An open robotics environment motivates students to learn the key concepts of artificial neural networks and reinforcement learning

Abstract

Educational robotics is a widely recognized tool to motivate students and concretize abstract and complex topics, such as artificial intelligence in computing education curricula. Lego Mindstorms series is one of the most popular robotics platform due to its flexibility and relatively cheap price. We used Lego Mindstorms EV3 robots with a novel Open Learning Environment for Artificial Intelligence (OLE-AI) to teach concepts of reinforcement learning and artificial neural networks (ANNs) to computer science students. OLE-AI uses a white box approach to expose internal structures of an ANN to students. Results from the pilot study with OLE-AI indicate that the participating students were able to deepend their knowledge about AI topics through a practical and open exercise that involved them in controlling EV3 robots by manipulating the ANN and Q-Learning algorithm.