Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network
The symmetric positive definite (SPD) matrix has attracted much attention in classification problems because of its remarkable performance, which is due to the underlying structure of the Riemannian manifold with non-negative curvature as well as the use of non-linear geometric metrics, which have a...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-05-01
|
Series: | Entropy |
Subjects: | |
Online Access: | https://www.mdpi.com/1099-4300/22/5/585 |
id |
doaj-1bf0aaa3fb0e4275b1ced0c3751aedc8 |
---|---|
record_format |
Article |
spelling |
doaj-1bf0aaa3fb0e4275b1ced0c3751aedc82020-11-25T03:03:49ZengMDPI AGEntropy1099-43002020-05-012258558510.3390/e22050585Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral NetworkJiangyi Wang0Xiaoqiang Hua1Xinwu Zeng2School of Meteorology and Oceanography, National University of Defence Technology, Changsha 410073, ChinaSchool of Meteorology and Oceanography, National University of Defence Technology, Changsha 410073, ChinaSchool of Meteorology and Oceanography, National University of Defence Technology, Changsha 410073, ChinaThe symmetric positive definite (SPD) matrix has attracted much attention in classification problems because of its remarkable performance, which is due to the underlying structure of the Riemannian manifold with non-negative curvature as well as the use of non-linear geometric metrics, which have a stronger ability to distinguish SPD matrices and reduce information loss compared to the Euclidean metric. In this paper, we propose a spectral-based SPD matrix signal detection method with deep learning that uses time-frequency spectra to construct SPD matrices and then exploits a deep SPD matrix learning network to detect the target signal. Using this approach, the signal detection problem is transformed into a binary classification problem on a manifold to judge whether the input sample has target signal or not. Two matrix models are applied, namely, an SPD matrix based on spectral covariance and an SPD matrix based on spectral transformation. A simulated-signal dataset and a semi-physical simulated-signal dataset are used to demonstrate that the spectral-based SPD matrix signal detection method with deep learning has a gain of 1.7–3.3 dB under appropriate conditions. The results show that our proposed method achieves better detection performances than its state-of-the-art spectral counterparts that use convolutional neural networks.https://www.mdpi.com/1099-4300/22/5/585signal detectionSPD matrix constructionSPD matrix learningneural networks |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Jiangyi Wang Xiaoqiang Hua Xinwu Zeng |
spellingShingle |
Jiangyi Wang Xiaoqiang Hua Xinwu Zeng Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network Entropy signal detection SPD matrix construction SPD matrix learning neural networks |
author_facet |
Jiangyi Wang Xiaoqiang Hua Xinwu Zeng |
author_sort |
Jiangyi Wang |
title |
Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network |
title_short |
Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network |
title_full |
Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network |
title_fullStr |
Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network |
title_full_unstemmed |
Spectral-Based SPD Matrix Representation for Signal Detection Using a Deep Neutral Network |
title_sort |
spectral-based spd matrix representation for signal detection using a deep neutral network |
publisher |
MDPI AG |
series |
Entropy |
issn |
1099-4300 |
publishDate |
2020-05-01 |
description |
The symmetric positive definite (SPD) matrix has attracted much attention in classification problems because of its remarkable performance, which is due to the underlying structure of the Riemannian manifold with non-negative curvature as well as the use of non-linear geometric metrics, which have a stronger ability to distinguish SPD matrices and reduce information loss compared to the Euclidean metric. In this paper, we propose a spectral-based SPD matrix signal detection method with deep learning that uses time-frequency spectra to construct SPD matrices and then exploits a deep SPD matrix learning network to detect the target signal. Using this approach, the signal detection problem is transformed into a binary classification problem on a manifold to judge whether the input sample has target signal or not. Two matrix models are applied, namely, an SPD matrix based on spectral covariance and an SPD matrix based on spectral transformation. A simulated-signal dataset and a semi-physical simulated-signal dataset are used to demonstrate that the spectral-based SPD matrix signal detection method with deep learning has a gain of 1.7–3.3 dB under appropriate conditions. The results show that our proposed method achieves better detection performances than its state-of-the-art spectral counterparts that use convolutional neural networks. |
topic |
signal detection SPD matrix construction SPD matrix learning neural networks |
url |
https://www.mdpi.com/1099-4300/22/5/585 |
work_keys_str_mv |
AT jiangyiwang spectralbasedspdmatrixrepresentationforsignaldetectionusingadeepneutralnetwork AT xiaoqianghua spectralbasedspdmatrixrepresentationforsignaldetectionusingadeepneutralnetwork AT xinwuzeng spectralbasedspdmatrixrepresentationforsignaldetectionusingadeepneutralnetwork |
_version_ |
1724684417620770816 |