Florian Aubin – Phd’s defense – oct. 2, 2020

The first detection of gravitational waves from compact binary coalescences occured duringthe O1 and O2 observing runs between 2015 and 2017, using the LIGO and Virgo detectors.On August 17, 2017, the observation of gravitational waves from the coalescence of twoneutron stars triggered a successful multi-messager follow-up of the source. The detectionof electromagnetic counterparts leads to major results in astrophysics, fundamental physics,and cosmology. After a detector upgrade period, the O3 run took place between April 1,2019 and March 27, 2020. In order to optimize the chances of finding counterparts, thecollaboration use low latency searches which, since O3, automatically send public alerts incase of significant candidates. The work presented in this manuscript focuses on one of thesepipelines, developed within the Virgo collaboration, called MBTA. One of the challenges forthese analyses is to deal in real time with non-Gaussian and non-stationary noise producedby a network of detectors with heterogeneous sensitivities. In prepartion for O3, it wasnecessary to identify the problems that emerged during the previous run. Thanks to thisexperience, the noise rejection technics used by MBTA have been reworked and new, moreprogressive tools were developed. During this new observing run, the MBTA pipeline wasrun and improved, in order to process the data in the best possible way. Finally, MBTAwas able to contribute, along with the other low-latency analyses of the collaboration, toidentifying online 56 candidates from possible astrophysical origin.