基于终端流量预测的低地球轨道卫星互联网资源分配策略
LEO satellite Internet resource allocation strategy based on terminal traffic prediction
沈斐 1吕承丞 2张嘉璇 2阮小婷2
作者信息
- 1. 中国科学院上海微系统与信息技术研究所无线传感网与通信重点实验室,上海 200050
- 2. 中国科学院上海微系统与信息技术研究所无线传感网与通信重点实验室,上海 200050;中国科学院大学,北京 101408
- 折叠
摘要
针对地面网络存在覆盖盲区和卫星网络通信资源利用率低等问题,提出了基于终端流量预测的低地球轨道(LEO)卫星互联网资源分配策略.该策略利用真实数据集提出改进LSTM-ARIMA算法,准确预测地面区域未来一定时间内产生的数据流量,通过Stackelberg博弈构建差分化数据传输和任务卸载2种通信模型.综合考虑数据处理时延和能耗,通过求解纳什均衡,得到用户通过LEO卫星互联网传输数据或卸载任务的最优比率,以及卫星提供网络服务的最优定价.仿真结果表明,所提策略在数据传输服务中收益能提高约40%,在任务卸载服务中收益能提高约50%.
Abstract
A resource allocation strategy for the low earth orbit(LEO)satellite Internet based on terminal traffic predic-tion was proposed to address the problems of blind coverage spots in ground network and the low resource utilization of satellite network.An improved LSTM-ARIMA algorithm was proposed with real datasets by the strategy to accurately predict the data traffic generated in the ground area over a certain period of time in the future.Two communication mod-els,differentiated data transmission and task offloading were constructed through Stackelberg games,taking into the data processing latency and energy consumption account.By solving the Nash equilibrium,the optimal ratio for users to trans-mit data or unload tasks through the LEO satellite Internet,as well as the optimal pricing for satellites to provide network services,were obtained.Extensive simulation results verify that the proposed strategy can increase the revenue by ap-proximately 40%in data transmission services and 50%in task offloading services.
关键词
低地球轨道卫星互联网/数据流量预测/资源分配/数据传输/任务卸载Key words
LEO satellite Internet/data flow prediction/resource allocation/data transmission/task offloading引用本文复制引用
基金项目
国家重点研发计划基金资助项目(2019YFE0120700)
上海市科技创新行动计划基金资助项目(22511100800)
上海市科技创新行动计划基金资助项目(22511100500)
出版年
2024
国家科技期刊平台
NSTL
万方数据