Parametric Method Based PSD Estimation using Gaussian Window

Non-parametric methods of Spectrum Estimation Such as Periodogram, Modified Periodogram, Welch, Bartlett and Blackman-Tukey are generally used but are not always efficient in finding out the power spectral densities. These methods lacks due to their limitations like windowing of the autocorrelation sequence, spectral leakage, poor resolution and incapability to include the available information about the process into the estimation procedure. In these cases parametric approach of spectrum estimation outperforms the nonparametric ones and helps in producing high resolution spectral estimates. Also, parametric methods are more accurate and computationally more efficient than the non-parametric methods. As Gaussian window is an important window in digital signal processing applications. So in the proposed work, Burg’s, Yule-Walker, Covariance and Modified Covariance methods are used to estimate the PSD of the Gaussian window function. The simulation is done in MATLAB and compared to analyze the power in different spectral components using different methods.