High-resolution direction-finding algorithm using Qth−order cumulants tensor: Q−TMUSIC.

To resolve more sources than sensors, the 2 q − MUSIC algorithm has been developed recently by rearranging the 2 q th − order circular cumulants tensor into a suitable size matrix. However, in 2 q − MUSIC, the angular resolution and estimation accuracy have not attained their optimum values due to the loss of differential information by matricizing the cumulants tensor. This paper proposes a subspace-based MUSIC-like algorithm, namely Q th − order cumulants-based tensor MUSIC algorithm for the direction of arrival (DOA) estimation of statistically independent non-Gaussian sources by exploiting the Q th − order cumulants tensors, Q ∈ { 3 , 4 , 5 , ⋯ } , directly using multilinear algebra tools. The proposed algorithm offers an improved DOA estimation performance in terms of angular resolution and estimation accuracy compared to their equivalent matrix-based Q −MUSIC. Moreover, the computational complexity of the proposed Q -TMUSIC is lesser than that of their equivalent matrix-based Q −MUSIC for Q > 4. However, the proposed Q −TMUSIC can identify atmost (M − 1) sources using an array of M −sensors than that of (M ⌊ Q 2 ⌋ − 1) with their equivalent matrix based Q −MUSIC. [ABSTRACT FROM AUTHOR]