Learning model-free robot control by a Monte Carlo EM algorithm.
Nikos VlassisMarc ToussaintGeorgios KontesSavas PiperidisPublished in: Auton. Robots (2009)
Keyphrases
- monte carlo
- em algorithm
- robot control
- model free
- reinforcement learning
- expectation maximization
- maximum likelihood
- temporal difference
- learning process
- parameter estimation
- mixture model
- learning algorithm
- temporal difference learning
- learning tasks
- function approximation
- mobile robot
- policy evaluation
- maximum a posteriori
- importance sampling
- unsupervised learning
- generative model
- autonomous robots
- particle filter
- markov chain
- policy iteration
- genetic algorithm
- reinforcement learning methods