AI technology used in drones for public safety

Keeping everyone safe is in the best interest of individuals, companies and governments. For one university, Artificial Intelligence (AI) and safety were combined to tackle a danger that occurs on many Australian beaches – sharks.

The University of Technology Sydney (UTS) and the Ripper Group created a shark-spotting drone to protect beach-goers in Australia. Speaking at the Industrial Internet 4.0 Summit in Sydney in February, UTS professor Michael Blumenstein explained how the university collaborated with industry to provide solutions to issues using AI.

“It just so happens that there’s a lot of AI sitting in universities. We will work with industry to create an environment of engagement.” He said UTS undertakes research that is very transferable to industry and is able to be commercialised.

“We collaborated with the Ripper Group to develop the world’s first shark-spotting drones to be able to detect sharks from video surveillance imagery in the water using a unmanned aerial vehicle (UAV).

“The processing of the video, which is data, needs to be trained and needs to be autonomous. So we used an AI learning algorithm for that.

“The technology is now being deployed at over 15 beaches in New South Wales and Queensland and by the end of this year, I suspect it will be up to about 56 because the Ripper Group has just got another contract with Westpac,” said Blumenstein.

With applications such as this one, there is room to use it in different fields. “It will be able to attract not just sharks but many other animals in the water. It will also be able to detect any danger. It’s there to protect beach goers and save lives.

“This a great example of AI that’s actually been applied for the purpose of a social good application,” he said.

Room for improvement
While advances in technology are helping people build products such as the shark-spotting drone, Blumenstein explained that there is plenty of research needed to push AI further.

“In 2019 we are pretty much seeing the top of a peak that’s been building for about a year where people are actually saying they are employing it in their business, in industry and in university.

“But actually, there’s a lot of things that AI can’t do.”

One sticking point for AI is a lack of human touch. Where AI is at now, humans are able to detect when they are speaking to a computer and when they are speaking to another human.

“Will there be a time where the human can’t detect the difference of whether the conversation is coming from a computer or a human being? To this day we haven’t passed the test yet.

“Even with the simplest things we take for granted, such as speech and being able to understand the fact that we can communicate as humans, regular computers still haven’t got the perfect grasp of that yet.”

Blumenstein and other industry experts shared their views on the current state of Industry 4.0, challenges business and industry face, and how technology can advance industry – including manufacturing.