Revelation and Elimination of Redundant Spectral Searches in MUSIC

In this letter, the spatial-spectral peak search process of the classical multiple signal classification (MUSIC) algorithm is revealed to inherently involve 50% computational redundancy through an analysis of the MUSIC spectrum function, i.e., only half-spatial-domain spectral searches are essential to construct the complete full-spatial-domain spectrum. Thus, the spectral-search framework of MUSIC is refined by the proposed half spectral search MUSIC (HSS-MUSIC), which eliminates the redundant searches and preserves identical outputs of MUSIC. Derivations and simulations demonstrate that compared with MUSIC, the proposed HSS-MUSIC does not exhibit any deterioration in both array structure requirement and DOA estimation performance, but achieves an approximately 50% reduction in computational load.