Time-varying feedback particle filter☆
Abstract
Feedback particle filter is a novel Monte Carlo algorithm with identically distributed particles evolving under feedback control structure, such that the Kullback-Leibler divergence between the actual posterior of the state and the common posterior of any particle can be minimized. In this work, we consider the time -varying linear systems and explicitly analyze the errors between the optimal solution obtained by Kalman filter and the estimates given by feedback particle filter and the optimal transportation particle filter, respectively. These theoretical analyses are also supported by the numerical simulation, where we compare the performances of particle filter, feedback particle filter, optimal transportation particle filter and Kalman filter. (c) 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
