Abstract:This paper proposes an innovative global solution which is a pioneering work applying automated machine learning algorithms to remarkable precision sparse underwater direction-of-arrival (DOA) estimation that views the subaquatic sparse-sampling DOA estimation problem as a classification prediction task. The proposed solution, termed automated multi-layer perceptron discriminative neural network (AutoMPDNN), is built upon a Bayesian optimization framework. AutoMPDNN transforms sparsely sampled time-domain signals into the complex domain, preserving essential components in a one-source single-snapshot scenario. Leveraging Bayesian optimization principles, the algorithm embeds necessary hyperparameters into the loss function, effectively defining it as a maximum likelihood problem using the upper confidence bound function and incorporating sparse signal features. We also explore the model space architecture and introduce variants of AutoMPDNN, denoted as AutoMPDNNs_ln (n = 2,3,4). Through a series of plane wave simulation experiments, it is demonstrated that AutoMPDNN achieves the highest prediction performance for one-source single-snapshot scenarios compared to classical DOA estimation algorithms that incorporate sparse representation approaches, as well as contemporary deep learning DOA methods under varying conditions.

What problem does this paper attempt to address?

The paper primarily addresses the problem of underwater Direction-of-Arrival (DOA) estimation, especially in scenarios involving one-source single-snapshot and sparse sampling conditions for high-precision DOA estimation. Traditional DOA estimation methods often perform poorly under complex interference environments, rapidly moving or non-static targets, and spatial sampling limitations. To solve the aforementioned issues, the researchers propose an innovative global solution, namely the Automated Multi-Layer Perceptron Discriminative Neural Network (AutoMPDNN) based on a Bayesian optimization framework. This method treats the DOA estimation problem with sparse underwater sampling as a classification prediction task and utilizes machine learning algorithms to achieve high-precision DOA estimation. Specifically, the main contributions of AutoMPDNN include: 1. **Signal Reconstruction and Feature Extraction**: Reconstructing the sparsely sampled time-domain signals through signal processing techniques to extract key feature information, adapting to the one-source single-snapshot scenario. 2. **Design of Multi-Layer Perceptron Discriminative Neural Network (MPDNN)**: Constructing a lightweight multi-layer perceptron network model based on supervised learning for end-to-end DOA estimation. 3. **Automated Hyperparameter Optimization**: Using Bayesian optimization strategies to automatically adjust the hyperparameters in MPDNN, thereby achieving overall model optimization. This method incorporates hyperparameters as penalty terms in the loss function, defining the problem through maximum likelihood estimation. Through a series of plane wave simulation experiments, the paper demonstrates that AutoMPDNN exhibits higher predictive performance compared to traditional DOA estimation algorithms and modern deep learning DOA methods in one-source single-snapshot scenarios under different signal-to-noise ratio (SNR) conditions. This indicates that AutoMPDNN can effectively address the challenges of sparse data and shows high accuracy and real-time performance in practical applications.

An automated multi-layer perceptron discriminative neural network based on Bayesian optimization achieves high-precision one-source single-snapshot direction-of-arrival estimation

A high-resolution method for direction of arrival estimation based on an improved self-attention module

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

Performance Analysis of Deep Neural Networks for Direction of Arrival Estimation of Multiple Sources

Neural network-based DOA estimation for distributed sources in massive MIMO systems

Direction of Arrival Estimation Based on DNN and CNN

SDOA-Net: An Efficient Deep-Learning-Based DOA Estimation Network for Imperfect Array

LDnADMM-Net: A Denoising Unfolded Deep Neural Network for Direction-of-Arrival Estimations in A Low Signal-to-Noise Ratio

A Two-Stage Multi-Layer Perceptron for High-Resolution DOA Estimation

CAE-CNN-Based DOA Estimation Method for Low-Elevation-Angle Target

Performance Analysis of Direction of Arrival Estimation Based on Deep Learning.

SDOA-Net: An Efficient Deep Learning-Based DOA Estimation Network for Imperfect Array

Lightweight Deep Neural Network with Data Redundancy Removal and Regression for DOA Estimation in Sensor Array

Reliable underwater multi-target direction of arrival estimation with optimal transport using deep models

Deep learning approach for high-resolution DOA estimation

Direction-of-Arrival Estimation Method Based on Neural Network with Temporal Structure for Underwater Acoustic Vector Sensor Array

Vehicle Positioning With Deep-Learning-Based Direction-of-Arrival Estimation of Incoherently Distributed Sources

MDTCNet: Multi-Task Classifications Network and TCNN for Direction of Arrival Estimation

An off-grid direction-of-arrival estimator based on sparse Bayesian learning with three-stage hierarchical Laplace priors

Off-Grid Direction of Arrival Estimation Using Sparse Bayesian Inference.

A3R-Net: adaptive attention aggregation residual network for sparse DOA estimation