【论文阅读】LLM4CP: Adapting Large Language Models for Channel Prediction(2024)
摘要
Channel prediction(信道预测) is an effective approach(有效方法) for reducing the feedback(减少反馈) or estimation overhead(估计开销) in massive multi-input multi-output(大规模多输入输出) (m-MIMO) systems. However, existing channel prediction methods(现有的信道预测方法) lack precision(缺乏精度) due to(由于) model mismatch errors(模型失配误差) or network generalization issues(网络泛华问题). Large language models (LLMs) have demonstrated(显示出) powerful modeling(强大的建模) and generalization abilities(泛化能力), and have been successfully applied(应用) to cross-modal tasks(跨模态任务), including the time series analysis(时间序列分析). Leveraging the expressive power of LLMs(利用LLM的表达能力), we propose a pre-trained(预训练) LLM-empowered channel prediction(信道预测) (LLM4CP) method