Deep Reinforcement Learning of Dexterous Pre-grasp Manipulation for Human-like Functional Categorical Grasping.
Dmytro PavlichenkoSven BehnkePublished in: CoRR (2023)
Keyphrases
- manipulation tasks
- reinforcement learning
- human robot interaction
- object manipulation
- real robot
- robot navigation
- robot control
- motion planning
- function approximation
- humanoid robot
- human hand
- categorical data
- gesture recognition
- service robots
- robotic systems
- markov decision processes
- state space
- haptic feedback
- model free
- reinforcement learning algorithms
- real time
- learning algorithm
- learning process
- virtual environment
- dynamic programming
- optimal policy
- search algorithm
- transfer learning
- human activities
- policy search
- neural network