Federated Learning-Based Codebook Design for Massive MIMO Communication System

Jing Jiang, Runqiu Han, Ben Liu,Dan Feng

Advances in Natural Computation, Fuzzy Systems and Knowledge DiscoveryLecture Notes on Data Engineering and Communications Technologies(2022)

引用 0|浏览0
暂无评分
摘要
Codebook design is critical to acquire the performance gains of massive MIMO wireless communication systems. However, the previous codebook designs did not consider the changeable wireless environment. Inspired by the processing capabilities for big data of federated learning, we propose a federated learning (FL) based codebook design scheme. Model training only collects user’s gradient. We design a convolutional neural network in which the input is channel data and the codebook is generated at the output. Simulation results show that the achievable rate of the proposed FL-based codebook design scheme is better than the traditional codebook design scheme.
更多
查看译文
关键词
massive mimo communication system,codebook design,learning-based
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要