Hierarchical search for compact binary coalescences in the Advanced LIGO's first two observing runs
Detection of many compact binary coalescences (CBCs) is one of the primary goals of the present and future ground-based gravitational-wave (GW) detectors. While increasing the detectors' sensitivities will be crucial in achieving this, efficient data analysis strategies can play a vital role. With given computational power in hand, efficient data analysis techniques can expand the size and dimensionality of the parameter space to search for a variety of GW sources. Matched filtering based analyses that depend on modeled signals to produce adequate signal-to-noise ratios for signal detection may miss them if the parameter space is too restrained. Specifically, the CBC search is currently limited to non-precessing binaries only, where the spins of the components are either aligned or anti-aligned to the orbital angular momentum. A hierarchical search for CBCs is thus well motivated. The first stage of this search is performed by matched filtering coarsely sampled data with a coarse template bank to look for candidate events. These candidates are then followed up for a finer search around the vicinity of an event's parameter space. Performing such a search leads to enormous savings in computational cost. Here we report the first successful implementation of the hierarchical search as a PyCBC-based production pipeline to perform a complete analysis of LIGO observing runs. With this, we analyze Advanced LIGO's first and second observing run data. We recover all the events detected by the PyCBC (flat) search in the first GW catalog, GWTC-1, published by the LIGO-Virgo collaboration, with nearly the same significance using a scaled background. In the analysis, we get an impressive factor of 20 reduction in computation compared to the flat search. With a standard injection study, we show that the sensitivity of the hierarchical search remains comparable to the flat search within the error bars.