Dynamics-Enabled Safe Deep Reinforcement Learning: Case Study on Active Suspension Control.
Zhaojian LiTianshu ChuUros KalabicPublished in: CCTA (2019)
Keyphrases
- case study
- reinforcement learning
- control problems
- optimal control
- highly nonlinear
- robot control
- control system
- air fuel ratio
- control policy
- control strategy
- control method
- active control
- real world
- data sets
- real time
- model free
- control strategies
- action selection
- markov decision processes
- dynamic programming
- nonlinear dynamics
- function approximation
- lessons learned
- adaptive control
- reinforcement learning algorithms
- learning problems
- machine learning
- hybrid intelligent