Editorial: Satellite Terrestrial Networks: Technologies, Security and Applications

The ongoing evolution toward sixth-generation (6G) communications necessitates the deep integration of terrestrial and non-terrestrial infrastructures into the seamless Space–Air–Ground–Sea Integrated Network. Low Earth orbit (LEO) mega-constellations, combined with advanced terrestrial networks, offer unprecedented opportunities for global coverage, resilient connectivity, and low-latency services. However, this paradigm also introduces formidable challenges in terms of spectrum efficiency, dynamic resource management, mobility support, security, and end-to-end quality of service. This Special Issue of Electronics, “Satellite Terrestrial Networks: Technologies, Security and Applications”, brings together twelve research contributions that collectively advance the state of the art in these domains.

Recent research has highlighted that STNs will play a decisive role in the evolution from 5G toward 6G networks, where the integration of non-terrestrial components is no longer optional but rather a fundamental design principle [1]. A number of survey works have emphasized that efficient spectrum sharing and interference management remain among the most critical technical challenges [2]. Furthermore, the rapid growth of low Earth orbit (LEO) mega-constellations has accelerated the need for scalable routing and resource allocation schemes capable of coping with highly dynamic topologies [3,4].

In addition to communication aspects, computing and intelligence are increasingly being integrated into STNs. The paradigm of edge computing, initially studied in terrestrial mobile networks, has been extended to LEO satellite systems, enabling low-latency task execution and real-time data processing [5]. Meanwhile, machine learning methods, including reinforcement learning and graph neural networks, have demonstrated their potential in optimizing resource allocation, routing, and beam management in dynamic environments [6,7]. Signal processing research has also advanced toward robust synchronization and anti-jamming techniques, which are crucial for ensuring reliability in contested environments [2].

Security and resilience are equally pressing issues in STNs. Beyond conventional cryptographic measures, recent studies have proposed physical-layer security mechanisms and frequency-hopping strategies to defend against jamming and eavesdropping [8]. At the system level, the rise of cloud-native platforms and network function virtualization provides new opportunities for the flexible deployment and orchestration of communication and computing functions across satellite and terrestrial segments [9].

Against this background, the twelve studies included in this Special Issue provide concrete solutions and experimental validations that align with these global research trends. Collectively, they offer both theoretical insights and practical frameworks that bring satellite–terrestrial networks closer to deployment in real-world scenarios.

Motivated by the growing role of artificial intelligence in communication networks, Sun et al. (Contributor 1) introduce SCNOC-Agentic, an LLM-driven framework for satellite network operation and control. The system integrates intent refinement, multi-agent workflows, long-term memory, and graph-based retrieval, achieving significant improvements in network task planning, fault analysis, and resource optimization. Zhang et al. (Contributor 2) address the resource management problem in beam-hopping-based satellite systems, proposing a progressive decomposition framework for joint beam scheduling and power/frequency allocation. Their approach, supported by geographic isolation mechanisms, ensures efficient coexistence between GEO and LEO systems while enhancing throughput. Song et al. (Contributor 3) tackled the challenge of Doppler frequency-offset estimation in LEO systems. By leveraging beam pointing information instead of GNSS support, the authors developed an accurate and lightweight estimation method suitable for highly dynamic non-terrestrial networks.

Machine learning-based methods are also investigated for interference detection in GNSS. Baldini and Bonavitacola (Contributor 4) evaluate the impact of signal bit-depth on the accuracy of wireless interference classification, demonstrating the trade-offs between hardware sampling constraints and ML model performance. Edge computing and task offloading are studied by Li et al. (Contributor 5), who propose a joint optimization framework for task scheduling and communication resource allocation in LEO satellite edge networks. Their results show improvements in energy efficiency, task completion rates, and overall system cost. Pan et al. (Contributor 6) focus on burst spread spectrum signals and introduced a novel signal structure that eliminates synchronization overheads while improving acquisition reliability, stealth, and resource efficiency.

The coexistence problem between terrestrial IMT base stations and satellite services is examined by Jia et al. (Contributor 7). Through a power control and intervention algorithm combined with deep reinforcement learning, they demonstrate how harmful interference can be mitigated while preserving terrestrial throughput. Wang et al. (Contributor 8) contribute to routing in satellite self-organizing networks, proposing SQL-CBRP, a cluster-based routing protocol evaluated on the OMNeT++ platform. Their results show reduced delay and packet loss compared to classical methods under high loads. Han et al. (Contributor 9) further advance routing optimization by introducing a deep reinforcement learning-based multipath routing algorithm for LEO mega-constellations. Their multipath discovery and graph-neural-network-based traffic scheduling achieve considerable improvements in throughput, reliability, and delay over shortest-path schemes.

In the context of beam management, Liu and Pan (Contributor 10) investigate optimal beamwidth settings for quasi earth-fixed cells in LEO systems, showing that the dynamic adjustment of beamwidth maximizes uplink coverage probability and outperforms 3GPP baseline schemes. Anti-jamming mechanisms are studied by Yu et al. (Contributor 11), who propose a capacity enhancement method for frequency-hopping systems. By adapting hopping rates, the method improves capacity without compromising jamming resistance. Finally, Shi et al. (Contributor 12) introduce ComEdge, a cloud-native platform for integrated computing and communication in satellite–terrestrial networks. By employing microservices, containers, and service mesh technologies, the platform provides a flexible environment for real-world deployment and demonstrates the feasibility of AI-driven network resource management.

The diversity of these studies highlights the breadth of ongoing research in STNs. Several themes are evident across the contributions: the need for intelligent automation through machine learning and AI, the challenge of efficient coexistence between heterogeneous systems, the focus on robustness against interference and jamming, and the shift toward practical, deployment-ready platforms. Nonetheless, significant challenges remain, including the validation of the proposed solutions under real-world conditions, the balance between AI-driven efficiency and explainability, the energy constraints of spaceborne platforms, and the adaptation of algorithms to time-varying and uncertain environments. Addressing these challenges will be crucial for building scalable, secure, and efficient satellite–terrestrial networks that are capable of supporting next-generation applications.