Physics-Based Dexterous Manipulations with Estimated Hand Poses and Residual Reinforcement Learning.
Guillermo Garcia-HernandoEdward JohnsTae-Kyun KimPublished in: IROS (2020)
Keyphrases
- reinforcement learning
- residual error
- human hand
- state space
- reinforcement learning algorithms
- dynamic programming
- markov decision processes
- temporal difference
- learning algorithm
- deformable models
- object manipulation
- humanoid robot
- function approximation
- policy search
- temporal difference learning
- action selection
- data sets
- learning problems
- genetic algorithm
- machine learning