Wednesday 2 January 2019 11:53 am

Back to the future: UK scientists complete testing of AI-powered autonomous Mars robot

To boldly go where no other robot has gone before, the UK Space Agency have finished testing a Mars rover that will be able to make its own decisions on where to explore, enabling it to travel much further than existing robots.

The Martian robot will use artificial intelligence (AI) software, developed by researchers at King's College London and Airbus, to travel over a kilometre without human interaction. 

Currently, hand-guided robots can only travel a few dozen metres per day due to the time it takes to receive commands from Earth.

Having a mind of its own will mean rovers can deliver more scientific returns per mission, collecting samples and analysing the landscape to determine factors such as whether life has ever existed on the Red Planet.

"Mars is a very difficult planet to land safely on, so it’s essential to maximise the discoveries from each successful touchdown," said Catherine Mealing-Jones, director of growth at the UK Space Agency.

"New autonomous robot technology like this will help to further unlock Mars’ mysteries and I’m delighted that the UK is a key player in this cutting-edge field."

The testing was carried out in the Martian-like Sahara Desert throughout December on a robot called Sherpa, with other scientists from Thales Alenia Space, the University of Strathclyde and GMV-UK also taking part.

The technology will be loaded on to a new UK-built rover to be launched as part of the Exomars programme, which will land on Mars in 2020. The name of the robot has yet to be announced, after being decided in a public competition last year.

The UK Space Agency is the second largest European contributor to Exomars, having invested £260m in the mission programme and a further £14m on the instruments.

Aside from the autonomous AI framework, researchers also tested a variety of new technologies including data fusion systems, a plug-and-play sensor suite and an open-source operating system for controlling the robot.