Interpretable and Efficient Beamforming-Based Deep Learning for Single Snapshot DOA Estimation

TL;DR

This paper presents an interpretable deep learning approach for single snapshot DOA estimation that combines classical beamforming with deep learning, achieving high accuracy, efficiency, and interpretability in real-time scenarios.

Contribution

It introduces a deep-MPDR network translating MPDR beamforming into deep learning, improving generalization, efficiency, and interpretability over existing methods.

Findings

Outperforms classical and deep learning methods in accuracy and inference time.

Demonstrates robustness on simulated and real-world datasets.

Offers a more interpretable DOA estimation framework.

Abstract

We introduce an interpretable deep learning approach for direction of arrival (DOA) estimation with a single snapshot. Classical subspace-based methods like MUSIC and ESPRIT use spatial smoothing on uniform linear arrays for single snapshot DOA estimation but face drawbacks in reduced array aperture and inapplicability to sparse arrays. Single-snapshot methods such as compressive sensing and iterative adaptation approach (IAA) encounter challenges with high computational costs and slow convergence, hampering real-time use. Recent deep learning DOA methods offer promising accuracy and speed. However, the practical deployment of deep networks is hindered by their black-box nature. To address this, we propose a deep-MPDR network translating minimum power distortionless response (MPDR)-type beamformer into deep learning, enhancing generalization and efficiency. Comprehensive experiments conducted using both simulated and real-world datasets substantiate its dominance in terms of inference time and accuracy in comparison to conventional methods. Moreover, it excels in terms of efficiency, generalizability, and interpretability when contrasted with other deep learning DOA estimation networks.