High-Resolution Multiple-Source 2D DOA Estimation Using Convolutional Neural Network With Robustness to Array Imperfections

In this paper, we present an efficient convolutional neural network (CNN)-based model to estimate both elevation and azimuth arrival angles of multiple sources with high resolution (small source angular separation). The sources are considered closely spaced in both elevation and azimuth up to 0.5&am...

Full description

Saved in:
Bibliographic Details
Main Authors: Tarek Sallam, Qun Wang, Ahmed M. Attiya
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10713319/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we present an efficient convolutional neural network (CNN)-based model to estimate both elevation and azimuth arrival angles of multiple sources with high resolution (small source angular separation). The sources are considered closely spaced in both elevation and azimuth up to 0.5&#x00B0; resolution in both directions. Ten sources are supposed to be received by an <inline-formula> <tex-math notation="LaTeX">$8\times 8$ </tex-math></inline-formula> antenna array. Compared to multiple signal classification (MUSIC) algorithm and radial basis function neural network (RBFNN), numerical results indicate that the CNN model exhibits notable precision in two-dimensional (2D) direction of arrival (DOA) estimation of multiple-source with high resolution, along with less than one-tenth second execution time. Finally, it maintains its superior performance even in the presence of array imperfections and significantly reduced SNR.
ISSN:2169-3536