How do you find very small gravitational-wave signals in the ocean of background noise that’s always present in our LIGO detectors?
We use a technique called matched filtering, originally developed in the early 40s to optimize detections in radars and sonars. The idea is that your search is a lot more sensitive if you know what you’re looking for: have you ever noticed that when you are in a noisy room, you can hear much easier somebody speaking your name rather than any other name? That’s because your brain is matching a known pattern (your name) to the background noise.
Matched filtering is exactly the same. We start with a template for the gravitational-wave signal that we are looking for, let’s say from the collision of two black holes like those in the first event we detected (GW150914). We can then compute the correlation of this template with the detector’s output, at all times. Even if the signal is buried in noise, the value of this correlation, the signal-to-noise ratio, will spike when there is something in the signal that matches the template, as shown in the animation.
The problem is, one template isn’t enough: black holes and neutron stars come in all shapes and flavors: masses, sky localization, orientation and many more parameters. We need to cover the entire parameter space with a tessellation of many many templates. In the last LIGO and Virgo science run the compact binary coalescence (CBC) searches used several millions of templates! Each one continuously correlated to the detectors’ output thousands of times per second. That’s a lot of computing power!
I post every week simple explanations of the science and engineering behind the discovery and detection of gravitational waves. Follow and connect!
Leave a Reply