Low-complexity Fusion Filtering for Continuous-Discrete Systems

TL;DR

This paper introduces a low-complexity distributed fusion filtering algorithm for mixed continuous-discrete multisensory systems, utilizing recursive equations and the covariance intersection method to efficiently compute cross-covariances.

Contribution

It proposes a novel recursive approach for local cross-covariances and integrates the covariance intersection algorithm for efficient distributed fusion filtering.

Findings

The algorithm reduces computational complexity in distributed filtering.

Theoretical analysis confirms the effectiveness of the proposed method.

Numerical examples demonstrate improved performance in multisensory systems.

Abstract

In this paper, low-complexity distributed fusion filtering algorithm for mixed continuous-discrete multisensory dynamic systems is proposed. To implement the algorithm a new recursive equations for local cross-covariances are derived. To achieve an effective fusion filtering the covariance intersection (CI) algorithm is used. The CI algorithm is useful due to its low-computational complexity for calculation of a big number of cross-covariances between local estimates and matrix weights. Theoretical and numerical examples demonstrate the effectiveness of the covariance intersection algorithm in distributed fusion filtering.