The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
A aquisição de informações de estado de canal (CSI) no lado do transmissor é um grande desafio em sistemas MIMO massivos para permitir transmissões de alta eficiência. Para resolver este problema, vários esquemas de feedback CSI foram propostos, incluindo esquemas de feedback limitado com quantização vetorial baseada em livro de códigos e feedback explícito de matriz de canal. Devido às limitações da capacidade do canal de feedback, um problema comum nestes esquemas é a representação eficiente do CSI com um número limitado de bits no lado do receptor, e a sua reconstrução precisa com base nos bits de feedback do receptor no lado do transmissor. Recentemente, inspiradas por aplicações bem-sucedidas em muitos campos, as tecnologias de aprendizagem profunda (DL) para aquisição de CSI têm recebido considerável interesse de pesquisa tanto da academia quanto da indústria. Considerando o mecanismo de feedback prático das novas redes de rádio (NR) de 5ª geração (5G), propomos dois esquemas de implementação de inteligência artificial para CSI (AI4CSI), o receptor baseado em DL e o design ponta a ponta, respectivamente. Os esquemas AI4CSI propostos foram avaliados em redes 5G NR em termos de eficiência de espectro (SE), sobrecarga de feedback e complexidade computacional, e comparados com esquemas legados. Para demonstrar se esses esquemas podem ser usados em cenários da vida real, tanto os dados do canal modelados quanto os canais medidos na prática foram utilizados em nossas investigações. Quando a aquisição CSI baseada em DL é aplicada apenas ao receptor, que tem pouco impacto na interface aérea, ela fornece ganho SE de aproximadamente 25% em um nível moderado de sobrecarga de feedback. É viável implantá-lo nas atuais redes 5G durante as evoluções do 5G. Para as melhorias CSI baseadas em DL de ponta a ponta, as avaliações também demonstraram seu ganho adicional de desempenho em SE, que é de 6% a 26% em comparação com receptores baseados em DL e de 33% a 58% em comparação com esquemas CSI legados. Considerando o seu grande impacto no design de interfaces aéreas, será uma tecnologia candidata para redes de 6ª geração (6G), nas quais uma interface aérea projetada por inteligência artificial pode ser utilizada.
Xin WANG
DOCOMO Beijing Communications Laboratories, Co. Ltd.
Xiaolin HOU
DOCOMO Beijing Communications Laboratories, Co. Ltd.
Lan CHEN
DOCOMO Beijing Communications Laboratories, Co. Ltd.
Yoshihisa KISHIYAMA
NTT DOCOMO, INC.
Takahiro ASAI
NTT DOCOMO, INC.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Xin WANG, Xiaolin HOU, Lan CHEN, Yoshihisa KISHIYAMA, Takahiro ASAI, "Deep Learning-Based Massive MIMO CSI Acquisition for 5G Evolution and 6G" in IEICE TRANSACTIONS on Communications,
vol. E105-B, no. 12, pp. 1559-1568, December 2022, doi: 10.1587/transcom.2022EBP3009.
Abstract: Channel state information (CSI) acquisition at the transmitter side is a major challenge in massive MIMO systems for enabling high-efficiency transmissions. To address this issue, various CSI feedback schemes have been proposed, including limited feedback schemes with codebook-based vector quantization and explicit channel matrix feedback. Owing to the limitations of feedback channel capacity, a common issue in these schemes is the efficient representation of the CSI with a limited number of bits at the receiver side, and its accurate reconstruction based on the feedback bits from the receiver at the transmitter side. Recently, inspired by successful applications in many fields, deep learning (DL) technologies for CSI acquisition have received considerable research interest from both academia and industry. Considering the practical feedback mechanism of 5th generation (5G) New radio (NR) networks, we propose two implementation schemes for artificial intelligence for CSI (AI4CSI), the DL-based receiver and end-to-end design, respectively. The proposed AI4CSI schemes were evaluated in 5G NR networks in terms of spectrum efficiency (SE), feedback overhead, and computational complexity, and compared with legacy schemes. To demonstrate whether these schemes can be used in real-life scenarios, both the modeled-based channel data and practically measured channels were used in our investigations. When DL-based CSI acquisition is applied to the receiver only, which has little air interface impact, it provides approximately 25% SE gain at a moderate feedback overhead level. It is feasible to deploy it in current 5G networks during 5G evolutions. For the end-to-end DL-based CSI enhancements, the evaluations also demonstrated their additional performance gain on SE, which is 6%-26% compared with DL-based receivers and 33%-58% compared with legacy CSI schemes. Considering its large impact on air-interface design, it will be a candidate technology for 6th generation (6G) networks, in which an air interface designed by artificial intelligence can be used.
