Reinforcement Learning with Adaptive Curriculum Dynamics Randomization for Fault-Tolerant Robot Control.
Wataru OkamotoHiroshi KeraKazuhiko KawamotoPublished in: CoRR (2021)
Keyphrases
- robot control
- fault tolerant
- reinforcement learning
- fault tolerance
- distributed systems
- mobile robot
- autonomous robots
- load balancing
- subsumption architecture
- unstructured environments
- state machine
- dynamic model
- motion control
- safety critical
- pid controller
- function approximation
- dynamical systems
- learning algorithm
- adaptive systems
- fuzzy logic
- artificial intelligence