Joint MWC and off-grid sparse Bayesian learning based on Augmented Multi-Subarray Dilated Nested Array

Guannan Li

doi:10.1016/j.phycom.2025.102882

2025-10-01

PRE

Guannan Li *

何继爱 (Jiai He)

Youfei Hao

DOI：10.1016/j.phycom.2025.102882

原文链接

原文求助

摘要

En 中文

In the case of excessive mutual coupling (MC) between finite elements in Non-Uniform Linear Arrays(NLAs) for DOA estimation, concerning the problems of high Nyquist sampling data volume, computational complexity and low angle estimation accuracy in current wideband signal processing techniques, this paper proposes a Modulated Wideband Converter-Off-Grid Sparse Bayesian Learning (MWC-OGSBL) joint estimation method based on sub-Nyquist sampling and Augmented Multi-Subarray Dilated Nested Array (AMDNA). The method first maintains the Degrees of Freedom (DOF) of the Multi-Subarray Dilated Nested Array (MDNA) unchanged, reduces the MC by migrating the dense elements, and performs sparse sampling on the received signal using AMDNA and MWC to reduce the data volume. Subsequently, we consider the noise variance to be a component of the unknown signal of interest in order to automatically determine its value using Bayesian inference. Finally, the locations of grid points are handled as modifiable parameters, iteratively refined using the Expectation Maximisation (EM) algorithm to eliminate modeling errors caused by the gaps between the grid points and achieve accurate angle estimation. The experimental results show that compared with other NLA and wideband signal estimation algorithms, the proposed method can improve positioning accuracy, effectively reduce experimental running time, and have better convergence in environments with fewer snapshots and low signal-to-noise ratio (SNR).

Keyword:

Augmented Multi-Subarray Dilated Nested

Array (AMDNA)

Mutual coupling

Modulated Wideband Converter (MWC)

Off-grid sparse Bayesian learning

DOA estimation