MUSIC Algorithm with Unknown Source Number Estimation Using AIC Criterion

This article introduces the MUSIC (Multiple Signal Classification) algorithm, a signal processing technique used for estimating unknown source numbers to enhance signal analysis and understanding. To improve estimation accuracy, we integrate the MUSIC algorithm with the AIC (Akaike Information Criterion) approach. The implementation involves first applying AIC criterion to estimate the number of signal sources, then utilizing this estimate within the MUSIC algorithm framework for more precise results. The algorithm workflow typically includes: computing the sample covariance matrix from received signals, performing eigenvalue decomposition to identify signal and noise subspaces, using AIC to determine the optimal source number by minimizing information loss, and finally applying MUSIC spatial spectrum estimation. This technique finds broad applications across various domains including communications systems, radar signal processing, and sonar array processing, particularly in direction-of-arrival estimation scenarios where the number of incoming signals is unknown.