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 win-dows 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?19 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 interfer-ometer 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 χ2 for detecting gravitational waves and estimating their parameters (direction of their source, etc.) are also presented.%以引力波探测为基础的引力波天文学是一门正在崛起的新兴交叉学科,它是继以电磁辐射为探测手段的传统天文学之后,人类观测宇宙的一个新窗口,对研究宇宙的起源和演化,拓展天文学的研究领域都有极其重要的意义.激光干涉引力波探测器的出现,更开辟了引力波探测的新纪元.引力波数据处理与分析已在世界各地迅速发展起来,为引力波天文学的研究提供了锐利的武器.系统地介绍了引力波数据分析中常用的工具软件,详细讨论了时间-频率分析、复合分析法、脉冲星计时分析法、匹配过滤器、模板、χ2检验、蒙特卡罗模拟等引力波数据分析中使用的基本方法.
展开▼
