The escalating demand for global connectivity has driven Low Earth Orbit (LEO) constellation networks to become indispensable components of next-generation communication systems. As these networks evolve to support massive device connections, they generate unprecedented volumes of access network traffic requests, making accurate access network traffic prediction critically important. However, the online prediction of traffic presents significant challenges. These challenges stem from the lack of real-time correlation information, limited satellite resources, and the influence of sudden traffic changes. Due to the regular layout and predictable movement of the LEO constellation, the traffic of adjacent satellites within the same orbit plane has similar trends and strong correlations. Furthermore, the real-time traffic information can be shared between adjacent satellites via Inter-Satellite Links (ISL). Inspired by the above, this paper, for the first time, proposes an Adjacent Satellite Correlation-based Single-Hop Prediction (ASC-SHP) method for LEO constellation network online traffic prediction. This method explores the traffic correlations between adjacent satellites and employs neural networks to extract and renew the spatiotemporal deviation features of traffic between adjacent satellites autonomously. Simulation results validate that the proposed ASC SHP method reduces prediction errors by over 82% on average compared to baseline methods, with performance improving as the constellation size increases. Further simulations show that ASC-SHP remains effective under ISL instability and satellite failure conditions, while reducing computational complexity by over 80% compared to baseline methods.

Beam hopping (BH) is crucial for managing communication resources in large-scale multi-satellite, multi-beam Low Earth Orbit (LEO) constellation networks. However, achieving high-dynamic multi-satellite cooperative BH in LEO mega-constellations while satisfying multi-dimensional resource allocation demands remains a significant challenge. To address the interdependence of long-term coverage stability and short-term resource volatility, this paper proposes a distributed Beam Hierarchical Multi-Agent Reinforcement Learning (HMARL) framework. Within this framework, a hierarchical Duplex Dueling Multi-agent Q-learning (QPLEX) architecture decouples the high-dimensional joint optimization problem into two tractable layers: satellite-level cell association and beam-level resource scheduling. The former handles coverage planning under load-balancing and interference constraints, while the latter manages dynamic resource allocation amid spatiotemporal traffic variations. Additionally, Chemical Reaction Optimization (CRO) and Cross-Layer Attention (CLA) mechanisms are introduced to address the discrete beam interference problem and the inter-layer information isolation challenge, respectively. Simulation results show that the proposed method, BH-HMARL, performs better than baseline methods. Further results confirm its scalability, with performance remaining stable despite variations in the number of satellites, beams, and cells. Moreover, BH-HMARL achieves real-time scheduling with execution time under 5ms per agent, satisfying slot duration constraints.

The traffic prediction method for Low Earth Orbit (LEO) satellite-based Internet of Things (IoT) provides prior conditions for addressing challenges in resource allocation and management of LEO satellite-based IoT. However, traditional methods rely on the pre-generation of terrestrial centers, leading to substantial computational complexity and delays when applying prediction results to LEO satellite-based IoT. Therefore, achieving autonomous LEO satellite-based IoT traffic prediction is essential. Unlike the spatiotemporal features of terrestrial IoT, LEO satellite-based IoT traffic is influenced by both satellite dynamics and terrestrial user behavior. Fortunately, real IoT traffic information can be shared between adjacent satellites through Inter-satellite Links (ISL). Inspired by the above, this paper proposes, for the first time, an autonomous LEO satellite-based IoT traffic prediction method. A Spatiotemporal Fusion Neural Network (SFNN) employing a four-layer hybrid neural network architecture is designed to accommodate the unique spatiotemporal features of LEO satellite-based IoT traffic. The simulation results demonstrate that the proposed method has low computational complexity and achieves a better prediction performance than the five baseline methods in LEO constellations. Further simulations demonstrate that the proposed method achieves higher accuracy in large-scale LEO constellations, and the algorithm remains applicable under ISL instability and satellite failure conditions.

The Shortest Distance Path (SDP) problem is a critical routing issue in communication networks, particularly in satellite networks. Typically, SDP is solved by graph-based iterative algorithms, while an explicit or analytic approach is challenging. However, considering the orbit dynamics and topology regularity, this paper proposes, for the first time, an explicit analytic phase-based algorithm STEPCLIMB to directly solve the SDP in low-Earth orbit (LEO) satellite networks. Based on the relationship between satellite phase and inter-satellite link distance, the SDP is modeled with the satellite phase, and SDP problem is converted into a total phase offset problem through theoretical derivations. Then STEPCLIMB is derived in two cases, respectively. Monte-Carlo simulations verify STEPCLIMB’s accuracy, which has zero error in the mono-valley case and has less than 0.1% error in the bi-valley case. The algorithm performs better in larger-scale constellations and can save over 99.4% computational cost compared to Dijkstra algorithm. Further, the SDP pattern and features in Starlink constellation are analyzed. The model proves that most inter-plane hops in the SDP occur successively, and the simulations further indicate that these hops prefer satellites in the higher latitude regions.