Bayesian Fusion of Data Partitioned Particle Estimates

TL;DR

This paper introduces a Bayesian data fusion technique that combines particle estimates from data subsets, enabling effective multi-sensor data integration with proven convergence and demonstrated success in orbit and tracking applications.

Contribution

It proposes a novel Bayesian fusion method for particle estimates from data subsets, with convergence guarantees and practical multi-sensor application demonstrations.

Findings

Method converges in the particle limit.

Effective in multi-sensor orbit determination.

Successful in bearings-only tracking.

Abstract

We present a Bayesian data fusion method to approximate a posterior distribution from an ensemble of particle estimates that only have access to subsets of the data. Our approach relies on approximate probabilistic inference of model parameters through Monte Carlo methods, followed by an update and resample scheme related to multiple importance sampling to combine information from the initial estimates. We show the method is convergent in the particle limit and directly suited to application on multi-sensor data fusion problems by demonstrating efficacy on a multi-sensor Keplerian orbit determination problem and a bearings-only tracking problem.