The lecture will provide a comprehensive summary of statistical inference using particle filters for positioning and tracking applications. The lecture will cover the motivation behind developing such frameworks, the underlying motion and measurement models, assumptions, and the Bayesian approach towards solving such an estimation problem. The talk will also highlight the importance of linearity, Gaussian, and Markovian assumptions, enabling a tractable solution framework also popularly termed as Kalman filter (KF). Insights will be provided into discrete time and continuous time KFs, including the steady state stability conditions. Further, to deal with non-linear systems, the talk with discuss the linearization based extended Kalman filter.
Reference: Niclas Bergman, “Recursive Bayesian Estimation Navigation and Tracking Applications”, Department of Electrical Engineering Linkoping University, Sweden