In this study, we apply the Analytical method of Blind Separation (ABS) of the cosmic microwave background (CMB) from foregrounds to estimate the CMB temperature power spectrum from multi-frequency microwave maps. We test the robustness of the ABS estimator and assess the accuracy of the power spectrum recovery by using realistic simulations based on the seven-frequency Planck data, including various frequency-dependent and spatially-varying foreground components (synchrotron, free-free, thermal dust and anomalous microwave emission), as well as an uncorrelated Gaussian-distributed instrumental noise. Considering no prior information about the foregrounds, the ABS estimator can analytically recover the CMB power spectrum over almost all scales with less than $0.5%$ error for maps where the Galactic plane region ($|b|<10^{circ}$) is masked out. To further test the flexibility and effectiveness of the ABS approach in a variety of situations, we apply the ABS to the simulated Planck maps in three cases: (1) without any mask, (2) imposing a two-times-stronger synchrotron emission and (3) including only the Galactic plane region ($|b|<10^{circ}$) in the analysis. In such extreme cases, the ABS approach can still provide an unbiased estimate of band powers at the level of 1 $murm{K}^2$ on average over all $ell$ range, and the recovered powers are consistent with the input values within 1-$sigma$ for most $ell$ bins.
تحميل البحث