A Scalable Track-Before-Detect Method With Poisson/Multi-Bernoulli Model

TL;DR

This paper introduces a scalable track-before-detect tracking method utilizing a Poisson/multi-Bernoulli model, which improves performance for low-observable objects while maintaining computational efficiency.

Contribution

The paper presents a novel scalable TBD tracking approach that approximates the multi-Bernoulli mixture posterior with a multi-Bernoulli pdf, enhancing efficiency and detection capabilities.

Findings

Significantly improved tracking performance over existing methods

Effective detection of low-observable objects

Reduced computational complexity through approximation and recycling

Abstract

We propose a scalable track-before-detect (TBD) tracking method based on a Poisson/multi-Bernoulli model. To limit computational complexity, we approximate the exact multi-Bernoulli mixture posterior probability density function (pdf) by a multi-Bernoulli pdf. Data association based on the sum-product algorithm and recycling of Bernoulli components enable the detection and tracking of low-observable objects with limited computational resources. Our simulation results demonstrate a significantly improved tracking performance compared to a state-of-the-art TBD method.