Reachability-based Trajectory Safeguard (RTS): A Safe and Fast Reinforcement Learning Safety Layer for Continuous Control.
Yifei Simon ShaoChen ChaoShreyas KousikRam VasudevanPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- state space
- control problems
- optimal control
- control system
- machine learning
- control policy
- multi layer
- control loop
- neural network
- control strategy
- dynamic programming
- multi agent
- upper layer
- function approximation
- data acquisition
- path planning
- learning classifier systems
- reinforcement learning algorithms
- mobile robot
- trajectory data
- action space
- continuous state spaces
- robotic control
- continuous state and action spaces
- fitted q iteration