Robot Learning in Partially Observable, Noisy, Continuous Worlds.
Reid BroadbentTodd PetersonPublished in: ICRA (2005)
Keyphrases
- partially observable
- decision problems
- state space
- markov decision processes
- dynamical systems
- reinforcement learning
- partial observability
- infinite horizon
- partially observable domains
- action models
- partial observations
- action space
- belief state
- markov decision problems
- partially observable environments
- hidden state
- heuristic search
- belief space
- utility function
- optimal control
- finite state
- reward function
- domain specific
- general purpose
- computational complexity