Retro-RL: Reinforcing Nominal Controller With Deep Reinforcement Learning for Tilting-Rotor Drones.
I Made Aswin NahrendraChristian TirtawardhanaByeongho YuEungchang Mason LeeHyun MyungPublished in: IEEE Robotics Autom. Lett. (2022)
Keyphrases
- reinforcement learning
- actor critic
- optimal control
- operating conditions
- control policy
- adaptive control
- function approximation
- reinforcement learning algorithms
- temporal difference
- model free
- rl algorithms
- control strategies
- state space
- induction motor
- control system
- optimal policy
- markov decision processes
- machine learning
- control strategy
- closed loop
- learning algorithm
- reinforcement learning methods
- approximate dynamic programming
- mathematical model
- control law
- fuzzy controller
- control scheme
- transfer learning
- learning capabilities
- action selection
- control algorithm
- function approximators
- multi agent
- temporal difference learning
- policy gradient
- neural network
- state and action spaces
- markov decision process
- initial state
- learning classifier systems
- learning problems
- learning process