Final presentation "Biologically-inspired Learing"
At the end of the term, the students of the lecture on “Biologically-inspired Learning for Humanoid Robots” presented their final projects. Their task was to program a Nao robot to learn to kick a ball at a soccer goal. The robot learned the task through reinforcement learning, which means that the robot received positive or negative feedback, depending on the outcome of each executed kick. Positive feedback was given when scoring a goal, negative feedback when scoring failed, for example when missing the goal. With enough trials, the Nao thus learned this task and scored (almost) every time. Congratulations for passing!
A video is on our YouTube channel: https://youtu.be/gdB985MmwxM