The World Robot Soccer Championship

The World Robot Soccer Championship

My group by UNSW Australia defeated staff B-Human out of Germany 3-1 final week to assert back Robocup SPL World Championships. Here is how we did this.

Each group employs the very same robots, so the contest is centered on applications and artificial intelligence (AI) advancement, not on hardware structure.

Both groups had powerful plans and well-tested code, but in the long run, our pace proved too quickly for the Germans to stay up with.

Late Push

We had three robots eliminate power as a consequence of heavy drops and were abruptly reduced to just two active robots on the area. Following our timeout however our bots got things back on course.

Our approach to winning was concentrated on harnessing our fast walking pace. Our robots can accelerate much quicker than many teams’ plus they attain a top speed of approximately 30 centimetres per minute (roughly 1 kmh). Even though this is slow in contrast to people, in robot football it is really fast.

After the ball is at our half of the area, we do not bother attempting to maneuver it to a teammate, we simply try to boot it up the opposite end of this area.

This contributes to us playing the majority of the sport within our competitor’s half of the area, where it is difficult for them to score goals, but simpler for us to score.

Listening For Whistles

The contest changes the rules annually to produce the games harder. This year, each of the finals matches were launched by a referee’s whistle, rather than the normal Wi-Fi message. This meant that the robots needed to listen out to the whistle before every kick.

We had been the only group to start our robots onto the referee’s whistle.

If just a robot observed a whistle, however, the other four did not, they determine that the a robot should have been incorrect, and do not play. If three from those five hear a whistle, however, the last two do not, they then pick both should have missed it and all of them start playing.

This vast majority vote system was critical in ensuring that our staff listened to this whistle faithfully.

Another of the most important reasons we believe we had been successful is the testing and development practices we now have as a staff.

We initiate the robot at exactly the identical set of places per time and week how fast it gets into the ball and shoots it to the objective. This fast highlights how successful the last week’s worth of growth has been and where we will need to improve. Additionally, it quickly reveals us some significant bugs we’ve introduced with the most recent set of modifications.

In addition, we play little practice matches of 5 versus 3 or 0 as we get nearer to contest. Primarily, this shows us just how well our robots rank and perform as a group, which can be important in a team game like football.

Second it shows how well we can perform in a shifting environment, playing a dwell resistance. It is always far more difficult to score when you have got competitions getting at all.

Looking Into The Future

The rest of the teams will be carefully observing our progress and focusing on their plans to conquer us. We’ll devote the next 12 weeks continuing to innovate and make sure that we deliver a much greater group to the contest next year.

Even though it’s a very long way off however, we’re anticipating using a home floor advantage in a couple of years time.

Comments are closed, but trackbacks and pingbacks are open.