that it requires far more testing. You have to do far more assurance on all of the systems that feed in to ensure repeatability can happen safely.”
Similarly, Keegan explained, the repeatability facilitated by automation has meant miners can maintain production with a limited available workforce, or ensure that their machines don’t overcut or undercut a seam.
“At the beginning of the [iron ore] boom in 2005, there was an exponential demand for steel, which meant that, if we were to continue to maintain our share of iron ore production, we had to do something different to be able to grow,” Keegan said. “If you look at different case studies around remote operations, a lot of people would say it’s important to improve safety – but it’s very important to give consistency in outcomes and operations.”
READ: 5 Australian autonomous systems already in service
Robots on the range
Reliability, consistency, safety: the operations around Australia that use autonomous systems look very different and service a broad range of industries, but they’re all designed to take advantage of these common goals.
For more than a decade, Dr Salah Sukkarieh, Professor of Robotics and Intelligent Systems at the University of Sydney’s Australian Centre for Robotics, has worked on perfecting agricultural robots that help farmers with such tasks as weeding, spraying pesticides, and monitoring and moving livestock.
His experience extends beyond that; he’s worked on intelligent robotic platforms for aerospace, ports, environmental monitoring, disaster response, mining and more.
Regardless of the domain in which he’s operating, Sukkarieh approaches the task of designing and working an autonomous system in much the same way each time.
“You sit there, and you define the same type of requirements and activities,” he said. “I’ll be looking at an environmental axis, I’ll be looking at a human-involvement axis, and I’ll be looking at a complexity-to-the-task axis.
“Along those three axes, you ask: What is it about the environment? What is it about the tasks? How much human involvement does there need to be? That frames the choices about what you might do with an autonomous system.”
When Sukkarieh first began working with robotics, while researching for his PhD, systems were less developed and required resourceful cross-disciplinary knowledge application. He brought together understanding of electronics, sensors, platforms and algorithms – and, eventually, experience with the domains in which his robots would operate, such as horticulture.
Today, the field has matured, and robotics can be used as an enabler rather than approached as a platform.
Someone who specialises in the AI aspect of a system, for instance, might not need in-depth knowledge of the robot that is operating their algorithm.
“If you’re a roboticist, you look at the fundamental requirements of what the autonomous system will need to do, or be like or act like, or perform or operate in. That changes the way you design the solution,” Sukkarieh said.
“If you focus more broadly on intelligence systems, you’re looking at that overall picture. Does the AI need to work in real time or is it OK offline? Do the sensors need to capture a lot of information or need to process that information in real time? Does the platform work in very rugged terrain – in which case it’ll need other sensors that tell me how its attitude is going, its position, its velocity.”
Placing the system’s requirements at the forefront leads to further questions. Is the primary goal of the platform information-gathering, or is it important for it to interact with its environment?
And while machine learning is not a new feature of autonomous systems, the increasing sophistication of AI is expanding their potential. Faster and more accurate algorithms offer the potential for systems to be engaged in reasoning tasks, rather than preset activities.
Sukkarieh describes the thought process an agricultural robot could go through today, all without a farmer’s input: “It might rain in three hours; harvesting will happen tomorrow; irrigation was switched on at 6:00 o’clock; pests in block A; therefore, I will spray block C.”
Interoperability is another factor that is growing in importance for autonomous systems, according to Colin Sheldon FIEAust CPEng, Manager, Asset Management and Performance at Aurecon and chair of Engineers Australia’s Mechanical College.
“Interoperability is going to keep developing and make it easier to interface, and that starts to democratise autonomous system development,” he said.
“There’s always been an argument for IP and proprietary-access systems, but it’s becoming more of an understanding from original [equipment manufacturers] and other manufacturers that this is coming, so they need to align with the standards, and develop their outputs and inputs so they work with other systems.”
Made for Australia
From Sukkarieh’s perspective, it’s no surprise autonomous systems have flourished in Australia.
“We’ve led the world in many ways,” he said. “It’s a large land, and it’s very much focused on primary industries. Nobody wants to work out in the middle of nowhere, and the weather conditions are challenging.
“If you had to say what the factors are that allow autonomy to flourish if you decide to go down that path, Australia has those right conditions.”
