Particle Filter as A Controlled Markov Chain For On-Line Parameter Estimation in General State Space Models.
George PoyiadjisSumeetpal S. SinghArnaud DoucetPublished in: ICASSP (3) (2006)
Keyphrases
- parameter estimation
- state space
- markov chain
- particle filter
- markov chain monte carlo
- estimation problems
- model selection
- monte carlo
- transition probabilities
- importance sampling
- gibbs sampling
- visual tracking
- object tracking
- random fields
- maximum likelihood
- markov random field
- particle filtering
- least squares
- state estimation
- motion model
- observation model
- dynamic programming
- kalman filter
- posterior density
- optimal policy
- em algorithm
- proposal distribution
- reinforcement learning
- approximate inference
- appearance model
- mean shift
- expectation maximization
- likelihood function
- bayesian inference
- posterior distribution
- search space
- maximum likelihood estimation
- maximum a posteriori
- mobile robot