Jiuguang Wang

Seán McLoone: Robot soccer

The football game starts, and the players dash across the field. With control of the ball, the striker readies to shoot and the goalkeeper tenses in anticipation. With a swift kick, the ball glides across the pitch towards the goals. These players aren’t humans though. They are two-foot tall, autonomous humanoid robots, programmed by a team of engineers and computer scientists in the National University of Ireland Maynooth to compete in the annual world cup of robot soccer, RoboCupEuroScientist interviewed Seán McLoone, head of the Department of Electronic Engineering, about the team, and watched a game with team members Tom Whelan and Fiachra Matthews.

Barney lines up for a penalty shootout

How did you get involved in the competition?

Professor Rick Middleton of the Hamilton Institute, NUIM was previously in the University of Newcastle in Australia where he had been very active in RoboCup. Drawing on this experience, myself, Adam Winstanley (Head of Department in Computer Science) and Rick Middleton were able to establish a team.

It was 2008 when we first participated in Robocup. Most teams were still struggling to get the robots running, as were we, but we actually won the competition. We had formed a joint team with Australia that year and the following year we went out on our own and formed the Robo Eireann team. That year, we did ok, but we didn’t set the world alight, more like a typical Irish soccer team’s performance.

In 2011 we had a strong team but didn’t do particularly well in the soccer tournament. There’s a sob story there in terms of hard luck and bad decisions – just like real soccer! However, we did win the Open Challenge competition. This is where teams demonstrate the robotics research they are doing to address the challenges of robot soccer.

What are the technical challenges?

The robot joints are actuated by motors – you control the current and voltages to make the motors move, but turning this movement into reliable and stable locomotion is a major challenge.

The eyes of the robot are basically a webcam. We take for granted that we look at something and see objects – it’s not as simple as that! There’s software to turn an image into recognition of a ball, field lines, another robot, the goalposts. You need to know that “the ball is five metres ahead of me, and I need to kick the ball in that direction towards the goals”. The mathematics needed to achieve this is quite challenging.

Team play is also a key challenge. At the moment most teams don’t have particularly sophisticated team play. Our robots are still in that category, where it’s get the ball, and not quite hoof it up the pitch, but not far from it! They do have some level of awareness– a robot will recognise that it is nearest the ball, so it becomes the striker and the others will be defenders or supporting players.

It’s amazing, even though the focus is research, you do get really competitive. You get people shouting at the robots, telling them “Kick the ball, kick the ball!” even though you know they’re not listening to you at all because they’re not designed to. People get very passionate even though the robots are moving really slowly and a game is really painful to watch sometimes, because the ball might sit in the corner of the pitch for three minutes because the robots can’t see [it] and they’re wandering around crashing into each other and falling over.

The limited processing power available on board the robots presents an engineering challenge for the teams.

Do those reactions make you optimistic about the integration of robots into humans’ lives?

Once [robots] have actions and behaviours like humans, you start treating them like humans – no different to the way people talk to their pets and think they understand them. But it’s the social aspects of the interaction that will take a bit longer. A robot being sympathetic – no one will believe that to be genuine for a good while yet. Once we master that, they’ll be much more acceptable.

PhD student Tom Whelan with the 2011 Open Challenge trophy

What is it about robots that fascinates people?

I guess from a personal perspective, it’s that you can recreate a human effectively as a machine. The next stage is, can you transfer your own consciousness into a machine — then it is a way of becoming immortal. The other side of it is that robots are the ultimate servant and you don’t have any ethical issues about treating them badly.

[In Robo Cup,] there are also robots which don’t look humanoid. They’re basically cubes on wheels that can fire a ping pong ball around, and it’s almost like watching a squash game because the ball travels around so fast. Humans wouldn’t be able to beat those sort of computerised systems.

It’s a bit of a lofty aim. There are many practical issues to overcome. For example, the weight of battery that the robots have to carry around to be able to sustain gameplay for 90 minutes means that it just becomes impractical with current technology. By 2050, with the advances that people will make in AI programming, I’d be quite confident that from a team play and software perspective robots will be up to the challenge. You just have to look at chess playing computers – they are getting to the level where they can compete against the best in the world. I’m looking forward to being around long enough to see that but I wouldn’t bet my house on them winning.

Robocup video

Featured image credit: CC BY-SA 2.0 by Jiuguang Wang

EuroScientist is looking for contributors!

If you would like to write guest posts in EuroScientist magazine, send us your suggestions of articles at office@euroscientist.com.

Martin McKenna

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.