Australians for years have tried to come up with creative solutions to rid our land of pests.  One of the most creative was the introduction of the cane toad to get rid of the cane beetle.  "And we all know how that went."

The latest and world first is the most innovative solution of all, robots. Drs Matthew Dunbabin and Feras Dayoub from Queensland University of Technology’s Institute of Future Environments have developed the COTSbot, the world’s first robot designed to control the Great Barrier Reef’s crown-of-thorns starfish population.

Dr Dunbabin explains, “COTSbot is a revolutionary robot system designed specifically to visually identify crown of thorn starfish and administer a single injection to help control the numbers that we currently face on the Great Barrier Reef.

“COTS (crown of thorn starfish) lives naturally in the Great Barrier Reef but in recent years they have become quite prevalent in outbreak numbers across the Great Barrier Reef and have been attributed to 40 per cent  of the total coral decline that has been occurring second only to cyclones. It is an invasive pest that is destroying large areas of coral,” said Dr Dunbabin.

With divers being able to disseminate and differentiate the coral landscape and identify what is and is not a COTS, Dr Dunbabin states the robot is not trying to compete with divers.

“Divers are incredible at what they can do, the problem is the scale of this problem is so huge that there’s just not enough divers out there at the moment to be able to make a full impact across the entire reef, so we have tried to create a tool that can replicate current diver capabilities, at least a subset of those, where we can target only crown of thorn star fish.

“We are not trying to out compete a diver,  we’re trying to help control this pest and we do that specifically by creating a computer vision algorithm that is very specific to crown of thorn star fish. We have done this by training it on hundreds of thousands images of this particular crown of thorns star fish and the system learns what is a star fish and what is not and if we have any doubts we can get the robot to collect an image and a human can verify what we are looking at,” said Dr Dunbabin.

With a visual data base of COTS amassed during the process, Dr Dunbabin said You Tube is where they began the robots training of the sophisticated recognition system.

“There is so much coral that looks like COTS and it is hard to distinguish between the two.  We have collected images and have a visual data base of COTS,” he said.

Once the COTS is identified by the robot, the eradicating process is a single injection system into the COTS after which there is a reaction within 24 hours and then it dies.

The lethal injection is a bial salt created by James Cook University and replaces a multi injection system.  The robot uses an extended arm to administer a single injection fatal dose of bile salts.  There is no residual effect on the Great Barrier Reef and adopts the Great Barrier Reef Marine Park Authority’s COTS Control Guides.

The robot is around 1.3 metres and weights less that 30kg.  It is yellow with five motors and can administer up to 200 bial salt injections for up to eight hours.

Researchers are also currently working on a lighting system as COTS come out at night and feed on coral in the dark and will be particularly useful as divers are not able to dive in the dark.

Dr Dunbabin advised that divers are not to be eliminated, they are dextrous and achieve things which are well beyond the capabilities of the robot.  This is a tool to eliminate a great bulk of the COTS and keep populations under control.

After six months of intensive training, COTSbot passed its first sea trial with flying colours in Moreton Bay this week.

The next challenge is to see how it navigates strong water currents and more complex terrain in open ocean.

“We are currently doing reef trials now and then a five month evaluation,” said Dr Dunbabin.

Researchers are hoping to have the robot fully operational on the reef early next year.

The machine is the first robot in the world that uses real-time image processing, that does all its’ processing on board and the first robot designed to eradicate a naturally occurring pest.