Dual apodization in the detection of a sinusoidal signal under the effect of interference

Presence of a sinusoidal interference signal can inhibit the detection of another sinusoidal signal because of its sidelobes. By the use of conventional window functions, sidelobes can be suppressed with the price of frequency resolution loss. Spatially variant apodization (SVA), which uses an infinite number of window functions, can reduce spectral leakage by preserving the frequency resolution of the unwindowed signal. Dual apodization (DA) is a simple version of SVA, in which only two window funtions are used. In this work SVA and DA methods are applied to the problem of detection of a sinusoidal signal under sinusoidal interference. The probability density function for DA output is presented and the theoretical threshold calculation is given. Detection performance comparisons between DA, SVA and conventional windows are given. By the help of the simulations, it is shown that SVA and DA methods outperform the conventional windows.