CNN-Aided Self-Interference Estimation for In-Band Full-Duplex Systems
Modern radio access technologies approach Shannon’s limit, necessitating innovative methods for enhanced spectral efficiency. In-band full-duplex (IBFD) can double the spectral efficiency, enabling simultaneous transmission and reception over the same time-frequency resource. IBFD faces the challenge of self-interference, which has to be canceled by up to 100 dB. This paper estimates the loopback channel through convolutional neural networks (CNNs), which leverage the natural signal structure of wireless channels, effectively mapping time-frequency features. We test the method via simulations in two measured channels, showing cancelations of up to 52 dB.
Authors
- Zhihong Hong (ORCID: https://orcid.org/0000-0001-5653-8043)
- Jon Montalbán (ORCID: https://orcid.org/0000-0003-0309-3401)
- Iñigo Bilbao (ORCID: https://orcid.org/0000-0003-1166-0568)
- Pablo Angueira (ORCID: https://orcid.org/0000-0002-5188-8412)
- Yiyan Wu (ORCID: https://orcid.org/0000-0001-8890-5389)
- Eneko Iradier (ORCID: https://orcid.org/0000-0002-0424-3857)
Institutions
- University of the Basque Country (ES)
- Communications Research Centre Canada (CA)
Publication Details
- Journal
- IEEE Transactions on Cognitive Communications and Networking
- Published
- 2025-02-21
- DOI
- https://doi.org/10.1109/tccn.2025.3544838
- Primary Topic
- Full-Duplex Wireless Communications
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- European Commission
- Eusko Jaurlaritza
- Euskal Herriko Unibertsitatea
- European Regional Development Fund
- Agencia Estatal de Investigación
- CHIST-ERA