Active Improvement of Control Policies with Bayesian Gaussian Mixture Model.
Hakan GirginEmmanuel PignatNoémie JaquierSylvain CalinonPublished in: IROS (2020)
Keyphrases
- gaussian mixture model
- control policies
- maximum likelihood
- mixture model
- optimal policy
- em algorithm
- feature vectors
- control policy
- motion control
- expectation maximization
- background subtraction
- gaussian mixture modeling
- finite horizon
- reinforcement learning
- control system
- control strategies
- bayesian networks
- feature space
- action space
- probabilistic model
- long run
- similarity measure
- learning algorithm