A Novel Three-Layer GRU Equalizer for High-Speed PAM-4 Transmission Beyond 300Gbps
A novel three-layer gated recurrent unit neural network (TL-GRU) equalizer for high-speed intensity modulation direct-detection (IM/DD) PAM-4 systems is proposed. Based on this novel equalization scheme, the BER performance of the 200 Gbaud (400 Gbps) PAM-4 can be decreased into hard-decision forward error correction (HD-FEC) threshold 3.8× 10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−3</sup>, which is much better than least mean squares (LMS), polynomial nonlinear equalization (PNE) and gated recurrent unit (GRU) equalizers. The TL-GRU highlights its potential as a promising signal equalization solution for 1.6T and 3.2T short-reach optical interconnect systems.
Authors
- Jian Sun (ORCID: https://orcid.org/0000-0001-6530-9007)
- Jun Qin (ORCID: https://orcid.org/0000-0002-8168-5034)
- Jie Shi (ORCID: https://orcid.org/0000-0002-1760-0462)
- L. F. Chen
- Sun Yu (ORCID: https://orcid.org/0000-0002-5180-6608)
- Qian Wang (ORCID: https://orcid.org/0000-0002-8967-8525)
- Yueqin Li
- Junxiong Tan
- Junde Lu (ORCID: https://orcid.org/0009-0003-3072-6195)
Institutions
- Beijing Information Science & Technology University (CN)
Publication Details
- Published
- 2024-11-02
- DOI
- https://doi.org/10.1109/acp/ipoc63121.2024.10809964
- Primary Topic
- Coding theory and cryptography
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China