Humanoid robots up to 60 centimeters compete against each other in the RoboCup humanoid league. Video cameras mounted on the robots’ heads help them to recognize their position on the field, as well as the position of the ball and the opponent. The robots must move like humans on two legs and must be able to move toward the ball – a tennis ball. Kicking a ball is easy for humans; for robots it requires a high degree of motor coordination. In Magdeburg the FUmanoids were even able to demonstrate how to throw a ball with both arms into the field, and they were able to dribble the ball between the goal posts. For this performance, they were declared winners of the “Technical Challenges.”
Building robots that can play various games is just one step in the development toward production and service robots of the future. There are already robots that play billiards, computers that master chess (for example, Deep Blue from IBM), and even robots that kick real soccer balls to test their dynamic properties for sports firms. Robot soccer is particularly interesting because team coordination plays a role. The ball must reach the goal, and that is the objective of an entire team.
The FUmanoids are built by students of Freie Universität Berlin as part of a project in the Artificial Intelligence Group. Each year the students design new robots and write new software, for example, for computer vision and communication. Freie Universität Berlin will participate in the World Cup in Istanbul July 5 – 11 with new types of robots that can run faster and more securely toward the ball. These robots were not used in the preparation tournaments (where the robots from last year were used), so the Berlin team is expected to be even more effective in July. This will be necessary because the competition from the Far East is very strong. The FUmanoids will be demonstrated on the Dahlem campus during the Long Night of Science on May 28.
For further information, please contact:
Raúl Rojas, Department of Mathematics and Computer Science, Freie Universität Berlin, Artificial Intelligence Group, Tel.: +49 (0)30 / 838-75130, Email: Raul.Rojas@fu-berlin.de