An Improved Random Matrix Prediction Model for Manoeuvring Extended Targets

TL;DR

This paper introduces an enhanced prediction update method for extended target tracking using a generalized non-central inverse Wishart distribution, improving accuracy and flexibility over existing models.

Contribution

It develops a novel prediction update incorporating a non-central inverse Wishart distribution and a new tuning parameter, requiring only one divergence minimization, thus advancing extended target tracking methods.

Findings

Improved tracking accuracy over state-of-the-art methods

Simplified prediction update requiring only one divergence minimization

Enhanced model flexibility with an additional tuning parameter

Abstract

This paper proposes an improved prediction update for extended target tracking with the random matrix model. A key innovation is to employ a generalised non-central inverse Wishart distribution to model the state transition density of the target extent; resulting in a prediction update that accounts for kinematic state dependent transformations. Moreover, the proposed prediction update offers an additional tuning parameter c.f. previous works, requires only a single Kullback-Leibler divergence minimisation, and improves overall target tracking performance when compared to state-of-the-art alternatives.