On-line Reinforcement Learning for Trajectory Following with Unknown Faults.
Yves SohegeGregory M. ProvanPublished in: AICS (2018)
Keyphrases
- reinforcement learning
- function approximation
- optimal policy
- fault detection
- initially unknown
- markov decision processes
- fault diagnosis
- learning algorithm
- model free
- state space
- dynamic programming
- machine learning
- action selection
- robotic control
- multiple faults
- multi agent
- optimal control
- learning process
- test cases
- spatio temporal
- model based diagnosis
- supervised learning
- reinforcement learning algorithms
- trajectory data
- root cause
- temporal difference learning
- multi agent reinforcement learning