An algorithm for detection of extreme mass ratio inspirals in LISA data

TL;DR

This paper presents a new stochastic search algorithm that efficiently detects and estimates parameters of extreme mass ratio inspiral signals in LISA data, overcoming computational challenges of traditional methods.

Contribution

The authors introduce a constrained Metropolis-Hastings algorithm tailored for EMRI detection in LISA data, enabling practical analysis of complex signals.

Findings

Successfully detected EMRI signals in simulated LISA data

Achieved accurate parameter estimation of isolated EMRIs

Demonstrated algorithm's effectiveness on Mock LISA Data Challenge datasets

Abstract

The gravitational wave signal from a compact object spiralling toward a massive black hole (MBH) is thought to be one of the most difficult sources to detect in the LISA data stream. Due to the large parameter space of possible signals and many orbital cycles spent in the sensitivity band of LISA, it has been estimated previously that of the order of 10^{35} templates would be required for a fully coherent search with a template grid, which is computationally impossible. Here we describe an algorithm based on a constrained Metropolis-Hastings stochastic search which allows us to find and accurately estimate parameters of isolated EMRI signals buried in Gaussian instrumental noise. We illustrate the effectiveness of the algorithm with results from searches of the Mock LISA Data Challenge round 1B data sets.