Artificial intelligence is allowing drones to spot and assist swimmers in distress. Ripper Group, an Australian drone services company, today announced it is adding world first autonomous capabilities to the drones used by the Westpac Rescue Helicopter program.

“Little Ripper” drones are already in use on Australian beaches and have successfully assisted several live rescues. The flying drones can help spot and monitor swimmers as well as drop life saving devices. Up until now they have required a human pilot to remotely control the device from the shore and spot swimmers. 

New “Smart Artificial Intelligence” developed in conjunction with UTS unveiled by the company today at the Amazon Web Services Public Sector Summit in Canberra means Ripper Group drones can now also operate and identify swimmers autonomously.

Ripper Group cofounder Dr Paul Scully-Power said the new ability, which includes very low latency, means in the future long range drones would be able to monitor Australia’s un-patrolled beaches – where swimmers usually get in trouble.

A mobile app, released today, lets rescuers access video feeds from the drones when they detect a swimmer in distress using AI.

“To spot someone in distress, with artificial intelligence you can do much better than a human,” Scully Power told Which-50. “So that makes it totally unique.”

Ripper Group also uses drones to identify and monitor large marine life on Australian beaches. Some of the drones are even equipped with audio devices to warn surfers of nearby sharks. 

Training the software to spot sharks, including distinguishing them from whales, dolphins and crocodiles, allows long term monitoring – currently in use – helped build the foundations for identifying swimmers in distress.

“There’s certain signs you can pick up,” Scully-Power explained. “Like splashing around too much, [swimmers] waving their arm. You teach the artificial intelligence if they do a series of things like this then they are probably in distress.”

The software uses cloud computing for what Scully-Power calls “command and control”.

“Right now, while we are developing, all the video goes in to AWS and we have artificial intelligence looking over all that. The next step beyond that is having what I call ‘Smart Artificial Intelligence’: it will only transmit to the cloud incidents that we need to respond to.”

Previous post

Facebook will show users what data it's collecting on them around the web to target ads

Next post

85 per cent of Zendesk's systems are managed outside IT — but don't call it Shadow IT