Robot Challenges as an Opportunity for Students

Progress in computer science, in particular research on pattern recognition, gained speed considerably when public challenges were announced and competitions were organized. Examples are the NIST data set of postal addresses to test optical character recognition or segmentation challenges which measure performance of visual recognition tasks. Interpretation of sensor data and vision are core topics in robotics. Robotic challenges, such as RoboCup, serve as testbeds for solutions to vision problems as well as robotic architectures. Teams of the Koblenz University have participated in such competitions for almost two decades. Led by a few researchers, teams of several students solved the tasks set in these challenges quite successfully. In this contribution, we outline the history of these achievements. We describe how the students were selected, how the teams were built, and which structure of the teams was most effective. We give examples of how the academic freedom, financial and personal support, passionate researchers, and strategic planning led to several first places in international robotic challenges.