URL: https://global.ieice.org/en_transactions/communications/10.1587/transcom.2022EBP3009/_p
Copiar
@ARTICLE{e105-b_12_1559,
author={Xin WANG, Xiaolin HOU, Lan CHEN, Yoshihisa KISHIYAMA, Takahiro ASAI, },
journal={IEICE TRANSACTIONS on Communications},
title={Deep Learning-Based Massive MIMO CSI Acquisition for 5G Evolution and 6G},
year={2022},
volume={E105-B},
number={12},
pages={1559-1568},
abstract={Channel state information (CSI) acquisition at the transmitter side is a major challenge in massive MIMO systems for enabling high-efficiency transmissions. To address this issue, various CSI feedback schemes have been proposed, including limited feedback schemes with codebook-based vector quantization and explicit channel matrix feedback. Owing to the limitations of feedback channel capacity, a common issue in these schemes is the efficient representation of the CSI with a limited number of bits at the receiver side, and its accurate reconstruction based on the feedback bits from the receiver at the transmitter side. Recently, inspired by successful applications in many fields, deep learning (DL) technologies for CSI acquisition have received considerable research interest from both academia and industry. Considering the practical feedback mechanism of 5th generation (5G) New radio (NR) networks, we propose two implementation schemes for artificial intelligence for CSI (AI4CSI), the DL-based receiver and end-to-end design, respectively. The proposed AI4CSI schemes were evaluated in 5G NR networks in terms of spectrum efficiency (SE), feedback overhead, and computational complexity, and compared with legacy schemes. To demonstrate whether these schemes can be used in real-life scenarios, both the modeled-based channel data and practically measured channels were used in our investigations. When DL-based CSI acquisition is applied to the receiver only, which has little air interface impact, it provides approximately 25% SE gain at a moderate feedback overhead level. It is feasible to deploy it in current 5G networks during 5G evolutions. For the end-to-end DL-based CSI enhancements, the evaluations also demonstrated their additional performance gain on SE, which is 6%-26% compared with DL-based receivers and 33%-58% compared with legacy CSI schemes. Considering its large impact on air-interface design, it will be a candidate technology for 6th generation (6G) networks, in which an air interface designed by artificial intelligence can be used.},
keywords={},
doi={10.1587/transcom.2022EBP3009},
ISSN={1745-1345},
month={December},}
Copiar
TY - JOUR
TI - Deep Learning-Based Massive MIMO CSI Acquisition for 5G Evolution and 6G
T2 - IEICE TRANSACTIONS on Communications
SP - 1559
EP - 1568
AU - Xin WANG
AU - Xiaolin HOU
AU - Lan CHEN
AU - Yoshihisa KISHIYAMA
AU - Takahiro ASAI
PY - 2022
DO - 10.1587/transcom.2022EBP3009
JO - IEICE TRANSACTIONS on Communications
SN - 1745-1345
VL - E105-B
IS - 12
JA - IEICE TRANSACTIONS on Communications
Y1 - December 2022
AB - Channel state information (CSI) acquisition at the transmitter side is a major challenge in massive MIMO systems for enabling high-efficiency transmissions. To address this issue, various CSI feedback schemes have been proposed, including limited feedback schemes with codebook-based vector quantization and explicit channel matrix feedback. Owing to the limitations of feedback channel capacity, a common issue in these schemes is the efficient representation of the CSI with a limited number of bits at the receiver side, and its accurate reconstruction based on the feedback bits from the receiver at the transmitter side. Recently, inspired by successful applications in many fields, deep learning (DL) technologies for CSI acquisition have received considerable research interest from both academia and industry. Considering the practical feedback mechanism of 5th generation (5G) New radio (NR) networks, we propose two implementation schemes for artificial intelligence for CSI (AI4CSI), the DL-based receiver and end-to-end design, respectively. The proposed AI4CSI schemes were evaluated in 5G NR networks in terms of spectrum efficiency (SE), feedback overhead, and computational complexity, and compared with legacy schemes. To demonstrate whether these schemes can be used in real-life scenarios, both the modeled-based channel data and practically measured channels were used in our investigations. When DL-based CSI acquisition is applied to the receiver only, which has little air interface impact, it provides approximately 25% SE gain at a moderate feedback overhead level. It is feasible to deploy it in current 5G networks during 5G evolutions. For the end-to-end DL-based CSI enhancements, the evaluations also demonstrated their additional performance gain on SE, which is 6%-26% compared with DL-based receivers and 33%-58% compared with legacy CSI schemes. Considering its large impact on air-interface design, it will be a candidate technology for 6th generation (6G) networks, in which an air interface designed by artificial intelligence can be used.
ER -