Windowing and random weighting based cubature RTS smoothing for target tracking

TL;DR

This paper introduces a novel adaptive smoothing algorithm combining windowing and random weighting techniques to improve target tracking accuracy by dynamically estimating noise statistics.

Contribution

It proposes the WRWACRTS algorithm, integrating windowing and random weighting with convergence proof, enhancing nonlinear smoothing over traditional methods.

Findings

Demonstrates improved tracking accuracy in numerical simulations.

Provides convergence proof using Lyapunov functions.

Outperforms conventional cubature RTS smoothers.

Abstract

This paper presents windowing and random weighting (WRW) based adaptive cubature Rauch Tung Striebel (CRTS) smoother (WRWACRTS). The Unscented KF (WRWUKF) has already existed as an alternative to nonlinear smoothing solutions. In the proposed method, both windowing and random weighted estimation methods are combined together and used to estimate the noise statistics. Subsequently, the weights of each window are adjusting randomly, and update the process and measurement noise covariances matrices at each epoch. The developed WRWACRTS algorithm overcomes the limitation of the conventional CKS. The Lyapunov function-based approach is used to investigate the convergence proof of the WRWACRTS algorithm. A numerical example is shown to demonstrate the performance of the proposed algorithm.