Unity-rate coding improves the iterative detection convergence of autoencoder-aided communication systems
Unity-rate coding improves the iterative detection convergence of autoencoder-aided communication systems
A forward error correction (FEC) and unity-rate coded (URC) autoencoder (AE)-assisted communication system is proposed for the first time, which relies on soft iterative decoding for attaining a vanishingly low error probability. The AE-demapper is specifically designed for directly calculating the extrinsic logarithmic likelihood ratios (LLRs), which can be directly entered into the URC decoder for soft iterative decoding. This avoids the potential degradation due to the conversion of symbol probabilities to bit LLRs. A comprehensive capacity analysis of the AE is performed, which demonstrates the capacity advantage of the AE-aided constellation design over its conventional quadrature amplitude modulation (QAM)/phase shift keying (PSK) counterpart. Furthermore, we carry out its EXtrinsic Information Transfer (EXIT) chart analysis, which indicates that as a benefit of our URC, the EXIT curve always reaches the [1,1] point of perfect convergence, leading to a vanishingly low error probability. More explicitly, our bit error ratio (BER) and block error ratio (BLER) results demonstrate that the proposed FEC-URC-AE system achieves significant iterative gains both in additive white Gaussian noise (AWGN) and Rayleigh channels, outperforming both its model-based FEC-AE and its conventional coded QAM/QPSK counterparts.
Codes, Communication systems, Decoding, EXtrinsic Information Transfer (EXIT) chart, Forward error correction, Iterative decoding, Receivers, Training, Unity-rate code (URC), autoencoder (AE), iterative detection and decoding
5037-5047
Xiang, Luping
56d951c0-455e-4a67-b167-f6c8233343b1
Xu, Chao
5710a067-6320-4f5a-8689-7881f6c46252
Zhang, Xiaoyu
ba39d478-405a-4f41-955b-300e464048b3
Luong, Thien Van
e95afcc7-65ce-46b2-a002-d36e24c90b97
Maunder, Robert
76099323-7d58-4732-a98f-22a662ccba6c
Yang, Lie-Liang
ae425648-d9a3-4b7d-8abd-b3cfea375bc7
Hanzo, Lajos
66e7266f-3066-4fc0-8391-e000acce71a1
1 May 2022
Xiang, Luping
56d951c0-455e-4a67-b167-f6c8233343b1
Xu, Chao
5710a067-6320-4f5a-8689-7881f6c46252
Zhang, Xiaoyu
ba39d478-405a-4f41-955b-300e464048b3
Luong, Thien Van
e95afcc7-65ce-46b2-a002-d36e24c90b97
Maunder, Robert
76099323-7d58-4732-a98f-22a662ccba6c
Yang, Lie-Liang
ae425648-d9a3-4b7d-8abd-b3cfea375bc7
Hanzo, Lajos
66e7266f-3066-4fc0-8391-e000acce71a1
Xiang, Luping, Xu, Chao, Zhang, Xiaoyu, Luong, Thien Van, Maunder, Robert, Yang, Lie-Liang and Hanzo, Lajos
(2022)
Unity-rate coding improves the iterative detection convergence of autoencoder-aided communication systems.
IEEE Transactions on Vehicular Technology, 71 (5), .
(doi:10.1109/TVT.2022.3151246).
Abstract
A forward error correction (FEC) and unity-rate coded (URC) autoencoder (AE)-assisted communication system is proposed for the first time, which relies on soft iterative decoding for attaining a vanishingly low error probability. The AE-demapper is specifically designed for directly calculating the extrinsic logarithmic likelihood ratios (LLRs), which can be directly entered into the URC decoder for soft iterative decoding. This avoids the potential degradation due to the conversion of symbol probabilities to bit LLRs. A comprehensive capacity analysis of the AE is performed, which demonstrates the capacity advantage of the AE-aided constellation design over its conventional quadrature amplitude modulation (QAM)/phase shift keying (PSK) counterpart. Furthermore, we carry out its EXtrinsic Information Transfer (EXIT) chart analysis, which indicates that as a benefit of our URC, the EXIT curve always reaches the [1,1] point of perfect convergence, leading to a vanishingly low error probability. More explicitly, our bit error ratio (BER) and block error ratio (BLER) results demonstrate that the proposed FEC-URC-AE system achieves significant iterative gains both in additive white Gaussian noise (AWGN) and Rayleigh channels, outperforming both its model-based FEC-AE and its conventional coded QAM/QPSK counterparts.
Text
TVT_URC_FINAL
- Accepted Manuscript
More information
Accepted/In Press date: 9 February 2022
Published date: 1 May 2022
Keywords:
Codes, Communication systems, Decoding, EXtrinsic Information Transfer (EXIT) chart, Forward error correction, Iterative decoding, Receivers, Training, Unity-rate code (URC), autoencoder (AE), iterative detection and decoding
Identifiers
Local EPrints ID: 455037
URI: http://eprints.soton.ac.uk/id/eprint/455037
ISSN: 0018-9545
PURE UUID: 3c514bbe-6f06-4447-ac87-a752877f33e2
Catalogue record
Date deposited: 04 Mar 2022 17:38
Last modified: 20 Apr 2024 01:49
Export record
Altmetrics
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
