University of Hamburg - to research, to teach, to educate and form, to homepage
Hub for Crossdisciplinary Learning
Hub for Crossdisciplinary Learning

Photo: UHH/von Wieding

Hub for Crossdisciplinary Learning

Hamburg Bit-Bots: Learning robots – Bridging the Reality Gaps

The team of the Hamburg Bit-Bots has been founded more than ten years ago by interested and engaged students who organize themselves independently and who do research in different areas of robotic. New approaches are developed and evaluated and scientific results are presented at international conferences. There are more than ten active members at the moment. The student work group is part of the Department of Computer Science of the University of Hamburg and participates in at least one international robotic competition: the RoboCup. At the RoboCup robots compete different leagues. The long-term objective is to have a team of robots which is able to win against the human soccer world champion in 2050. On our way to reach that goal, we will do important basic research in the fields of robotic, image recognition, human-robot-interaction and artificial intelligence. The Hamburg Bit-Bots have always been involved in researching emerging technologies. For example, one of the first papers on the by now very popular Convolutional Neural Networks in the RoboCup Domain has been written by our work group.

We continue to research a lot in the field of Computer Vision, where we focus on a completely learning-based system. Furthermore, we examine the field of Motion on more natural, stable and less manually predefined motion sequences. Current research look into reward-based learning methods for running and other motions. We also contribute to the standardization and reuse of frequently used software components in Robotic. That enis why we engage in cooperations with other scientists and developers to improve the widespread robot software surrounding ROS. After a longer period of working in simulation due to Corona, the next step is the transformation of the knowledge gained into the real world. We especially focus on the following two points: Our Walking which has been improved by Reinforcement Learning and a new, more robust image recognition with depth perception. These are all progresses in the field of artificial intelligence which will lead to more natural, more robust and more secure robots and computer systems in the future.

More information:

Github:https://github.com/bit-bots
Website: https: //bit-bots.de

Research group

  • Valerie Bartel
  • Florian Vahl
  • Timon Engelke
  • Jan-Niklas Feldhusen
  • Ash Brandt
  • Joern Griepenburg

Mentor

  • Niklas Fiedler