Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation

Marcio L. Lima de Oliveira; Marco J.G. Bekooij

doi:10.1109/RADAR54928.2023.10371066

Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation

Marcio L. Lima de Oliveira^*
, Marco J.G. Bekooij

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

2 Citations (Scopus)

114 Downloads (Pure)

Abstract

In this paper, we discuss the fusion of deep neural networks with the Maximum Likelihood Estimation (MLE) algorithm for estimating both the Direction Of Arrival (DOA) of MIMO radars and the number of sources for a single-snapshot scenario using a Minimum Redundancy antenna Array (MRA) with imperfections and evaluating it with synthetic and real-world data. The combination of Deep Learning (DL) with the classical MLE seems to be a viable solution for this problem, as it is less computationally intensive than a classical MLE while not losing generalization and being better at estimating the number of sources. In both our experiments, using synthetic and real-world data, our method performs close to MLE and appears to be a deployable solution for real scenarios. Besides reducing the computational load of MLE, the novelty of our model lies in the fact that the deep learning model learns the gain and phase errors as a function of the direction of arrival instead of applying a simple calibration.

Original language	English
Title of host publication	2023 IEEE International Radar Conference, RADAR 2023
Place of Publication	Piscataway, NJ
Publisher	IEEE
ISBN (Electronic)	978-1-6654-8278-3
ISBN (Print)	978-1-6654-8279-0
DOIs	https://doi.org/10.1109/RADAR54928.2023.10371066
Publication status	Published - 28 Dec 2023
Event	2023 International Radar Conference, RADAR 2023 - Sydney, Australia Duration: 5 Nov 2023 → 10 Nov 2023 https://www.radar2023.org/

Publication series

Name	Proceedings of the IEEE Radar Conference (RADAR)
Publisher	IEEE
Volume	2023
ISSN (Print)	1097-5764
ISSN (Electronic)	2375-5318

Conference

Conference	2023 International Radar Conference, RADAR 2023
Abbreviated title	RADAR 2023
Country/Territory	Australia
City	Sydney
Period	5/11/23 → 10/11/23
Internet address	https://www.radar2023.org/

Keywords

2024 OA procedure
Deep Learning (DL)
Direction of arrival
Machine Learning (ML)
Maximum likelihood estimation
Antenna imperfection mitigation

Access to Document

10.1109/RADAR54928.2023.10371066

ArticleFinal published version, 261 KBLicence: Taverne

Cite this

@inproceedings{1c05359e0dd04283b5ba8250a4fab661,

title = "Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation",

abstract = "In this paper, we discuss the fusion of deep neural networks with the Maximum Likelihood Estimation (MLE) algorithm for estimating both the Direction Of Arrival (DOA) of MIMO radars and the number of sources for a single-snapshot scenario using a Minimum Redundancy antenna Array (MRA) with imperfections and evaluating it with synthetic and real-world data. The combination of Deep Learning (DL) with the classical MLE seems to be a viable solution for this problem, as it is less computationally intensive than a classical MLE while not losing generalization and being better at estimating the number of sources. In both our experiments, using synthetic and real-world data, our method performs close to MLE and appears to be a deployable solution for real scenarios. Besides reducing the computational load of MLE, the novelty of our model lies in the fact that the deep learning model learns the gain and phase errors as a function of the direction of arrival instead of applying a simple calibration.",

keywords = "2024 OA procedure, Deep Learning (DL), Direction of arrival, Machine Learning (ML), Maximum likelihood estimation, Antenna imperfection mitigation",

author = "\{Lima de Oliveira\}, \{Marcio L.\} and Bekooij, \{Marco J.G.\}",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 International Radar Conference, RADAR 2023, RADAR 2023 ; Conference date: 05-11-2023 Through 10-11-2023",

year = "2023",

month = dec,

day = "28",

doi = "10.1109/RADAR54928.2023.10371066",

language = "English",

isbn = "978-1-6654-8279-0",

series = "Proceedings of the IEEE Radar Conference (RADAR)",

publisher = "IEEE",

booktitle = "2023 IEEE International Radar Conference, RADAR 2023",

address = "United States",

url = "https://www.radar2023.org/",

}

Lima de Oliveira, ML & Bekooij, MJG 2023, Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation. in 2023 IEEE International Radar Conference, RADAR 2023. Proceedings of the IEEE Radar Conference (RADAR), vol. 2023, IEEE, Piscataway, NJ, 2023 International Radar Conference, RADAR 2023, Sydney, Australia, 5/11/23. https://doi.org/10.1109/RADAR54928.2023.10371066

Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation. / Lima de Oliveira, Marcio L.; Bekooij, Marco J.G.
2023 IEEE International Radar Conference, RADAR 2023. Piscataway, NJ: IEEE, 2023. (Proceedings of the IEEE Radar Conference (RADAR); Vol. 2023).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation

AU - Lima de Oliveira, Marcio L.

AU - Bekooij, Marco J.G.

PY - 2023/12/28

Y1 - 2023/12/28

N2 - In this paper, we discuss the fusion of deep neural networks with the Maximum Likelihood Estimation (MLE) algorithm for estimating both the Direction Of Arrival (DOA) of MIMO radars and the number of sources for a single-snapshot scenario using a Minimum Redundancy antenna Array (MRA) with imperfections and evaluating it with synthetic and real-world data. The combination of Deep Learning (DL) with the classical MLE seems to be a viable solution for this problem, as it is less computationally intensive than a classical MLE while not losing generalization and being better at estimating the number of sources. In both our experiments, using synthetic and real-world data, our method performs close to MLE and appears to be a deployable solution for real scenarios. Besides reducing the computational load of MLE, the novelty of our model lies in the fact that the deep learning model learns the gain and phase errors as a function of the direction of arrival instead of applying a simple calibration.

AB - In this paper, we discuss the fusion of deep neural networks with the Maximum Likelihood Estimation (MLE) algorithm for estimating both the Direction Of Arrival (DOA) of MIMO radars and the number of sources for a single-snapshot scenario using a Minimum Redundancy antenna Array (MRA) with imperfections and evaluating it with synthetic and real-world data. The combination of Deep Learning (DL) with the classical MLE seems to be a viable solution for this problem, as it is less computationally intensive than a classical MLE while not losing generalization and being better at estimating the number of sources. In both our experiments, using synthetic and real-world data, our method performs close to MLE and appears to be a deployable solution for real scenarios. Besides reducing the computational load of MLE, the novelty of our model lies in the fact that the deep learning model learns the gain and phase errors as a function of the direction of arrival instead of applying a simple calibration.

KW - 2024 OA procedure

KW - Deep Learning (DL)

KW - Direction of arrival

KW - Machine Learning (ML)

KW - Maximum likelihood estimation

KW - Antenna imperfection mitigation

UR - https://www.scopus.com/pages/publications/85182728340

U2 - 10.1109/RADAR54928.2023.10371066

DO - 10.1109/RADAR54928.2023.10371066

M3 - Conference contribution

AN - SCOPUS:85182728340

SN - 978-1-6654-8279-0

T3 - Proceedings of the IEEE Radar Conference (RADAR)

BT - 2023 IEEE International Radar Conference, RADAR 2023

PB - IEEE

CY - Piscataway, NJ

T2 - 2023 International Radar Conference, RADAR 2023

Y2 - 5 November 2023 through 10 November 2023

ER -

Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this