A Square-Root Factorized Multiplicative Extension to the Particle Flow Filter
Abstract
Particle-based methods for recursive Bayesian estimation provide a means to obtain precise solutions in challenging problems. The subset of particle flow filters compute the Bayesian update by modeling the motion of the samples from their a priori locations through the state space to represent the a posteriori beliefs. This work expands the filtering capabilities of the particle flow estimation framework, specifically in the context of the recently developed information flow variant. A means of incorporating attitude estimation mathematically consistent with the underlying parameterization is developed in addition to a square-root factorized flow model to promote numerically stable uncertainty representation. An entropy-based convergence control for the flow is also presented that circumvents the need for ad hoc selection of an underweighting-like parameter. The resulting square-root factorized, multiplicative information flow filter with convergence control is applied to a lunar descent-to-landing navigation simulation to demonstrate the new capabilities and resulting performance.
Recommended Citation
K. C. Ward and K. J. Demars, "A Square-Root Factorized Multiplicative Extension to the Particle Flow Filter," Advances in the Astronautical Sciences, vol. 175, pp. 2723 - 2741, Springer, Jan 2021.
Department(s)
Mechanical and Aerospace Engineering
International Standard Book Number (ISBN)
978-087703675-3
International Standard Serial Number (ISSN)
0065-3438
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2024 Springer, All rights reserved.
Publication Date
01 Jan 2021
