arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.01972cs.ITcs.LGmath.IT
收藏

混合神经编码调制:设计与训练方法

Hybrid Neural Coded Modulation: Design and Training Methods

Hallym University(翰林大学);KIOST(韩国海洋科学技术院);Hanyang University(汉阳大学)

Sung Hoon Lim, Jiyong Han, Wonjong Noh, Yujae Song, Sang-Woon Jeon

摘要:

提出了一种由内码和外码组成的混合编码调制方案。外码可以是具有有效软解码能力的任何标准二进制线性码(例如,低密度奇偶校验(LDPC)码)。内部代码使用深度神经网络(DNN)来设计,深度神经网络(DNN)获取信道编码比特并输出调制符号。为了训练DNN,我们建议使用受广义互信息启发的损失函数。所得到的星座图被示出优于具有5G标准LDPC码的调制阶数16和64的传统的基于正交幅度调制(QAM)的编码方案。

英文摘要:

We propose a hybrid coded modulation scheme which composes of inner and outer codes. The outer-code can be any standard binary linear code with efficient soft decoding capability (e.g. low-density parity-check (LDPC) codes). The inner code is designed using a deep neural network (DNN) which takes the channel coded bits and outputs modulated symbols. For training the DNN, we propose to use a loss function that is inspired by the generalized mutual information. The resulting constellations are shown to outperform the conventional quadrature amplitude modulation (QAM) based coding scheme for modulation order 16 and 64 with 5G standard LDPC codes.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