Bit-level Optimized Neural Network for Multi-antenna Channel Quantization

Chao Lu; Wei Xu; Shi Jin; Kezhi Wang

doi:10.1109/LWC.2019.2942908

Bit-level Optimized Neural Network for Multi-antenna Channel Quantization

Chao Lu, Wei Xu, Shi Jin, Kezhi Wang

School of Computer Science

Research output: Contribution to journal › Article › peer-review

63 Citations (Scopus)

40 Downloads (Pure)

Abstract

Quantized channel state information (CSI) plays a critical role in precoding design which helps reap the merits of multiple-input multiple-output (MIMO) technology. In order to reduce the overhead of CSI feedback, we propose a deep learning based CSI quantization method by developing a joint convolutional residual network (JC-ResNet) which benefits MIMO channel feature extraction and recovery from the perspective of bit-level quantization performance. Experiments show that our proposed method substantially improves the performance.

Original language	English
Pages (from-to)	87-90
Number of pages	4
Journal	IEEE Wireless Communications Letters
Volume	9
Issue number	1
Early online date	20 Sept 2019
DOIs	https://doi.org/10.1109/LWC.2019.2942908
Publication status	Published - 9 Jan 2020

Keywords

Channel state information (CSI)
quantization
multiple-input multiple-output (MIMO).

Access to Document

10.1109/LWC.2019.2942908

08845636Accepted author manuscript, 287 KB

Cite this

@article{3f4e1005d30b4be5b6c935b1ea39e422,

title = "Bit-level Optimized Neural Network for Multi-antenna Channel Quantization",

abstract = "Quantized channel state information (CSI) plays a critical role in precoding design which helps reap the merits of multiple-input multiple-output (MIMO) technology. In order to reduce the overhead of CSI feedback, we propose a deep learning based CSI quantization method by developing a joint convolutional residual network (JC-ResNet) which benefits MIMO channel feature extraction and recovery from the perspective of bit-level quantization performance. Experiments show that our proposed method substantially improves the performance.",

keywords = "Channel state information (CSI), quantization, multiple-input multiple-output (MIMO).",

author = "Chao Lu and Wei Xu and Shi Jin and Kezhi Wang",

note = "Research funded by National Basic Research Program of China (973 Program) (2018YFA0701602), Natural Science Foundation of Jiangsu Province (BK20190012), National Natural Science Foundation of China (61871109), Royal Society",

year = "2020",

month = jan,

day = "9",

doi = "10.1109/LWC.2019.2942908",

language = "English",

volume = "9",

pages = "87--90",

journal = "IEEE Wireless Communications Letters",

issn = "2162-2337",

publisher = "IEEE",

number = "1",

}

TY - JOUR

T1 - Bit-level Optimized Neural Network for Multi-antenna Channel Quantization

AU - Lu, Chao

AU - Xu, Wei

AU - Jin, Shi

AU - Wang, Kezhi

N1 - Research funded by National Basic Research Program of China (973 Program) (2018YFA0701602), Natural Science Foundation of Jiangsu Province (BK20190012), National Natural Science Foundation of China (61871109), Royal Society

PY - 2020/1/9

Y1 - 2020/1/9

N2 - Quantized channel state information (CSI) plays a critical role in precoding design which helps reap the merits of multiple-input multiple-output (MIMO) technology. In order to reduce the overhead of CSI feedback, we propose a deep learning based CSI quantization method by developing a joint convolutional residual network (JC-ResNet) which benefits MIMO channel feature extraction and recovery from the perspective of bit-level quantization performance. Experiments show that our proposed method substantially improves the performance.

AB - Quantized channel state information (CSI) plays a critical role in precoding design which helps reap the merits of multiple-input multiple-output (MIMO) technology. In order to reduce the overhead of CSI feedback, we propose a deep learning based CSI quantization method by developing a joint convolutional residual network (JC-ResNet) which benefits MIMO channel feature extraction and recovery from the perspective of bit-level quantization performance. Experiments show that our proposed method substantially improves the performance.

KW - Channel state information (CSI)

KW - quantization

KW - multiple-input multiple-output (MIMO).

UR - https://www.scopus.com/pages/publications/85078446297

U2 - 10.1109/LWC.2019.2942908

DO - 10.1109/LWC.2019.2942908

M3 - Article

SN - 2162-2337

VL - 9

SP - 87

EP - 90

JO - IEEE Wireless Communications Letters

JF - IEEE Wireless Communications Letters

IS - 1

ER -

Bit-level Optimized Neural Network for Multi-antenna Channel Quantization

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this