论文信息 - Lego Mindstorms Robots as a Platform for Teaching Reinforcement Learning

Lego Mindstorms Robots as a Platform for Teaching Reinforcement Learning

This paper examines the suitability of Lego Mindstorms robotic kits as a platform for teaching the concepts of reinforcement learning. The reinforcement learning algorithm Sarsa was implemented on board an autonomous Mindstorms robot, and applied to two learning tasks. The reasons behind the differing results obtained on these two tasks are discussed, and several issues related to the suitability of Mindstorms as a platform for teaching or experimentation with reinforcement learning are identified.

Peter Vamplew | P. Vamplew

[1] Mahesan Niranjan,et al. On-line Q-learning using connectionist systems , 1994 .

[2] Alexander Zelinsky,et al. Q-Learning in Continuous State and Action Spaces , 1999, Australian Joint Conference on Artificial Intelligence.

[3] Sridhar Mahadevan,et al. Automatic Programming of Behavior-Based Robots Using Reinforcement Learning , 1991, Artif. Intell..

[4] Deepak Kumar,et al. A robot laboratory for teaching artificial intelligence , 1998, SIGCSE '98.

[5] Elizabeth Sklar,et al. RoboCupJunior: Learning with Educational Robotics , 2003, AI Mag..

[6] Minoru Asada,et al. Purposive Behavior Acquisition for a Real Robot by Vision-Based Reinforcement Learning , 2005, Machine Learning.

[7] Peter Stone,et al. Reinforcement Learning for 3 vs. 2 Keepaway , 2000, RoboCup.

[8] Karl Sims,et al. Evolving virtual creatures , 1994, SIGGRAPH.

[9] Richard S. Sutton,et al. Generalization in ReinforcementLearning : Successful Examples UsingSparse Coarse , 1996 .

[10] Thomas G. Dietterich,et al. Advances in neural information processing systems : proceedings of the ... conference , 1989 .

[11] Richard S. Sutton,et al. Learning to predict by the methods of temporal differences , 1988, Machine Learning.

[12] Elizabeth Sklar,et al. RoboCupJunior: A Vehicle for Enhancing Technical Literacy , 2002, AAAI Mobile Robot Competition.