To Learn or Not to Learn:Deep Learning Assisted Wireless Modem Design

doi:10.12142/ZTECOM.201904002

ZTE Communications ›› 2019, Vol. 17 ›› Issue (4): 3-11.DOI: 10.12142/ZTECOM.201904002

收稿日期:2019-09-19 出版日期:2019-12-25 发布日期:2020-04-16

To Learn or Not to Learn:Deep Learning Assisted Wireless Modem Design

XUE Songyan¹, LI Ang¹, WANG Jinfei¹, YI Na¹, MA Yi¹(), Rahim TAFAZOLLI¹, Terence DODGSON²

^1.Institute for Communication Systems, University of Surrey, Guildford, GU2 7XH, the United Kingdom
^2.Airbus Defense and Space, Portsmouth, PO3 5PU, the United Kingdom

Received:2019-09-19 Online:2019-12-25 Published:2020-04-16
About author:XUE Songyan received the B.E. degree from University of Electronic Science and Technology of China and M.S. degree from University of Surrey, the United Kingdom. He is currently pursuing the Ph.D. degree at the Institute of Communication Systems (ICS), University of Surrey. His research interests center around deep learning applications in wireless communication physical layer.|LI Ang received the B.E. degree from Northwestern Polytechnical University, China, and M.S. degree from University of Surrey. He is currently pursuing the Ph.D. degree at the Institute of Communication Systems (ICS), University of Surrey. His research interests focus on the multi-carrier waveform design for future physical layer, machine learning and non-linear transceiver optimization.|WANG Jin fei received the B.S. degree from University of Science and Technology of China, and M.S. degree from University of Surrey. He is currently pursuing the Ph.D. degree at the Institute of Communication Systems (ICS), University of Surrey. His research interests include mobile edge computing and URLLC.|YI Na is the Founder and Director of DEEPGO Ltd. She is also a Senior Research Fellow in Institute for Communication Systems (ICS), University of Surrey. She has had 15-year experience in coordination of international research projects funded by European Commission or EPSRC. She has established rich cooperation across EU, China and other parts of Asia. Her current research interest focuses on machine learning for wireless networks, large-scale 5G pilot platform, and C2X communications.|MA Yis the Head of AI Wireless Group within the Institute of Communication Systems (ICS), University of Surrey, the United Kingdom. He has authored and co-authored 150+ IEEE journal and conference papers as well as 5 patents in the field of multiuser information theory, signal processing and machine learning for wireless communications.y.ma@surrey.ac.uk|Rahim Tafazolliis the Director of Institute for Communications (ICS) and the Director of 5G Innovation Centre (5GIC), University of Surrey. He is the Regius Professor of Electrical Engineering in the United Kingdom. He has authored and co-authored 600+ peer-reviewed journal and conference papers, and been the co-inventor of 20+ patents in the field of mobile and satellite communications.|Terence Dodgson is qualified to Ph.D. level in image processing/electrical and electronic engineering/telecommunications and has a substantial technical background plus wide domain knowledge in the area of digital communications systems, research and algorithm development, implementation, system design and validation (in modern mobile communications systems such as GSM, UMTS, HSDPA, LTE, satellite communications systems and image processing systems), including artificial intelligence. He currently works for Airbus in the Modems Advanced Development Group.
Supported by:
EU H2020 5G-DRIVE Programme(814956);Airbus Defense and Space;the UK 5G Innovation Centre (5GIC)

摘要/Abstract

Abstract:

Deep learning is driving a radical paradigm shift in wireless communications, all the way from the application layer down to the physical layer. Despite this, there is an ongoing debate as to what additional values artificial intelligence (or machine learning) could bring to us, particularly on the physical layer design; and what penalties there may have? These questions motivate a fundamental rethinking of the wireless modem design in the artificial intelligence era. Through several physical-layer case studies, we argue for a significant role that machine learning could play, for instance in parallel error-control coding and decoding, channel equalization, interference cancellation, as well as multiuser and multiantenna detection. In addition, we discuss the fundamental bottlenecks of machine learning as well as their potential solutions in this paper.

Key words: deep learning, neural networks, machine learning, modulation and coding

. [J]. ZTE Communications, 2019, 17(4): 3-11.

XUE Songyan, LI Ang, WANG Jinfei, YI Na, MA Yi, Rahim TAFAZOLLI, Terence DODGSON. To Learn or Not to Learn:Deep Learning Assisted Wireless Modem Design[J]. ZTE Communications, 2019, 17(4): 3-11.

图/表 9

Figure 1. A simple example of Level B communication problem (semantic communication).

Figure 2. Block diagram of the physical-layer (PHY) procedure of point-to-point communication.

Figure 3. An artiflcial neural network (AN-assisted parallel decoder: (a) two-stage parallel decoding and (b) AN-assisted sub- codeword decoder.There are three hidden layers, with each employing Rectified Linear Unit (ReLU) activation function; the output layer is equipped with sigmoid activaltion function, which outputs the estimated original information bits.

Figure 4. Bit reliability and latency evaluation for decoding of halfrate non-recursive convolutional codes with a codeword length of 64.

Figure 5. Block diagram of the deep artifcial neural network (deep- ANN) assisted lassification- and-then-multiuser detection (CAT-MUD).

Figure 6. BLER of classification-and-then -multiuser-detection (CATMUD) as a function of Eb/N0 (dB) over Rayleigh fading channels.

