On the framework of stochastic gravitational wave background(SGWB) by compact binary systems, we studied the strain spectra of SGWB produced by cosmological cataclysmic variables(CV). For this we reviewed the empirical properties of CVs by using newly published CV catalogue and calculated the cosmological densities of CVs considering the galaxy luminosity function and cosmic stellar birth rate function. Assuming the secular evolution of CVs, we calculated the time scale of CV gravitational wave(GW) radiation and derived formulae for the strain spectra of SGWB by cosmological CVs.

On the formulation frameworks of linearly perturbed spacetime and weak gravitational lensing(WGL) we studied the statistical properties of a bundle of light rays propagating through stochastic gravitational wave background(SGWB). For this we considered the SGWB as tensor perturbations of linearly perturbed Friedmann spacetime. Using the solution of null geodesic deviation equation(NGDE) we related the convergence, shear and rotation deformation spectra of WGL with the strain spectra of SGWB. Adopting the astrophysical and cosmological SGWB strain spectra which were already known we investigated the approximated spectral forms of convergence, shear and rotation of WGL.

Many gravitational wave detectors are now being built or under operation throughout the world. In particular, LIGO has taken scientific data several times, although current sensitivity is not sufficient to detect the weak signals routinely. However, the sensitivities have been improving steadily over past years so that the real detection will take place in the near future. Data analysis is another important area in detecting the gravitational wave signal. We have carried out the basic research in order to implement data analysis software in Korea@home environment. We first studied the LIGO Science Collaboration Algorithm Library(LAL) software package, and extracted the module that can generate the virtual data of gravitational wave detector. Since burst sources such as merging binaries of neutron stars and black holes are likely to be detected first, we have concentrated on the simulation of such signals. This module can generate pure gravitational wave forms, noise suitable for LIGO, and combination of the signal and noise. In order to detect the gravitational signal embedded in the noisy data, we have written a simple program that employs 'matched filtering' method which is very effective in detecting the signal with known waveform. We found that this method works extremely well.