From Video Game to Real Robot: The Transfer between Action Spaces.
Janne KarttunenAnssi KanervistoVille HautamäkiVille KyrkiPublished in: CoRR (2019)
Keyphrases
- video games
- real robot
- action space
- reinforcement learning
- state space
- skill learning
- markov decision processes
- mobile robot
- robot arm
- state and action spaces
- simulated robot
- learning experience
- game play
- computer games
- real valued
- multi robot
- educational games
- serious games
- stochastic processes
- robotic systems
- real world
- markov decision process
- action selection
- single agent
- game design
- continuous action
- motion control
- dynamic programming
- learning process
- multi agent
- learning algorithm
- machine learning
- control law
- humanoid robot
- mathematical model
- pose estimation
- optimal policy
- data sets
- real time