Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM

Mehmet Ali Aygul; Mahmoud Nazzal; Ali Riza Ekti; Ali Gorcin; Daniel Benevides Da Costa; Hasan Fehmi Ates; Huseyin Arslan

doi:10.1109/VTC2020-Spring48590.2020.9129001

Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM

Mehmet Ali Aygul, Mahmoud Nazzal, Ali Riza Ekti, Ali Gorcin, Daniel Benevides Da Costa, Hasan Fehmi Ates, Huseyin Arslan

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

13 Citations (Scopus)

Abstract

The identification of spectrum opportunities is a pivotal requirement for efficient spectrum utilization in cognitive radio systems. Spectrum prediction offers a convenient means for revealing such opportunities based on the previously obtained occupancies. As spectrum occupancy states are correlated over time, spectrum prediction is often cast as a predictable time-series process using classical or deep learning-based models. However, this variety of methods exploits time-domain correlation and overlooks the existing correlation over frequency. In this paper, differently from previous works, we investigate a more realistic scenario by exploiting correlation over time and frequency through a 2D-long short-term memory (LSTM) model. Extensive experimental results show a performance improvement over conventional spectrum prediction methods in terms of accuracy and computational complexity. These observations are validated over the real-world spectrum measurements, assuming a frequency range between 832-862 MHz where most of the telecom operators in Turkey have private uplink bands.

Original language	English
Title of host publication	2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781728152073
DOIs	https://doi.org/10.1109/VTC2020-Spring48590.2020.9129001
Publication status	Published - May 2020
Externally published	Yes
Event	91st IEEE Vehicular Technology Conference, VTC Spring 2020 - Antwerp, Belgium Duration: 25 May 2020 → 28 May 2020

Publication series

Name	IEEE Vehicular Technology Conference
Volume	2020-May
ISSN (Print)	1550-2252

Conference

Conference	91st IEEE Vehicular Technology Conference, VTC Spring 2020
Country/Territory	Belgium
City	Antwerp
Period	25/05/20 → 28/05/20

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

Funding

ACKNOWLEDGEMENT This publication was made possible by NPRP12S-0225-190152 from the Qatar National Research Fund (a member of The Qatar Foundation). The statements made herein are solely the responsibility of the author[s]. The work of D. B. da Costa was supported by the Scientific and Technological Research Council of Turkey (TUBITAK) 2221 Programme.

Funders	Funder number
TUBITAK
Qatar National Research Fund	NPRP12S-0225-190152
Türkiye Bilimsel ve Teknolojik Araştirma Kurumu

Keywords

Deep learning
frequency correlation
real-world spectrum measurement
spectrum occupancy prediction

Access to Document

10.1109/VTC2020-Spring48590.2020.9129001

Cite this

Aygul, M. A., Nazzal, M., Ekti, A. R., Gorcin, A., Da Costa, D. B., Ates, H. F., & Arslan, H. (2020). Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM. In 2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings Article 9129001 (IEEE Vehicular Technology Conference; Vol. 2020-May). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VTC2020-Spring48590.2020.9129001

@inproceedings{671b8cc79e9a45b3ba9ab29c176c5ede,

title = "Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM",

abstract = "The identification of spectrum opportunities is a pivotal requirement for efficient spectrum utilization in cognitive radio systems. Spectrum prediction offers a convenient means for revealing such opportunities based on the previously obtained occupancies. As spectrum occupancy states are correlated over time, spectrum prediction is often cast as a predictable time-series process using classical or deep learning-based models. However, this variety of methods exploits time-domain correlation and overlooks the existing correlation over frequency. In this paper, differently from previous works, we investigate a more realistic scenario by exploiting correlation over time and frequency through a 2D-long short-term memory (LSTM) model. Extensive experimental results show a performance improvement over conventional spectrum prediction methods in terms of accuracy and computational complexity. These observations are validated over the real-world spectrum measurements, assuming a frequency range between 832-862 MHz where most of the telecom operators in Turkey have private uplink bands.",

keywords = "Deep learning, frequency correlation, real-world spectrum measurement, spectrum occupancy prediction",

