Direction Finding Using Higher Order Statistics Without Redundancy

In the last decade, $2q$ -MUSIC, an extension of MUSIC algorithm to an arbitrary even order $2q$ , has been proposed to process direction-finding problems. Despite of its better performance compared with MUSIC, it suffers from the high computational complexity, thus limiting its practical application. This letter proposes a method called Non-Redundant- $2q$ -MUSIC using $2q$ th order statistics for uniform linear arrays and uniform rectangular arrays. The proposed method lowers the computational complexity effectively by removing the redundancy of the virtual array and the $2q$ th-order cumulant matrix. And root-MUSIC can be applied to avoid calculating the pseudo-spectrum. It is illustrated in both theoretical proof and computer simulations that the proposed method performs properly and effectively.