Faster implementation of the hierarchical search algorithm for detection of gravitational waves from inspiraling compact binaries

Sengupta, Anand S. ; Dhurandhar, Sanjeev ; Lazzarini, Albert (2003) Faster implementation of the hierarchical search algorithm for detection of gravitational waves from inspiraling compact binaries Physical Review D - Particles, Fields, Gravitation and Cosmology, 67 (8). 082004_1-082004_14. ISSN 1550-7998

Full text not available from this repository.

Official URL: http://prd.aps.org/abstract/PRD/v67/i8/e082004

Related URL: http://dx.doi.org/10.1103/PhysRevD.67.082004

Abstract

The first scientific runs of kilometer scale laser interferometric detectors such as LIGO are under way. Data from these detectors will be used to look for signatures of gravitational waves from astrophysical objects such as inspiraling neutron-star-black-hole binaries using matched filtering. The computational resources required for online flat-search implementation of the matched filtering are large if searches are carried out for a small total mass. A flat search is implemented by constructing a single discrete grid of densely populated template waveforms spanning the dynamical parameters— masses, spins — which are correlated with the interferometer data. The correlations over the kinematical parameters can be maximized a prioriwithout constructing a template bank over them. Mohanty and Dhurandhar showed that a significant reduction in computational resources can be accomplished by using a hierarchy of such template banks where candidate events triggered by a sparsely populated grid are followed up by the regular, dense flat-search grid. The estimated speedup in this method was a factor ~25 over the flat search. In this paper we report an improved implementation of the hierarchical search, wherein we extend the domain of hierarchy to an extra dimension — namely, the time of arrival of the signal in the bandwidth of the interferometer. This is accomplished by lowering the Nyquist sampling rate of the signal in the trigger stage. We show that this leads to further improvement in the efficiency of data analysis and speeds up the online computation by a factor of ~65-70 over the flat search. We also take into account and discuss issues related to template placement, trigger thresholds, and other peculiar problems that do not arise in earlier implementation schemes of the hierarchical search. We present simulation results for 2PN waveforms embedded in the noise expected for initial LIGO detectors.

Item Type:Article
Source:Copyright of this article belongs to The American Physical Society.
ID Code:87740
Deposited On:21 Mar 2012 09:58
Last Modified:21 Mar 2012 09:58

Repository Staff Only: item control page