UTS researcher Dr. Dinuka Abeywardena has calibrated a tiny camera to develop an intelligent drone that doesn't rely on GPS for navigation. Credit: Andrew To.
An autonomous robot created by Australian researchers is providing engineers with high-definition images of the integrity of steel structures that support the iconic Sydney Harbour Bridge.
The robot - dubbed 'CROC' - was developed by researchers at Sydney's University of Technology (UTS) with funding and engineering support from the NSW Roads and Maritime Services.
More stringent health and safety requirements have meant that inspectors from Roads and Maritime Services can no longer inspect particular locations at certain heights or in confined spaces.
CROC uses a magnetic foot to climb steel walls, avoid obstacles and move through small spaces inside bridges and other steel structures such as transmission towers, oil rigs and ship hulls. It collects and records inspection data and compares it with information gathered from previous visits, UTS said.
Team leader, Peter Ward, said researchers around the world are developing inspection robots and some are in commercial use. However, they have "limited intelligence and face difficulties overcoming challenging obstacles," he said.
CROC is different as it needs sight to work out the distance of particular objects. The robot uses sophisticated algorithms to explore its environment and construct a map before figuring out where to go and how to get there, without hitting other objects on the way.
"Many other similar projects are just confined to a lab, whereas CROC is a deployable solution to a real world problem," said associate professor, Jaime Valls Miro.
"It is an industry driven project that is actually achieving something from the data it collects, including high-definition photos of the current status of the Sydney Harbour Bridge."
Meanwhile, UTS researcher Dr. Dinuka Abeywardena has calibrated a tiny camera to develop an intelligent drone that doesn't rely on GPS for navigation.
Dr. Abeywardena said GPS signals can be disrupted, particularly when drones are indoors.
"What we are doing is differently is improving the perception capabilities of the drone," he said. "We're using a camera as its eye so that it can explore its environment and collect data around it."
Images are processed by an on-board computer to estimate how it moves and flies, its speed and tilt and the direction of the wind, he said.
Dr. Abeywardena's team is looking for industry partners to test the drones and navigational algorithms that have been tested in the lab.
Source: CIO Australia
Sign up for CIO Asia eNewsletters.