Satellite-ground integrated networks (SGINs) are regarded as promising architectures for sensing heterogenous measurements, reducing network congestion and for providing pervasive intelligence services in support of terrestrial users. In SGINs, both low, medium and geostationary orbit based satellites are deployed for achieving global coverage and for supporting communication services for terrestrial users. However, the integration of task sensing, computation, communication and caching functionalities is quite challenging, which leads to low real-time task processing capabilities in dynamically fluctuating complex network environments. Hence, we propose an edge-intelligence-driven collaborative SGIN architecture and construct a framework relying on multiple planes, including the sensing plane, forwarding plane, control plane, intelligence plane and application plane. Furthermore, we construct an integrated sensing, communication, computation, caching, and intelligence service paradigm for SGINs. Moreover, a centralized and distributed integrated learning framework is established for computation offloading and network function virtualization. Our simulation results show that the proposed learning framework is superior to the existing baseline methods in terms of its average transmission rate. Finally, we list a suite of potential research directions and solutions.

中文翻译：

---

星地一体化网络的智能传感、通信、计算和缓存


卫星地面综合网络（SGIN）被认为是用于感知异构测量、减少网络拥塞以及提供普遍的情报服务以支持地面用户的有前景的架构。在SGIN中，部署了低、中和地球静止轨道卫星，以实现全球覆盖并支持地面用户的通信服务。然而，任务感知、计算、通信和缓存功能的集成相当具有挑战性，这导致在动态波动的复杂网络环境中实时任务处理能力较低。因此，我们提出了一种边缘智能驱动的协作SGIN架构，并构建了一个依赖于多个平面的框架，包括感知平面、转发平面、控制平面、智能平面和应用平面。此外，我们为 SGIN 构建了一个集成的传感、通信、计算、缓存和智能服务范例。此外，还建立了集中式和分布式集成学习框架，用于计算卸载和网络功能虚拟化。我们的模拟结果表明，所提出的学习框架在平均传输率方面优于现有的基线方法。最后，我们列出了一套潜在的研究方向和解决方案。