Multipath DOA Estimation via Vandermonde Constrained Block Term Decomposition
Li-Na Luo, Lei Wang, Deng-Chao Song, Xiao‐Feng Gong · 2024
Direction Of Arrival (DOA) estimation in multipath environment is a challenging problem in array signal processing. In this paper, we consider the two-dimensional Uniform Planar Array (UPA) and construct the received array data into a tensor following Block Term Decomposition (BTD) model to characterize the multipath structure. To exploit the special structure of the steering vectors, we proposed a DOA estimation algorithm based on BTD algorithm with Vandermonde constraint on factor matrices. In our algorithm, the existing BTD algorithms are used on the preprocessed data by hankelization and dimension merging of the original data, aiming to obtain the source signal matrix. The proposed Vandermonde Matrix Decomposition (VMD) algorithm is further used to obtain the two direction matrices. Besides, we analysis the working condition of our algorithm. Theoretical and experimental results demonstrate the proposed Vandermonde constrained BTD algorithm achieves high accuracy in multipath environment and have more relaxed identification conditions.