The collaborating gravitational-wave detectors LIGO, Virgo, and KAGRA have successfully used an AI method for the first time to extract signals from partially malfunctioning detectors.
In a new article in Physical Review Letters, the international LVK collaboration analyzed observations of two gravitational wave events, in September 2024 and February 2025. These were attributed to colliding black holes.
The first involved black holes with masses of 9 and 7 solar masses, located approximately 19 million light-years from Earth, and the second involved black holes with masses of 35 and 30 solar masses, located about 11 million light-years away. The massive distortion of spacetime itself generates waves in space that can still be detected on Earth.
Nikhef is part of the LVK collaboration and is closely involved with the Virgo detector in Pisa, Italy. The first gravitational wave was detected in 2015.
The precision of the two observations would have been impossible until recently, the LVK researchers write, because the LIGO detector in Hanford, Washington, was not properly calibrated in either case. The new technique makes it possible to use the gravitational waves themselves to identify corrections.
To ensure no signals are lost, even if the detectors are not properly calibrated, a method called Astrophysical Calibration has been developed. This involves comparing signals across different detectors. The observations from properly functioning detectors provide a reference for the received signal.
The signal from a detector that is functioning less well is compared to this, after which a correction for the poor calibration can be calculated. This allows the actual signal to be calculated after all and used in the analysis of all observations.
In a sense, the technique resembles a technology in the music industry called autotune, which corrects less-than-perfect vocal parts during live performances.
A version of “autotune” for gravitational wave detectors is now also being researched by the Gravitational Wave group at Nikhef, says program leader Chris Van Den Broeck, who is also a professor at Utrecht University.
Van Den Broeck: “This is an exciting first example of how our instruments can be calibrated using the gravitational waves themselves. That will certainly play a major role in the Einstein Telescope.”
A first application in this regard was an analysis of the various configurations of the future Einstein Telescope. ET is a significantly improved underground gravitational wave detector with detector arms approximately 10 kilometers long. Two variants are currently under study: a triangle consisting of three V-shaped detectors, or two L-shaped detectors at different locations.
Preliminary studies have shown that in the triangle variant, noise and other disturbances (so-called glitches) can be best suppressed through the clever combination of signals, even if one of the arms is temporarily not functioning perfectly.
