Proceedings Article10.1109/ICFEICT57213.2022.00035
DOA Estimation Using Beamspace-Based Deep Neural Network
Yuanjie Ji,Cai Wen,Yan Huang,Jinye Peng +3 more
- 01 Aug 2022
pp 153-158
TL;DR: In this article , the authors used deep learning to estimate the direction of arrival (DOA) in the virtual array beam space to adapt to the array imperfections, and a training method robust to various array-imperfections is proposed, that is, a spherical model is used to simulate the distribution of different array imperfection, and training sets are generated in the distribution.
read more
Abstract: Direction-of-arrival (DOA) estimation is to analyze the received sensor array data to determine the direction of the signal, so as to better perform beamforming or determine the target position. In the field of DOA estimation, traditional methods are parametric, while machine learning methods rely heavily on the consistency of test and training samples, and when there is array imperfection, their performance is extremely degraded. Therefore, this paper uses deep learning to solve the DOA estimation problem, and makes DOA estimation in the virtual array beam space to adapt to the array imperfections, and a training method robust to various array imperfections is proposed, that is, a spherical model is used to simulate the distribution of different array imperfections, and training sets are generated in the distribution. Simulation results show that the proposed method has good adaptability to array imperfections.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
References
Direction-of-Arrival Estimation Based on Deep Neural Networks With Robustness to Array Imperfections
TL;DR: A framework of the deep neural network to address the DOA estimation problem, so as to obtain good adaptation to array imperfections and enhanced generalization to unseen scenarios andSimulations are carried out to show that the proposed method performs satisfyingly in both generalization and imperfection adaptation.
Broadband DOA estimation using Convolutional neural networks trained with noise signals
TL;DR: In this paper, a convolution neural network (CNN) based classification method for broadband DOA estimation is proposed, where the phase component of the short-time Fourier transform coefficients of the received microphone signals are directly fed into the CNN and the features required for DOA estimations are learnt during training.
275
DeepMUSIC: Multiple Signal Classification via Deep Learning
TL;DR: In this article, a DL framework for multiple signal classification (DeepMUSIC) is proposed, where each CNN is fed with the array covariance matrix and it learns the MUSIC spectra of the corresponding angular subregion.
122
Time-Frequency Multi-Invariance ESPRIT for DOA Estimation
TL;DR: In this paper, a new ESPRIT-type algorithm is presented for estimating the direction of arrival (DOA), which reconstructs the received signal to form a time-frequency data model with a multi-invariance (MI) property, and then a $t{\hbox{-}}f$ MI Estimation of Signal Parameters via Rotational Invariance Technique (ESPRIT) algorithm is proposed.
82
DeepMUSIC: Multiple Signal Classification via Deep Learning
Ahmet M. Elbir
- 12 Mar 2020
TL;DR: It is shown, through simulations, that the proposed DeepMUSIC framework has superior estimation accuracy and exhibits less computational complexity in comparison with both DL- and non-DL-based techniques.
79