Trained AI cameras could greatly speed up & optimize future search & rescue operations over large areas.

AI can help SAR operations ‘stay frosty’. Source: AFP

Australian Air Force tests AI for search and rescue missions

  • AI and search & rescue missions have rarely been explored together… if at all
  • Australian Airforce AI team and their collaborators were challenged to do precisely that
  • The 1st AI cameras were used in ‘proof of concept’ trials recently to acclaimed success

Air and sea search and rescue (SAR) operations have largely been done the same way for over a hundred years, by using technology such as radar to get a general search area– and then using good ol’ unaided visual search to scan for objects in the search area.

At the same time, Australia is no stranger to leading artificial intelligence (AI) research for the national interest.

Wing Commander Michael Gan, the Australian Air Force Deputy Director of Artificial Intelligence (AI), said his team known as ‘Plan Jericho’ was challenged to augment the way SAR had been done for years using technology.

Like many other applications of AI and machine learning across industries, training algorithms to spot anomalies that the human-eye might miss could be hugely effective in SAR operations where vast stretches of empty ocean must be scoured for signs of life.

“Jericho saw the opportunity to use AI to augment and enhance SAR. The idea is to train a machine learning algorithm and AI sensors to complement existing visual search techniques,” said Wing Commander Gan. “Our vision is to give any aircraft and other defence platforms […] a low-cost improvised SAR capability.”

The Australian Air Force and the Australian Maritime College teamed up with the Queensland volunteer coastguard to test the results.

Dubbed AI-Search, a small camera is fixated to an aircraft window and looks for shapes and imagery out on the open sea – in particular, life rafts and lifeboats – that it has been trained to recognize using a combination of AI and machine learning.

The Australian Air Force and its trial partners have just completed the first set of ‘proof-of-concept’ trials off the coast of Stradbroke Island, near Queensland, Australia. “The first test flight worked brilliantly. It will be the first in a series to develop and evaluate the proof of concept,” said Wing Commander Gan.

AI-Search is a result of a collaborative joint project between Plan Jericho, the Warfare Innovation Navy Branch, Air Mobility Group’s 35SQN, and the University of Tasmania’s Australian Maritime College.

“Jericho enlisted the help of budding AI guru, Lieutenant Harry Hubbert, Warfare Innovation Navy Branch. We gave Harry a challenge – find an orange hull, in a large body of water, using AI, and do it in a month.

“Harry developed the algorithms in his own time within two weeks,” Gan said.

In order to test the algorithm and gather data to train the IA, the Australian airforce was enlisted to fly the AI-Search team and sensors in a C-27J aircraft over the Tamar River in Tasmania, capturing images of a range of life rafts, lifeboats and safety equipment.

The project appears to be the first time AI and machine learning capabilities have been applied to search and rescue efforts, although it is by no means the first AI project to involve the military and national security concerns.

Being applied towards potentially life-saving applications was hailed as a positive step in the right direction by several parties, including the Deputy Commander of Australian Volunteer Coastguard at Southport, Michael Hoffman.

“With this new technology, the Air Force is able to locate these rafts and vessels in the water and drop aid to them immediately, allowing us time to get out to them,” commented Hoffman.

“[The Air Force using AI cameras] can cover so much more area than we can, in a very short space of time– so for us, for them to locate and then be able to give us coordinates to it, it saves us hours and hours of searching.”