A morphology-independent data analysis method for detecting and characterizing gravitational wave echoes

The ability to directly detect gravitational waves has enabled us to empirically probe the nature of ultra-compact relativistic objects. Several alternatives to the black holes of classical general relativity have been proposed which do not have a horizon, in which case a newly formed object (e.g. as a result of binary merger) may emit echoes: bursts of gravitational radiation with varying amplitude and duration, but arriving at regular time intervals. Unlike in previous template-based approaches, we present a morphology-independent search method to find echoes in the data from gravitational wave detectors, based on a decomposition of the signal in terms of generalized wavelets consisting of multiple sine-Gaussians. The ability of the method to discriminate between echoes and instrumental noise is assessed by inserting into the noise two different signals: a train of sine-Gaussians, and an echoing signal from an extreme mass-ratio inspiral of a particle into a Schwarzschild vacuum spacetime, with reflective boundary conditions close to the horizon. We find that both types of signals are detectable for plausible signal-to-noise ratios in existing detectors and their near-future upgrades. Finally, we show how the algorithm can provide a characterization of the echoes in terms of the time between successive bursts, and damping and widening from one echo to the next.