Studies in Gravitational Wave Data Analysis

TL;DR

This thesis develops Fourier Transform methods for continuous gravitational wave data analysis, optimizing template matching for detection and parameter estimation in ground-based detectors over various observation times.

Contribution

It introduces generalized Fourier Transform techniques for continuous GW signals, including effects of Earth's motion, and analyzes template requirements for matched filtering.

Findings

Fourier Transform adapted for one-day and multi-year data sets.

Identified symmetry in template parameters affecting match quality.

Quantified the exponential increase in templates needed for higher fitting factors.

Abstract

This thesis is devoted to the investigations of gravitational wave (GW) data analysis from a continuous source e.g. a pulsar, a binary star system. The first Chapter is an introduction to gravitational wave and second Chapter is on the data analysis concept for the detection of GW. In third Chapter we developed the Fourier Transform (FT) of a continuous gravitational wave (CGW) for ground based laser interferometric detectors for the data set of one day observation time incorporating the effects arising due to rotational as well as orbital motion of the earth. The transform is applicable for arbitrary location of detector and source. In Chapter four we have generalized the FT for the data set for (i) one year observation time and (ii) arbitrary observation time. As an application of the transform we considered spin down and N-component signal analysis. In fifth Chapter we have made an analysis of the number of templates required for matched filter analysis as applicable to these sources. We have employed the concept of {\it Fitting Factor (FF)}; treating the source location as the parameters of the signal manifold and have studied the matching of the signal with templates corresponding to different source locations. We have investigated the variation of {\it FF} with source location and have noticed a symmetry in template parameters, $\theta_T$ and $\phi_T$. It has been found that the two different template values in source location, each in $\theta_T$ and $\phi_T$, have same {\it FF}. We have also computed the number of templates required assuming the noise power spectral density $S_n(f)$ to be flat. It is observed that higher {\it FF} requires exponentially increasing large number of templates. Appendix contains the source codes developed for the computation.