Performance of a Chirplet-based analysis for gravitational waves from binary black hole mergers
The gravitational wave (GW) signature of a binary black hole (BBH) coalescence is characterized by rapid frequency evolution in the late inspiral and merger phases. For a system with total mass larger than 100 M_sun, ground based GW detectors are sensitive to the merger phase, and the in-band whitened waveform is a short-duration transient lasting about 10-30 ms. For a symmetric mass system with total mass between 10 and 100 M_sun, the detector is sensitive instead to the inspiral phase and the in-band signal has a longer duration, between 30 ms - 3 s. Omega is a search algorithm for GW bursts that, with the assumption of locally stationary frequency evolution, uses sine-Gaussian wavelets as a template bank to decompose interferometer strain data. The local stationarity of sine-Gaussians induces a performance loss for the detection of lower mass BBH signatures, due to the mismatch between template and signal. We present the performance of a modified version of the Omega algorithm, Chirplet Omega, which allows a linear variation of frequency, to target BBH coalescences. The use of Chirplet-like templates enhances the measured signal-to-noise ratio due to less mismatch between template and data, and increases the detectability of lower mass BBH coalescences. We present the results of a performance study of Chirplet Omega in colored Gaussian noise at initial LIGO sensitivity.