Design of Dedicated Tilt-to-Length Calibration Maneuvers for LISA

TL;DR

This paper proposes and evaluates tilt-to-length calibration maneuvers for LISA, demonstrating that sinusoidal attitude modulations can accurately estimate coupling coefficients within 20 minutes, improving noise mitigation.

Contribution

It introduces a low-uncertainty tilt-to-length calibration method using rotation maneuvers, optimizing parameters for in-flight noise reduction in LISA.

Findings

Sinusoidal modulations of ~30 nrad at ~43 mHz are effective.

Multiple maneuvers at different frequencies can be performed simultaneously.

Calibration achieves sub-15 um/rad uncertainty after 20 minutes.

Abstract

Tilts of certain elements within a laser interferometer can undesirably couple into measurements as a form of noise, known as tilt-to-length (TTL) coupling. This TTL coupling is anticipated to be one of the primary noise sources in the Laser Interferometer Space Antenna (LISA) mission, after Time Delay Interferometry (TDI) is applied. Despite the careful interferometer design and calibration on the ground, TTL is likely to require in-flight mitigation through post-processing subtraction to achieve the necessary sensitivity. Past research has demonstrated TTL subtraction in simulations through the estimation of 24 linear coupling coefficients using a noise minimization approach. This paper investigates an approach based on performing rotation maneuvers for estimating coupling coefficients with low uncertainties. In this study, we evaluate the feasibility and optimal configurations of such maneuvers to identify the most efficient solutions. We assess the efficacy of TTL calibration maneuvers by modulating either the spacecraft attitude or the Moving Optical Sub-Assembly (MOSA) yaw angle. We found that sinusoidal signals with amplitudes of around 30 nrad and frequencies near 43 mHz are practical and nearly optimal choices for such modulations. Employing different frequencies generates uncorrelated signals, allowing for multiple maneuvers to be executed simultaneously. Our simulations enable us to estimate the TTL coefficients with precision below 15 um/rad (1-sigma, in free space) after a total maneuver time of 20 minutes. The results are compared to the estimation uncertainties that can be achieved without using maneuvers.