Modelling coloured residual noise in gravitational-wave signal processing

TL;DR

This paper presents a new signal processing model using Student's t distribution to handle unknown and heavy-tailed noise in gravitational-wave data, improving robustness over traditional Gaussian models.

Contribution

It introduces a novel t-distribution based model for non-white noise with unknown spectrum, enhancing robustness to outliers in gravitational-wave signal processing.

Findings

Model effectively captures heavy-tailed noise in gravitational-wave data.

Provides a natural extension to Gaussian noise models for uncertain spectra.

Demonstrates improved handling of outliers in real detector data.

Abstract

We introduce a signal processing model for signals in non-white noise, where the exact noise spectrum is a priori unknown. The model is based on a Student's t distribution and constitutes a natural generalization of the widely used normal (Gaussian) model. This way, it allows for uncertainty in the noise spectrum, or more generally is also able to accommodate outliers (heavy-tailed noise) in the data. Examples are given pertaining to data from gravitational wave detectors.