Week 3 Flashcards
How can a linear function of the state estimator be estimated?
Why can’t we use the same strategy for non-linear functions of state estimators? What should we do instead?
Please explain why Monte Carlo estimations work.
How is smooted density denoted? How is it defined in the case of a Kalman smoother?
Why is it important that (when doing a Monte Carlo simulation) that we draw from the smoothed density, instead of making independent draws?
What does the tilde over a variable denote? What does the superscript (i) denote?
Why are recursive methods preferred when generating from a smoothed density?
What does this lemma imply?
How is one draw of alpha tilde generated by Monte Carlo simulation?
How is the following simulated?
What are some well-known examples of non-linear non-Gaussian models? How can they be the most generally formulated?
How is the nonlinear Gaussian model (nLGM) formuled?
How can a Taylor expansion be applied to the nLGM model?
What is the Linearized nLGM model? What is the advantage of using such a model?
What is the Extended Kalman Filter and the Extened Kalman Smoother?
