Data conditioning for gravitational wave detectors: A Kalman filter for regressing suspension violin mode

TL;DR

This paper presents a Kalman filter-based method to model and subtract suspension violin mode vibrations from gravitational wave detector data, improving the clarity of potential gravitational wave signals.

Contribution

The paper introduces a novel Kalman filter approach for real-time estimation and removal of suspension violin mode vibrations in gravitational wave detectors.

Findings

Effective suppression of violin mode signals demonstrated in simulations.

Successful application of the filter to real LIGO prototype data.

Enhanced detector sensitivity by reducing narrow-band noise.

Abstract

Interferometric gravitational wave detectors operate by sensing the differential light travel time between free test masses. Correspondingly, they are sensitive to anything that changes the physical distance between the test masses, including physical motion of the masses themselves. In ground-based detectors the test masses are suspended as pendula and, consequently, thermal or other excitations of the suspension wires' violin modes lead to a strong, albeit narrow-band, ``signal'' in the detector wave-band that can confound attempts to observe gravitational waves. Here we describe the design of a Kalman filter that determines the time-dependent vibrational state of a detector's suspension ``violin'' modes from the detector output. From the estimated state we can predict that component of the detector output due to suspension excitations, thermal or otherwise, and subtractively remove those disturbances from the detector output. We demonstrate the filter's effectiveness both through numerical simulations and application to real data taken on the LIGO 40 M prototype detector.