Adaptive PCA-Based Spectral Estimation Method for MST Radar Signal Processing

Subspace methods founded on Eigen-value decomposition have been utilized to retrieve relevant data from extensive datasets. This work suggests the least variance spectral estimation technique named as Proposed Algorithm (PALG) for adaptive principle component based spectrum computation. Using PALG, we analyze the signal received by the MST (Mesosphere-Stratosphere-Troposphere) radar placed at NARL (National Atmospheric Research Laboratory) Gadanki in this proposed work. Additionally, we tested the PALG approach with simulated signals, like broadband signals, at various noise levels (a). The PALG performed better in identifying the number of frequencies in the simulated signal, especially in the case of a noise-corrupted signal. Finally, the suggested algorithm is applied to the MST radar data in order to estimate the Doppler frequency spectrum, which is then utilized to determine the wind, zonal, and meridional velocities. The PALG performs well at higher altitudes when compared to current techniques, and ground truth GPS data was used to validate the MST radar results.