Dynamic LSTM-Based Channel Quality Indicators Prediction for LEO Satellite Communication

doi:10.1109/LWC.2024.3414591

	Dynamic LSTM-Based Channel Quality Indicators Prediction for LEO Satellite Communication
	Li, Hongguang 1,2,3,4; Sui, Yuze 1,2,3,4; Liu, Yaoqi 1,2,3,4; Zhou, Yiqing 1,2,3,4; Shi, Jinglin 1,2,3,4
	2024-12-01
发表期刊	IEEE WIRELESS COMMUNICATIONS LETTERS
ISSN	2162-2337
卷号	13 期号:12 页码:3285-3289
摘要	The link environment of low Earth orbit (LEO) satellite communication system is time-varying due to the fast movement of the satellite, and the delay of the downlink makes channel quality indication (CQI) parameter fed back by the terminal is out of time, reducing the spectrum utilization. In this letter, we present a novel model based on dynamic long short-term memory (DLSTM) for CQI prediction. DLSTM consists of two LSTMs working in parallel. It updates LSTMs parameters by training alternately to achieve better matching of environmental features. During the entire visible window, the simulation results show that DLSTM is 63.78% lower than baseline's CQI root mean square error (RMSE). Predicting 1000 steps, DLSTM is 20.29% lower than Autoregressive integrated moving average (ARIMA), support vector regression (SVR) and LSTM on average. The cost of DLSTM is training time for updating parameters continuously. DLSTM shows the effect of LSTM updating parameters online, making it more practical.
关键词	Signal to noise ratio Interference Long short term memory Predictive models Training Low earth orbit satellites Hidden Markov models LEO LSTM CQI online training
DOI	10.1109/LWC.2024.3414591
收录类别	SCI
语种	英语
资助项目	Jiangsu Key Research and Development Programof China[BE2021013-2] ; Innovation Fundingof ICT, CAS[E261020] ; CAS Specific Research Assistant Funding Program
WOS研究方向	Computer Science ; Engineering ; Telecommunications
WOS类目	Computer Science, Information Systems ; Engineering, Electrical & Electronic ; Telecommunications
WOS记录号	WOS:001375692100006
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/41125
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Liu, Yaoqi
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, SKLP, Beijing 100190, Peoples R China 3.Univ Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 4.Zhongke Nanjing Mobile Commun & Comp Innovat Inst, Beijing 211135, Peoples R China
推荐引用方式 GB/T 7714	Li, Hongguang,Sui, Yuze,Liu, Yaoqi,et al. Dynamic LSTM-Based Channel Quality Indicators Prediction for LEO Satellite Communication[J]. IEEE WIRELESS COMMUNICATIONS LETTERS,2024,13(12):3285-3289.
APA	Li, Hongguang,Sui, Yuze,Liu, Yaoqi,Zhou, Yiqing,&Shi, Jinglin.(2024).Dynamic LSTM-Based Channel Quality Indicators Prediction for LEO Satellite Communication.IEEE WIRELESS COMMUNICATIONS LETTERS,13(12),3285-3289.
MLA	Li, Hongguang,et al."Dynamic LSTM-Based Channel Quality Indicators Prediction for LEO Satellite Communication".IEEE WIRELESS COMMUNICATIONS LETTERS 13.12(2024):3285-3289.