Skip to main navigation Skip to search Skip to main content

The Interacting Multiple Model Feedback Particle Filter for the State-dependent Switching Diffusion Systems

  • Yiyang Chen
  • , Ruoyu Wang
  • , Dengyu Yang
  • , Xue Luo*
  • *Corresponding author for this work
  • Beihang University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this paper, we develop the interacting multiple model-feedback particle filter (IMM-FPF) for the state-dependent switching diffusion systems (SD-SDS). Starting from the Kushner's equation of the hybrid state, we derive the evolutionary equations for the modes' and the continuous states' probabilities given the mode. The IMM is used to merge the estimate of the continuous states from all the modes. As for the implementation, the constant gain approximation is used in the FPF. At last, the numerical example of a three-mode SD-SDS is experimented. Both the accuracy and the efficiency are compared with the IMM-particle filter.

Original languageEnglish
Title of host publication11th 2025 International Conference on Control, Decision and Information Technologies, CoDIT 2025
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages242-247
Number of pages6
ISBN (Electronic)9798331503383
DOIs
StatePublished - 2025
Event11th International Conference on Control, Decision and Information Technologies, CoDIT 2025 - Split, Croatia
Duration: 15 Jul 202518 Jul 2025

Publication series

Name11th 2025 International Conference on Control, Decision and Information Technologies, CoDIT 2025

Conference

Conference11th International Conference on Control, Decision and Information Technologies, CoDIT 2025
Country/TerritoryCroatia
CitySplit
Period15/07/2518/07/25

Keywords

  • feedback particle filter
  • interacting multiple model
  • nonlinear filtering
  • switching diffusion process

Fingerprint

Dive into the research topics of 'The Interacting Multiple Model Feedback Particle Filter for the State-dependent Switching Diffusion Systems'. Together they form a unique fingerprint.

Cite this