by Annabelle Powell, Journalist, Infrastructure Magazine
Worldwide spending on artificial intelligence (AI) is increasing, with huge investments by government parties and the private sector into AI technologies. At the 2021 Critical Infrastructure: Digitisation Series Virtual Conference, Dr Sue Keay, Chair, Robotics Australia Group, explored how robotics and AI technologies are currently being implemented in the critical infrastructure sector, and its future applications.
Dr Keay was previously CEO of the Queensland AI Hub, and research director for cyber-physical systems at CSIRO’s Data61. She led the development of Australia’s first robotic roadmap, outlining how robotics and automation will impact every sector of the Australian economy.
Dr Keay said the level of interest and investment around the world in AI technologies can be attributed to a ‘tipping point’ in our understanding of AI.
“The reason that there is such a huge investment in AI and robotics at the moment is because many people believe we’ve reached a tipping point where these technologies, which have been around for more than 70 years, are finally reaching a point where they’re going to become much more useful and usable by all of us,” Dr Keay said.
“Now we have far more computer processing power in our handbags or in our pockets, in the forms of our mobile phones, than we ever had in the past.
“One of the other reasons that we are starting to see this huge increase in interest in artificial intelligence, thanks to all of this computer processing power, has been the advent of computer vision – a form of AI that allows computers or robots to automatically interpret the images that they are able to retrieve from cameras or video cameras.”
Dr Keay said we’ve now come to the point where we can place cameras on different objects and they can be interpreting information in real time, removing humans from dangerous situations.
Another pivotal development in AI is a type of machine learning known as ‘deep learning’; a way by which robots or computers learn how to do things, not by writing lines of specific code about what action the robot needs to take, but by being shown what they need to do.
Advantages of AI for Australian infrastructure
Australia was the first country in the world to automate its ports – an example of how implementing AI into infrastructure management can produce great results.
“Automated straddle carriers are in operation in the Brisbane and Botany Bay ports,” Dr Keay said. “No people are actually present on the port when the straddle carriers are in operation, they move them onto the back of the trucks and the trucks transport those goods to wherever they need to go.”
Australia is also a vast country with a significant amount of remote infrastructure that requires maintenance, so technology plays a key role in this vital work.
“We have thousands of kilometers of roadways, pipelines, rail lines, power lines, and bridges that all need to be inspected and maintained. We really don’t have enough people on the ground to be able to do all of this work without the help of technology,” Dr Keay said.
“Some of the ways that AI is being applied in infrastructure at the moment is for remote inspection to cover those thousands of kilometers of infrastructure that need to be inspected.
“To have a look at the deterioration of those structures, and then also do predictive and preventive maintenance, and finally to look at lifecycle optimisation so that we can track all of our assets and figure out when they need to be replaced and what particular aspects of those assets might need to be replaced,” Dr Keay said.
“The use of these technologies is producing such huge amounts of data that we can really use it to good effect, to have a thorough understanding of our assets and their lifecycles.”
Maintaining key Australian assets
One example of how this technology is being applied is on the Sydney Harbour Bridge, which has had more than 2,400 sensors deployed, including accelerometers.
“They’re all sort of linked into a daisy chain to connect them and they’re monitoring the vibrations that are occurring in the Sydney Harbour Bridge in real time.
Then machine learning algorithms work out whether those vibrations are indicative of there being potential cracks forming, or whether they’re simply the result of temperature and climate conditions of the day,” Dr Keay said.
As these machine learning models receive more data, they are able to learn and become better at predicting where there are likely to be areas that need to be considered for preventative maintenance.
An Australian company has also developed a ‘Fitbit’ for machines; a small device that can be put on any type of machine to inspect and perform vibration analysis to work out when there’s likely to be a failure of that piece of equipment.
“This can be very handy if you’re dealing with critical infrastructure as the failure of some of these instruments can have catastrophic consequences,” Dr Keay said.
Drones are also being used in Australia for the inspection of many of our assets, including solar and wind farms, and Dr Keay said the main objective is to make sure workers are removed from dangerous situations.
“The collection of all of this data from the numerous sensors that are now being deployed out in the environment allows us to build digital twins of much of our critical infrastructure.
“In the future, if we’re able to deploy our sensors and robots in a clever manner, we could hope that we don’t have to send people out to fix things on these remote oil rigs or out in remote areas of Australia, and that we will actually be able to deploy robots to do that same job,” Dr Keay said.