Knowledge Technology Paper wins "Best Paper Award" at the 20th Robocup International Symposium
4 July 2016, by Thomas Hummel
Photo: UHH/Knowledge Technology
The Paper "Speck, D., Barros, P., Weber, C. and Wermter, S. Ball Localization for Robocup Soccer using Convolutional Neural Networks" won the "Best Paper Award for Engineering Contribution" at the 20th Robocup International Symposium. In RoboCup, the football league for robots, the localization of the ball has become harder since 2016, as the ball is not unicoloured anymore, but has a pattern like the balls we know from human football. The paper presents an approach using Convolutional Neural Networks (CNN) to locate the new ball. The comittee members highlighted that the approach does not need any preprocessing and that this really emphasizes the idea of neural networks for localization tasks in robotics.