The Data Analysis in Gravitational Wave Detection

wang xiaoge,ERIC Lebigot,dou zhihui,cao junwei,wang yunyong,zhang fan,cai yongzhi,li muzi,zhu zonghong,qian jin,yin cong,wang jianbo,zhao wen,zhang yang,DAVID Blair,JU Li,ZHAO Chun-Nong,WEN Lin-Qing

DOI: https://doi.org/10.3969/j.issn.1000-8349.2016.01.04

2016-01-01

Abstract:Gravitational waves are an important prediction of general relativity, made in 1916 by Einstein. Yet, the deformation that they imprint on space has never been observed directly. The first such observation should soon be made with pulsar timing arrays or large laser interferometers, which are emphasized in this paper. The advanced interferometers will be fully operational in a few years. We review many important data analysis techniques used for these instruments.In the same way as the exploration of the electromagnetic spectrum opened new windows on the universe（radio waves, infrared, visible spectrum, ultraviolet, X-rays, gamma rays, etc.）, the ability to sense gravitational waves will open a new kind of astronomy. In fact, the shape of a gravitational wave encodes information about its physical origin, be it the coalescence of two very dense objects（neutron stars and/or black holes）, a supernova explosion, or other phenomena.Gravitational waves are very weak: the movement that the largest laser interferometers（LIGO and Virgo） are designed to observe has an amplitude of only about 10; m, which is measured as the relative variation（strain） of the length of arms that measure 3 to 4 kilometers. Achieving this requires a host of physical measurements that flow from sensors to computers. These measurements are analyzed, displayed and stored through a complex hardware and software system that are presented in this paper.Such a high precision measurement requires extremely sensitive experimental and data analysis techniques. The signal that continuously comes from a gravitational wave interferometer can be observed either as a function of time, or in the time-frequency domain, since this representation is adapted to the oscillatory nature of the gravitational waves. Different data analysis techniques are appropriate for each representation. For the time domain, we thus discuss matched filtering techniques, which can find a small signal of known form buried in noise. The time-frequency domain is presented in conjunction with techniques used for detecting bursts of gravitational waves, and unwanted glitches in the interferometer signal.Since multiple large gravitational wave interferometers are being made available, it is natural to take advantage of their conjoint measurements: a gravitational wave should be seen on all running detectors（with an amplitude that depends on their orientation）, whereas glitches in the interferometer signal should generally not happen at the same time in the same way. This forms the basis of the coincidence and coherent data analysis methods presented here. Other important computing techniques such as Monte-Carlo calculations and χ; for detecting gravitational waves and estimating their parameters（direction of their source, etc.） are also presented.