Model-Free control performance improvement using virtual reference feedback tuning and reinforcement Q-learning.
Mircea-Bogdan RadacRadu-Emil PrecupRaul-Cristian RomanPublished in: Int. J. Syst. Sci. (2017)
Keyphrases
- model free
- reinforcement learning
- impedance control
- reinforcement learning algorithms
- function approximation
- temporal difference
- visual servoing
- policy iteration
- optimal control
- control strategies
- adaptive control
- control system
- action selection
- reinforcement learning methods
- control strategy
- visual feedback
- average reward
- feedback loop
- state space
- machine learning
- least squares
- transfer learning
- markov decision processes
- learning problems
- force feedback
- markov decision problems
- policy evaluation
- control method
- continuous state and action spaces