RoboCup Soccer

Resources^{6 total}

• Project

  Real-time object detection and avoidance for autonomous Nao Robots performing in the Standard Platform League

  A classic object detection algorithm is combined with a relatively tiny deep neural network, which makes it possible to increase the reliability.

• Thesis

  Enhancing simulation images with GANs

  Automatic generation of realistic images makes it possible to generate enough data to train modern deep neural networks.

• Publication

  Finding the orientation of a soccer robot by looking at the surroundings

  By learning color patterns in the surroundings, a soccer robot can estimate reliably its heading.

• Conference paper

  Showcase demonstration of the AI techniques applied in RoboCup soccer

  For an audience of AI experts, the techniques behind whistle detection, localization and opponent detection are demonstrated.

• Conference paper

  Learning to start a soccer game by listening to the referee's whistle

  By using the right audio preprocessing techniques, all classic machine learning methods were able to recognize the whistle with high accuracy.

• Conference paper

  Anticipate an attackers behavior by analysing moving patterns.

  During a penalty shootout, a goalkeeper has to decide when to dive, and in which corner. Analysing the movement from optical flow helps to predict the moment.