Algorithm development for LISA

LISA is a space mission, adopted by ESA in 2024, to build and fly an interferometer to detect mHz-frequency gravitational waves.

LISA comprises three spacecraft, flying 2.5 million km apart in a triangular formation, with laser interferometers along each arm.

It will detect gravitational waves produced by the assembly and merger of supermassive black holes at high redshifts, and from the population of close compact binaries in the Milky Way. We propose a two-part project to increase the scientific return from the data that LISA will provide.

LISA will observe gravitational waves from the whole sky, and the principal challenge for the data analysis is separating faint but scientifically interesting signals from the background noise produced by unresolved Galactic sources, principally close binaries containing two white dwarfs.

The first aspect of the project is to produce a LISA model galaxy, containing a synthetic sample of compact binaries, computed using the most up-to-date stellar and binary physics and distributed in a galactic model that builds on results from the Gaia space mission and ground-based surveys.

The result will be an authoritative prediction of the Galactic population of LISA sources and resulting gravitational wave background, but also a quantification of the uncertainty.

This will provide a realistic simulated noise background for the analysis pipline and allow us to quantify how many of the binaries can be recovered from the noise.The second aspect of the project is to develop a testbench for the rapid prototyping and assessment of algorithms to analyse the LISA data.

Current algorithm development within the LISA consortium focusses on a Markov Chain Monte Carlo  method implemented within a block like structure which identifies and subtracts individual blocks of combined astrophysical sources.

This is not necessarily the optimal choice -- for example, it neglects to a large extent the correlations between individual sources. We will focus on developing algorithms for a global fit, based on our experience with the Gaia data analysis.

This will permit the creation of unbiased parameter fits for individual sources, characterisation of more, fainter binaries as well as variable sources such as EMRIs and transient sources from mergers of supermassive black hole binaries.

It will also allow us to assess the covariance between the signals extracted from different LISA sources.Finally, there is a strong synergy between the tools produced to characterise the LISA-visible compact binary population of the inner Milky Way and those required for prototyping the upcoming GaiaNIR mission.

GaiaNIR is a proposal for an  all-sky astrometric survey at infra-red wavelengths as a follow-on to the extremely successful Gaia mission. GaiaNIR will extend Gaia´s science into the dust-obscured bulge and galactic centre.

We will use the LISA model galaxy to predict the populations of faint stars -- including but not limited to binaries -- in the central Milky Way.

This is crucial to understand the impact of crowding on the design decisions around data downlink and scan rate to be made for GaiaNIR, and will also allow us to evaluate and plan for science cases where targets are observed both by LISA and GaiaNIR.