AIML scientists were part of a winning team in NASA’s global space robotics competition. Here’s why they got involved

Story written by Dr Sarah Keenihan, AIML

A group from The University of Adelaide was recently named as one of only 22 global teams to qualify for the Stage 2 competition round of the Space Robotics Challenge run by NASA in partnership with Space Center Houston. 

Among the contributing scientists were Australian Institute for Machine Learning (AIML) researchers Tom Rowntree, Andrew Du, Tat-Jun Chin, Sam Bahrami, Kiet To and Ian Reid. 

The Space Robotics Challenge is designed to identify and develop capability to advance autonomous robotic operations for space exploration missions on the surface of other worlds, such as the Moon and Mars.

“The competition is a great way for us to apply our machine learning and artificial intelligence capability to solve space problems,” said Tat-Jun, who is Professorial Chair of Sentient Satellites at AIML.

“It’s the sort of work that is really exciting, and maybe it will inspire new generations of researchers to study and become involved in machine learning, or the space industry.” 

Out of 114 registered teams, only 22 – including the one from University of Adelaide – successfully addressed the NASA needs and submitted qualifying entries, allowing them to move on to the final competition round.

“It was not a simple problem,” said Tat-Jun. 

“We were provided with a simulated moon environment, and we had to design algorithms to control a robot to move around on that surface.” 

But it wasn’t enough to just move around – the robot should also be able to find and identify valuable resources, and then collect and return them to a lunar processing plant. 

It’s the sort of activity that NASA and other space agencies and entities envisage for a future where space mining and remote operations on moons and planets is likely. Valuable resources might one day include water, fuel ingredients, or high value minerals. 

But why get involved? Tat-Jun said competitions provide important opportunities for AIML scientists. 

“This sort of work, participating in competitions like this, it’s a good match with the sort of academic work we already do,” he said. 

“But it extends it further, it allows us to further develop our skills and test capabilities.” 

“It also gives us a platform to show others what we’re capable of – it’s like a demonstration to share with other academics, collaborators, potential investors in technology and even just the general public who may be interested in our work,” said Tat-Jun. 

The final round of the Space Robotics Challenge will commence in late January.