Coherent Multi-detector search of GWaves
Archana Pai
Global network of Gravitational Wave (GW) interferometric antennas (LIGO-3no, VIRGO, KAGRA) would be up and running by 2015. With the projected sensitivities such antennas would observe GW events few per month from the compact binary systems with Neutron stars and Black holes. The GW observations would give information about the stellar mergers, centers of the Galaxies etc which otherwise is inaccessible using the electromagnetic probe. Due to the nature of the GW antennas as well as Einsteinian GR, a single antenna is not sufficient to characterise all the source parameters such as location and polarisation etc. Observation with Multi-detector antenna via the technique of Aperture synthesis which combines the delays and constructs an effective multi-detector antenna would give this information. However, this technique involves constructing phase matching templates in 5 (masses, source location and time) dimentional parameter space for non-spinning systems which further increases when each star has arbitrary spin.
Binary neutron stars are prominent sources of GWs. We detail the typical online computational requirement for a 3-detector LIGO-VIRGO network to detect such a system using all sky-search and searching only in the mass window of (1 − 10)M. The number of filters to extract the 3.5 post-Newtonian phase (in masses) with the phase-match between signal and template is 97 per cent are 105. All sky filters in time-delays are 106. This operation needs to be performed on each mass filter. Thus, the total filters gets multiplied and the online computational cost becomes closed to 20 TFlops on CPU (A. Pai, S. Dhurandhar, and S. Bose, Phys. Rev. D 64, 042004 (2001)). This number goes further up by a factor of 40 for the advanced detectors with broader bandwidth. Hence, till now the full sky-search with multi-detector antenna has not been implemented in the LIGO-VIRGO pipeline and is an open challenge in the community. Here, I propose to carry out two stage heirarchical scheme which would look for events on a coarser sky grid followed by a finer grid as a follow up. The preliminary calculations show that coarse grid would reduce the sky-patch by a factor of 5 and hence reduce the cost to 4 TFlop. Such a heirarchical scheme has not been demonstrated in the literature till now. The targetted search has been demonstrated recently (I. W. Harry and S. Fairhurst, Phys. Rev. D 83). This facility would allow to test such a heirarchical search on the off-line HPC facility we propose.Broader perspective – IISER-TVM is part of the LIGO Scientific Collaboration through Indian Intiative in Gravitational Wave Observatory(IndIGO) consortium. Members of the consortium have submitted a joint proposal of LIGO-India where one of three LIGO detectors are proposed to be stationed in India. If the LIGO-India proposal gets approval, the IISER-TVM would be able to access the LIGO-India data and the off-line test could be run into an online search.