Integrating POMDP and Reinforcement Learning for a Two Layer Simulated Robot Architecture.
Larry D. PyeattAdele E. HowePublished in: Agents (1999)
Keyphrases
- reinforcement learning
- simulated robot
- real robot
- multi layer
- state space
- continuous state
- partially observable
- markov decision processes
- function approximation
- partially observable markov decision processes
- optimal policy
- learning capabilities
- hidden state
- real time
- temporal difference
- model free
- reinforcement learning algorithms
- multi agent
- machine learning
- markov decision process
- policy evaluation
- learning algorithm
- middle layer
- partially observable markov decision process
- action selection
- real world
- partial observability
- markov models
- continuous state spaces
- dynamic programming
- data sets