Efficient gradient‐based dictionary optimisation method for time‐varying array incorporating with structure characteristic

Impact of antenna directivity on compressed sensing-based DOA estimation employing UCA antenna

“…where | • |, ∥ • ∥, and (•) H denote the absolute value of a scalar, Euclidean norm of a vector, and Hermitian transpose of a matrix, respectively; a 𝑖 and a 𝑗 are the 𝑖-th and 𝑗-th column vectors of A, respectively; 𝑅 𝑖, 𝑗 corresponds to the cross-correlation between their column vectors. However, the mutual coherence is not necessarily suitable for investigating the behavior of the CS-based reconstruction, because it is defined as the worst case of cross-correlation between the two column vectors [8]. Therefore, in the following section, we evaluate the cross-correlation itself, which constitutes the mutual coherence.…”

“…The CS-based method exploits sparsity of incident waves in the angular space domain, and is an attractive approach because the high-resolution estimation can be achieved by a single snapshot [3,4,5,6]. According to the CS theory, the property of a sensing matrix generally determines the reconstruction accuracy of sparse signals [7,8], which depends on the array antenna configuration in the DOA estimation. As for the antenna element arrangement, the uniform circular array (UCA) antenna [3] is effective for wide-range DOA estimation, compared with the uniform linear array (ULA) antenna [5,6].…”