Improving the Sensitivity of LISA

TL;DR

This paper develops a formalism using computational algebra to optimize LISA's sensitivity by analyzing data combinations, leading to significant improvements in signal-to-noise ratio through eigenvector-based methods and a rotating detector model.

Contribution

It introduces a systematic algebraic approach to identify optimal data combinations for enhancing LISA's sensitivity and SNR, including a novel network SNR concept.

Findings

Maximum SNR improvement up to 70% over Michelson at high frequencies.

Network SNR improves sensitivity by 40% to over 100%.

Switching data combinations can yield about 55-60% sensitivity enhancement.

Abstract

It has been shown in the past, that the six Doppler data streams obtained LISA configuration can be combined by appropriately delaying the data streams for cancelling the laser frequency noise. Raw laser noise is several orders of magnitude above the other noises and thus it is essential to bring it down to the level of shot, acceleration noises. A rigorous and systematic formalism using the techniques of computational commutative algebra was developed which generates all the data combinations cancelling the laser frequency noise. The relevant data combinations form a first module of syzygies. In this paper we use this formalism for optimisation of the LISA sensitivity by analysing the noise and signal covariance matrices. The signal covariance matrix, averaged over polarisations and directions, is calculated for binaries whose frequency changes at most adiabatically. We then present the extremal SNR curves for all the data combinations in the module. They correspond to the eigenvectors of the noise and signal covariance matrices. We construct LISA `network' SNR by combining the outputs of the eigenvectors which improves the LISA sensitivity substantially. The maximum SNR curve can yield an improvement upto 70 % over the Michelson, mainly at high frequencies, while the improvement using the network SNR ranges from 40 % to over 100 %. Finally, we describe a simple toy model, in which LISA rotates in a plane. In this analysis, we estimate the improvement in the LISA sensitivity, if one switches from one data combination to another as it rotates. Here the improvement in sensitivity, if one switches optimally over three cyclic data combinations of the eigenvector is about 55 % on an average over the LISA band-width. The corresponding SNR improvement is 60 %, if one maximises over the module.