RoboCup results in wins for local universities

As robotics researchers, students and enthusiasts gathered at the International Convention Centre in Sydney, home-town teams from universities in Sydney were hoping for simulated and solid success.

Two local universities came out with first place honours, University of Sydney (USYD) and the University of Technology Sydney (UTS), in the Soccer Simulation 2D League and the Social Robot Standard Platform League, respectively.

Demonstrating the breadth of applications for artificial intelligence (AI) and autonomous systems in 2019, RoboCup brought together teams from around the world to compete in soccer, industrial and rescue leagues, among others.

The team from USYD worked in partnership with researchers from CSIRO’s Data61 in building autonomous virtual robots, called “agents” to play two-dimensional virtual soccer.

Teams were unable to assist their “agents” with remote controls.

Up against the defending champions, the USYD and CSIRO team’s software, named “Fractals”, used virtual acoustic, physical and visual sensors to pass, shoot and score on the cyber playing field.

“We are so delighted that our team has won the Soccer Simulation 2D League at the 2019 RoboCup. The grand final was a very tense game which went into extra time after a goalless full-time, with Fractals scoring a winning goal with less than a minute to go,” said USYD Faculty of Engineering professor Mikhail Prokopenko.

The UTS team, named “UTS Unleashed!”, had to design a robot which could collaborate with people in realistic situations, and won the Social Robot Standard Platform League.

Robots in this league must be able to socially interact and adapt their behaviours.

All teams in this competition had to use a supplied robot, in this case provided by Softbank Robotics.

The robots had to perform the role of a party host, housekeeper and operate in a simulated restaurant environment, within which UTS Unleashed! achieved the highest point score, 800.

“With our research focus on developing safe and secure robots for human-centric environments, we are looking for ways social robots can operate with not just the intelligence to optimise and perform physical tasks but to work with humans in reliable and enjoyable ways,” said professor Mary-Anne Williams, who led the team.

These robots also had to be resilient in difficult situations, said Williams.

“We need robots that can keep operating in the home even when there are networks outages or other issues beyond the ability of the average person to fix. So we need to apply our technical expertise and ensure that a robot can keep running even if it’s being repaired remotely,” said Williams.

In the marquee league, the Standard Platform League, the team from the University of New South Wales, rUNSWift, came third, behind teams from two German universities.