A new unmanned, underwater robot could be the key in the latest efforts to track changes and help protect and manage one of Australia’s greatest treasures – the Great Barrier Reef. Thanks to the Google Impact Challenge Australia, that robot could soon become a reality.
The Great Barrier Reef Foundation has teamed up with Centre researchers Matthew Dunbabin and Feras Dayoub from the Queensland University of Technology (QUT) to enter the challenge, and has just been named one of ten finalists in this year’s challenge. As a finalist, the foundation is now in the running for one of four $750,000 grants.
Doing research to try to save the reef is nothing new for Dunbabin and Dayoub. In the past two years, they developed an underwater robot to help control the crown-of-thorns starfish (COTS), which is responsible for a significant decline in coral cover over the past few decades. Called the COTSbot, the underwater robot is able to identify a COTS and inject it with a solution to kill it.
Dunbabin and Dayoub now want to build a low-cost, more versatile version of the COTSbot. They’re calling it RangerBot.
“We learnt a lot from COTSbot, what works, what doesn’t work,” said Dr Dunbabin. “What we learnt will be brought into the new design.”
The COTSbot was designed with one purpose in mind. That won’t be the case with RangerBot.
“This robot will be more like a Swiss Army knife,” said Dr Dunbabin. “It can be used for multiple applications and will be easy to deploy.”
“The robot will be able to collect data from the reef, build maps, monitor changes, and provide scientists with rich, georeferenced information, which is very hard for divers to do,” said Dr Dayoub.
Dr Dunbabin is an Associate Investigator with the Australian Centre for Robotic Vision. Dr Dayoub is a Research Fellow with the Centre. The Centre’s mission is to combine robotics with computer vision and machine learning, giving robots the ability to see and function in the real world. The COTSbot is one of the Centre’s early successes.
“Computer vision and robotics are a natural combination,” said Dr Dayoub. “Images provide us with very rich information, and with all the advances happening now in machine learning, it opens the door for endless possibilities.”
Centre Director Peter Corke believes the RangerBot project would provide a low-cost alternative for monitoring the health of the reef.
“We know the reef is under stress but we need to how much stress and where that stress is,” said Prof Corke. “We need to measure what’s going on, but more importantly we need to measure what’s going on over time, to look for trends and seasonal patterns.”
“This is where robots have a great advantage, being able to work by autonomously collecting data without human supervision.”
Giving an underwater robot the ability to see is no easy task, though.
“You’ve got people now who say they do machine learning and computer vision, but they have a super computer behind them. I can’t run an extension lead out to my robot,” says Dr Dunbabin.
“That’s one of the key challenges we have, how to make a robust vision system that can run real-time on minimum power inside the robot.”
The other obstacle facing the team is the difficult environment of the reef itself.
“Our type of robot needs to navigate the currents and obstacles. We need to be close to the coral to collect the information we want, but not too close where we’re actually colliding with what we’re trying to protect,” said Dr Dayoub.
The Google Impact Challenge Australia was created to help not-for-profit organisations develop technologies that can help tackle the world’s biggest social challenges.
Drs Dunbabin and Dayoub hope they can help the Great Barrier Reef Foundation with its efforts to save the reef. Winning the Challenge would certainly help them develop a robot that could really make a difference.
“We would have the opportunity to create something new that could potentially change the lives of the one billion people who depend on the reef,” said Dr Dunbabin.
One of the prizes is the $750,000 People’s Choice Award, which the public can now vote for online.
Media Contact: Tim Macuga 07 3138 6741 or 0478 571 226 email@example.com
Australian Centre for Robotic Vision
2 George Street Brisbane, 4001
+61 7 3138 7549