Composable Deep Reinforcement Learning for Robotic Manipulation.
Tuomas HaarnojaVitchyr PongAurick ZhouMurtaza DalalPieter AbbeelSergey LevinePublished in: ICRA (2018)
Keyphrases
- reinforcement learning
- manipulation tasks
- real robot
- function approximation
- robotic control
- mobile robot
- perception action
- imitation learning
- real time
- multi agent
- robot navigation
- tactile sensing
- state space
- model free
- temporal difference
- action selection
- markov decision processes
- reinforcement learning algorithms
- robotic systems
- robot control
- human robot interaction
- motion planning
- multi robot
- dynamic programming
- machine learning
- real world
- reward function
- learning classifier systems
- learning capabilities
- optimal policy
- multi agent reinforcement learning
- belief nets
- learning algorithm