Training an Interactive Humanoid Robot Using Multimodal Deep Reinforcement Learning.
Heriberto CuayáhuitlGuillaume CoulyClément OlalaintyPublished in: CoRR (2016)
Keyphrases
- humanoid robot
- reinforcement learning
- multi modal
- motor skills
- real robot
- motion planning
- biologically inspired
- human robot interaction
- imitation learning
- supervised learning
- motion capture
- fully autonomous
- joint space
- computer graphics
- optimal policy
- deep architectures
- sensory feedback
- motor learning
- motor control
- real time
- body movements
- model free
- function approximation
- human motion
- state space
- multi agent
- learning algorithm
- machine learning
- manipulation tasks
- spatio temporal
- rough terrain
- walking speed