I have been involved in gravitational wave data analysis for searching a stochastic gravitational wave background (SGWB) in LIGO data from the outset. Since then, I have led, with Bruce Allen and Joe Romano, the task group for the Monte Carlo simulations of these signals to ascertain (with ``software injections'') the efficiency and biases of the detection pipeline and to estimate (with ``hardware'' injections) the effect of calibration errors on the detection statistic [<a href=''http://www.iop.org/EJ/abstract/0264-9381/20/17/311/''>Bose:2003nb</a>]. To a good extent, the observational results for a cosmological (and, therefore, isotropic) SGWB presented in the LIGO stochastic papers [Refs. <a href=''http://www.ligo.org/results/''>LIGO Papers</a>] depended on the successful execution of the above simulations and searches. A whole suite of routines now exist to simulate waveforms as well as to inject and search them in both real and simulated detector noise. I am now working with Ballmer et al. to extend this suite to work on directed searches of astrophysical SGWBs. Since hardware injections allow us to perform an ``end-to-end'' test of the full detection pipeline, the results from such injections are sought almost as soon as a science (data-taking) run begins. The simulation / search infrastructure that we have set up allows such quick turnarounds, while battling the setting up of data and calibration caches, and updating search routines to account for changes (e.g., analyzing different channels for gravitational wave content) occurring between science runs.
Signals from a host of astrophysical and cosmological sources, spanning a wide gamut ranging from hydrodynamic instabilities in neutron stars (such as r-modes) to particle production in the early universe, might show up in the data of the existing and upcoming detectors. Signals from a subset of these sources are expected to appear in these detectors as stochastic gravitational wave backgrounds (SGWBs). The detection of these backgrounds will help us in characterizing their sources. It will also be required by some detectors, such as the proposed space-based detector LISA, so that they can detect other gravitational wave signals. Recently, I formulated the problem of constructing a bank of search templates that discretely span the parameter space of a generic SGWB [<a href=''http://arxiv.org/abs/astro-ph/0504048''>Bose:2005fm</a>]. We applied it to the specific case of a class of cosmological SGWBs, known as the broken power-law models and derived how the template density varies in their three-dimensional parameter space. We also showed that for the LIGO 4km detector pair, with LIGO-I sensitivities, a few hundred templates will suffice to detect such a background while incurring an SNR loss of no more than 3%. This work is now being extended (in collaboration with Lazzarini, Dhurandhar, Gusev, et al.) to directed/localized searches for SGWBs.