移动通信2025，Vol.49Issue(6)：123-127,141,6.DOI:10.3969/j.issn.1006-1010.20250424-0003
面向大规模低轨星座的用户接入方法研究
User Access Method for Mega-Constellation Networks
摘要
Abstract
Mega-constellation networks composed of low Earth orbit(LEO)satellites serve as the critical infrastructure for 6G space-air-ground integrated networks,offering ubiquitous Internet access services for terrestrial users through global seamless communication coverage.To address the access selection challenges arising from the prevalent coverage overlap phenomenon in mega-constellation networks,this study formulates a weighted channel capacity maximization optimization problem to select the optimal access satellite from numerous available satellite nodes.During the user access selection phase,the quality-of-service requirements are comprehensively considered to achieve a balance between maximizing network throughput and ensuring service quality.Building upon this framework,we propose a joint optimization algorithm integrating graph attention mechanisms with deep reinforcement learning.Simulation results demonstrate that,compared with traditional random access schemes and deep reinforcement learning-based solutions,the proposed algorithm achieves significant improvements in both network throughput and access quality under Starlink constellation simulation scenario.关键词
巨型星座网络/非地面网络/资源优化/图神经网络/深度强化学习Key words
mega-constellation networks/non-terrestrial network/resource optimization/graph neural network/deep reinforcement learning分类
信息技术与安全科学引用本文复制引用
杨君一,林琦彬,范珂欣,安丽荣,郝中汉,张钦宇..面向大规模低轨星座的用户接入方法研究[J].移动通信,2025,49(6):123-127,141,6.基金项目
国家自然科学基金"深空通信信道模拟仿真器的研制"(62027802) （62027802）
国家自然科学基金"面向载人登月的空间信息传输理论与关键技术"(61831008) （61831008）
深圳市科技计划项目"基于硅基芯片的极高频段高效射频前端技术研究"(KQTD20210811090116029) （KQTD20210811090116029）
