AI 中文总结
针对地月飞行中地面站切换导致的遥测中断问题，通过构建冲突图、R-SCG模型等提出弹性频谱调度方法，仿真显示可恢复2.7小时遥测吞吐量。
AI 中文摘要
地球深空网络（DSN）终端之间的地面站切换每次事件会导致载波锁定重新获取延迟4至8分钟，在地月飞行期间引发周期性遥测中断。我们通过将时间切换序列建模为完美区间冲突图来解决这一限制。对于单航天器轨迹，我们确定添加1个备用信道（k=4）可实现静态离线信道预分配，消除地面站切换期间航天器应答器的重新调谐。为在意外大气衰落或地面站中断下维持链路可用性，我们构建了鲁棒空间通信图（R-SCG）模型，证明4个信道足以在单节点故障条件下维持不间断覆盖。对于未建模的链路中断，我们实现了摊销复杂度为O(1)的分布式局部重着色方案，并辅以随机森林状态分类用于预测性信道重新分配。此外，对于共址航天器集群，我们证明联合地空冲突图保持弦性，将所需频谱限制为c个并发资产对应3+c个信道。10天轨道仿真表明，与传统反应式切换方案相比，静态信道预分配可恢复多达2.7小时的遥测吞吐量。
英文摘要
Ground station handovers across Earth based Deep Space Network (DSN) terminals incur carrier lock reacquisition delays of 4 to 8 minutes per event, introducing periodic telemetry blackouts during cislunar flights. We address this limitation by formulating temporal handover sequences as perfect interval conflict graphs. For single spacecraft trajectories, we establish that adding a single reserve channel ($k=4$) enables static offline channel preallocation, eliminating spacecraft transponder retuning during station switches. To sustain link availability under unexpected atmospheric fades or station outages, we construct a Robust Space Communication Graph (R-SCG) model, demonstrating that four channels suffice to maintain uninterrupted coverage under single node failure conditions. For unmodeled link disruptions, we implement a distributed local recoloring scheme with $O(1)$ amortized complexity, augmented by Random Forest state classification for predictive channel reassignment. Furthermore, for colocated spacecraft clusters, we show that the joint ground space conflict graph retains chordality, bounding the required spectrum to $3+c$ channels for $c$ concurrent assets. A 10 day orbital simulation demonstrates that static channel preassignment recovers up to 2.7~hours of telemetry throughput relative to conventional reactive handover schemes.
Comments19 pages, 13 figures