Figure 7. Block diagram of deep-learning assisted multiuser multipleinput multiple-out (MU-MIMO) detection algorithms.

Figure 8. Block diagram of the DNN-PIC approach.

Figure 9. Average BER as a function of Eb/N0 for uncoded 4-by-8 multiuser multiple- input multiple-out (MU-SIMO) system with BPSK modulation.

参考文献 18

1	FG⁃NET⁃2030.Network 2030: Blueprint of TechnologyA, Applications and Market Drivers Towards the Year 2030 and Beyond[EB/OL]. (2019⁃05⁃20).
2	O’SHEA T J,ERPEK T,CLANCY T.Deep Learning Based MIMO Communications[EB/OL]. (2017-07-25).
3	NAOUM R S,AL⁃SULTANI Z N.Learning Vector Quantization (LVQ) and K⁃nearest Neighbor for Intrusion Classification [J].World of Computer Science and Information Technology Journal,2012,2(3):105-109
4	O’SHEA T J,HOYDIS J.An Introduction to Deep Learning for the Physical Layer [J].IEEE Transactions on Cognitive Communications and Networking,2017,3(4):563-575.DOI:10.1109/TCCN.2017.2758370 DOI
5	XUE S Y,MA Y,YI N,et al.Unsupervised Deep Learning for MU⁃SIMO Joint Transmitter and Noncoherent Receiver Design[J].IEEE Wireless Communication Letter,2019,8(1):177-180.DOI:10.1109/LWC.2018.2865563 DOI
6	BIRCANOGLU C,ARICA N.A Comparison of Activation Functions in Artificial Neural Networks [C]//Signal Processing and Communications Applications Conference (SIU),Izmir, Turkey,2018:1-4.DOI:10.1109/SIU.2018.8404724 DOI
7	HAMALAINEN A,HENRIKSSON J.Convolutional Decoding Using Recurrent Neural Networks [C]//International Joint Conference on Neural Networks.Washington, DC, USA,1999:3323-3327.DOI:10.1109/IJCNN.1999.836193 DOI
8	PAYANI A,FEKRI F.Decoding LDPC Codes on Binary Erasure Channels using Deep Recurrent Neural⁃Logic Layers [C]//IEEE International Symposium on Turbo Codes & Iterative Information Processing (ISTC),Hongkong, China,2018:1-5.DOI:10.1109/ISTC.2018.8625326 DOI
9	WANG J,MA Y,XUE S Y,et al.Parallel Decoding for Non⁃Recursive Convolutional Codes and its Enhancement Through Artificial Neural Networks [C]//IEEE International Workshop on Signal Processing Advances in Wireless Communications.Cannes, France,2019:1-1
10	LI A,MA Y,XUE S Y,et al.Unsupervised Deep Learning for Blind Multiuser Frequency Synchronization in OFDMA Uplink [C]//IEEE International Conference on Communications.Shanghai, China,2019:1-1.DOI:10.1109/ICC.2019.8761937 DOI
11	DE LUNA DUCOING J C,MA Y,YI N,et al.A Real⁃Complex Hybrid Modulation Approach for Scaling up Multiuser MIMO Detection [J].IEEE Transactions on Communications,2018,66(9):3916-3929.DOI:10.1109/TCOMM.2018.2833101 DOI
12	WOLNIANSKY P W,FOSCHINI G J,GOLDEN G D,et al.V⁃BLAST: An Architecture for Realizing Very High Data Rates over The Rich⁃scattering Wireless Channel [C]//URSI International Symposium on Signals, Systems, and Electronics.Conference Proceedings.Pisa, Italy,1998:295-300.DOI:10.1109/ISSSE.1998.738086 DOI
13	LIU T,LIU Y Y.Modified Fast Recursive Algorithm for Efficient MMSE⁃SIC Detection of the V⁃BLAST System [J].IEEE Transactions on Wireless Communications,2008,7(10):3713-3717.DOI:10.1109/T⁃WC.2008.070487 DOI
14	XUE S Y,MA Y,LI A,et al.On Unsupervised Deep Learning Solutions for Coherent MU⁃SIMO Detection in Fading Channels [C]//IEEE International Conference on Communications.Shanghai, China,2019:1-6.DOI:10.1109/ICC.2019.8761999 DOI
15	KENTON W.Overfitting [EB/OL]. [2019-07-02].
16	GEORGE D,HUERTA E A.Deep Neural Networks to Enable Real⁃time Multimessenger Astrophysics [J].Physical Review D,2018,97(4):044039.DOI:10.1103/PhysRevD.97.044039 DOI
17	DORNER S,CAMMERER S,HOYDIS J,et al.Deep Learning Based Communication Over the Air [J].IEEE Journal of Selected Topics in Signal Processing,2018,12(1):132-143.DOI:10.1109/JSTSP.2017.2784180 DOI
18	SAHOO D,PHAM Q,LU J,et al.Online Deep Learning: Learning Deep Neural Networks on the Fly [C]//International Joint Conferences on Artificial Intelligence Organization.Stockholm, Sweden,2018:2660-2666.DOI:10.24963/IJCAI.2018/369 DOI

To Learn or Not to Learn:Deep Learning Assisted Wireless Modem Design

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 18

相关文章 0

编辑推荐

Metrics