author = "Aygul, {Mehmet Ali} and Mahmoud Nazzal and Ekti, {Ali Riza} and Ali Gorcin and {Da Costa}, {Daniel Benevides} and Ates, {Hasan Fehmi} and Huseyin Arslan",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 91st IEEE Vehicular Technology Conference, VTC Spring 2020 ; Conference date: 25-05-2020 Through 28-05-2020",

year = "2020",

month = may,

doi = "10.1109/VTC2020-Spring48590.2020.9129001",

language = "English",

series = "IEEE Vehicular Technology Conference",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings",

address = "United States",

}

Aygul, MA, Nazzal, M, Ekti, AR, Gorcin, A, Da Costa, DB, Ates, HF & Arslan, H 2020, Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM. in 2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings., 9129001, IEEE Vehicular Technology Conference, vol. 2020-May, Institute of Electrical and Electronics Engineers Inc., 91st IEEE Vehicular Technology Conference, VTC Spring 2020, Antwerp, Belgium, 25/05/20. https://doi.org/10.1109/VTC2020-Spring48590.2020.9129001

Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM. / Aygul, Mehmet Ali; Nazzal, Mahmoud; Ekti, Ali Riza et al.
2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2020. 9129001 (IEEE Vehicular Technology Conference; Vol. 2020-May).

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

TY - GEN

T1 - Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM

AU - Aygul, Mehmet Ali

AU - Nazzal, Mahmoud

AU - Ekti, Ali Riza

AU - Gorcin, Ali

AU - Da Costa, Daniel Benevides

AU - Ates, Hasan Fehmi

AU - Arslan, Huseyin

PY - 2020/5

Y1 - 2020/5

N2 - The identification of spectrum opportunities is a pivotal requirement for efficient spectrum utilization in cognitive radio systems. Spectrum prediction offers a convenient means for revealing such opportunities based on the previously obtained occupancies. As spectrum occupancy states are correlated over time, spectrum prediction is often cast as a predictable time-series process using classical or deep learning-based models. However, this variety of methods exploits time-domain correlation and overlooks the existing correlation over frequency. In this paper, differently from previous works, we investigate a more realistic scenario by exploiting correlation over time and frequency through a 2D-long short-term memory (LSTM) model. Extensive experimental results show a performance improvement over conventional spectrum prediction methods in terms of accuracy and computational complexity. These observations are validated over the real-world spectrum measurements, assuming a frequency range between 832-862 MHz where most of the telecom operators in Turkey have private uplink bands.

AB - The identification of spectrum opportunities is a pivotal requirement for efficient spectrum utilization in cognitive radio systems. Spectrum prediction offers a convenient means for revealing such opportunities based on the previously obtained occupancies. As spectrum occupancy states are correlated over time, spectrum prediction is often cast as a predictable time-series process using classical or deep learning-based models. However, this variety of methods exploits time-domain correlation and overlooks the existing correlation over frequency. In this paper, differently from previous works, we investigate a more realistic scenario by exploiting correlation over time and frequency through a 2D-long short-term memory (LSTM) model. Extensive experimental results show a performance improvement over conventional spectrum prediction methods in terms of accuracy and computational complexity. These observations are validated over the real-world spectrum measurements, assuming a frequency range between 832-862 MHz where most of the telecom operators in Turkey have private uplink bands.

KW - Deep learning

KW - frequency correlation

KW - real-world spectrum measurement

KW - spectrum occupancy prediction

UR - http://www.scopus.com/inward/record.url?scp=85088293691&partnerID=8YFLogxK

U2 - 10.1109/VTC2020-Spring48590.2020.9129001

DO - 10.1109/VTC2020-Spring48590.2020.9129001

M3 - Conference contribution

AN - SCOPUS:85088293691

T3 - IEEE Vehicular Technology Conference

BT - 2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 91st IEEE Vehicular Technology Conference, VTC Spring 2020

Y2 - 25 May 2020 through 28 May 2020

ER -

Aygul MA, Nazzal M, Ekti AR, Gorcin A, Da Costa DB, Ates HF et al. Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM. In 2020 IEEE 91st Vehicular Technology Conference, VTC Spring 2020 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2020. 9129001. (IEEE Vehicular Technology Conference). doi: 10.1109/VTC2020-Spring48590.2020.9129001

Spectrum Occupancy Prediction Exploiting Time and Frequency Correlations Through 2D-LSTM

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this