Classical and Modern Spectral Estimation Methods: Two Major Categories of Spectral Estimation Techniques
Power spectrum estimation has broad applications across various disciplines and application domains, receiving significant attention. In the "Modern Signal Processing" course, we studied two primary spectral estimation methods: classical spectral estimation and modern spectral estimation. Classical spectral estimation, based on Fourier transform, offers high computational efficiency but suffers from low spectral resolution and severe side lobe leakage, making it suitable for long sequences. To overcome these limitations, researchers developed modern spectral estimation methods based on parametric models of stochastic processes, including maximum likelihood estimation, maximum entropy method, AR model approach, and predictive filtering techniques. Modern spectral estimation provides higher accuracy for short sequences, complementing classical methods. After thorough study, I selected modern spectral estimation for implementation, particularly focusing on AR model parameter estimation using techniques like the Yule-Walker equations or Burg's algorithm for efficient computation.