Constraint likelihood analysis for a network of gravitational wave detectors

We propose a coherent method for detection and reconstruction of gravitational wave signals with a network of interferometric detectors. The method is derived by using the likelihood ratio functional for unknown signal waveforms. In the likelihood analysis, the global maximum of the likelihood ratio over the space of waveforms is used as the detection statistic. We identify a problem with this approach. In the case of an aligned pair of detectors, the detection statistic depends on the cross correlation between the detectors as expected, but this dependence disappears even for infinitesimally small misalignments. We solve the problem by applying constraints on the likelihood functional and obtain a new class of statistics. The resulting method can be applied to data from a network consisting of any number of detectors with arbitrary detector orientations. The method allows us reconstruction of the source coordinates and the waveforms of two polarization components of a gravitational wave. We study the performance of the method with numerical simulations and find the reconstruction of the source coordinates to be more accurate than in the standard likelihood method.