According to Dr Sue Keay, the founder and chair of Robotics Australia Group, the country has always had a great reputation in the realm of field robotics, which is concerned with robust systems that typically work outdoors in unstructured environments.
“The good news for Australia is that we specialise where robotics needs to be,” she said. “Australia had to get very good at developing technologies that help us overcome remote distances and all of the requirements to service the huge amounts of infrastructure that we have.”
That has meant developing technologies that can operate in remote areas without being energy intensive – preferably relying on long-life batteries – and creating systems that can operate in communications-denied environments.
“These are all things fairly unique to the Australian context that have been important for Australian roboticists to solve for,” Keay said. “Fortuitously, that happens to be the direction robotics is heading in, now that we’re moving towards self-driving cars, and humanoid robots that need to be able to operate in unstructured environments and behave reliably and effectively.”
Sheldon notes that the immense task of managing complexity and risk makes autonomous solutions particularly attractive, especially for mining.
“In terms of the scale of the equipment and potential related incidents, there’s a lot of time spent managing the safety of the person operating the equipment,” he said.
A helping hand
As autonomous systems become increasingly sophisticated, Sukkarieh is noticing that the range of industries to which they can be applied is expanding.
“As technology advances, it also gets cheaper in various forms, which opens up more solutions to industries that originally couldn’t afford it,” he said. “Mining probably helped facilitate a lot of the knowledge and technology that went into agriculture, and now I think there is potential in environmental monitoring [because] it’s very similar to agriculture.”
Keegan said the future of autonomous systems is “up to our imagination” and expects these newer sectors to feed ideas back into some of the more established industries.
“What’s exciting is that there are other non-traditional players thinking about this and challenging the norms,” she said. “If we think particularly about the copper sector and the fact that we need to be able to learn how to mine much deeper and do that in a safe way, it’s up to us to design that future operation and make it real.”
That means new technologies, new approaches and perhaps even closer integration between the mining and space sectors.
“We are at a huge inflection point with robotics, automation and AI,” Keegan said. “The next five years are going to be super-exciting.”
Aerial taxis
As the 2032 Brisbane Olympic Games draw closer, officials in South East Queensland are turning to the skies to ease the sprawling region’s congestion problems.
In 2022, the South East Queensland Council of Mayors signed a partnership with Wisk Aero, a fully owned subsidiary of Boeing dedicated to electric vertical take-off and landing (eVTOL) flight – or aerial taxis.
Dr Kai Li Lim, of the University Queensland’s Dow Centre for Sustainable Engineering Innovation, said the plans are far from fantasy.
“[The technology] is moving quickly enough for us to see a really feasible implementation by the Brisbane Olympics,” he said. “It will most likely be marketed as a premium product that caters to the more affluent folks, and then after that – once you get market traction and scale – you can make it more affordable.”
While eVTOLs can be operated autonomously like a drone, Lim expects that, at first, they will have human pilots.
“This is more of a regulatory and safety issue than a technological issue, because the technology is quite established,” he said. “Once you start to put passengers in the vehicle, the stakes are a lot higher, and that’s why you want to have a human inside there.”
Instead, technological challenges revolve around creating more energy-efficient batteries that don’t add too much weight to the payload. The taxis would also need supporting infrastructure.
“Building vertiports, for example, and having the right grid capacity. Especially when talking about urban areas, how do you secure land for the vertiports? Then, before selecting them, you have to look into electrical grid capacity, because even if you’re talking about a few eVTOLs, you’re talking about megawatts of power.”
Although it will be important for the government to oversee any new aerial transportation modes, Lim is optimistic that the regulators will get it right, particularly considering eVTOLs are much quieter than helicopters, the most comparable vehicle.
“There has been a lot of consultation and discussion in relation to how we’re going to regulate eVTOLs in the country. We have always been a more risk-averse country when it comes to our policies, and we do put safety first, so it’s just having those mechanisms to ensure that these services can be rolled out.
This article was originally published in the November 2025 edition of create with the headline “Automatic for the people”.
ON-DEMAND WEBINAR: Intelligent robotics for civil infrastructure construction and maintenance
