Demand-Driven LEO Satellite Constellation for Dynamic Satellite Internet of Things

Low-earth-orbit (LEO) satellite constellation design aims to achieve full coverage of the Satellite Internet of Things (SIoT) by optimizing satellite orbits across regions. The time-varying channel conditions pose significant challenges to ensuring communication quality at all times, whereas existing solutions do not simultaneously consider dynamic communication demands and channel condition changes. To tackle these challenges, this paper proposes an LEO satellite constellation design scheme that takes into account dynamic communication service demands and channel conditions. Initially, a dynamic SIoT model based on LEO satellite network coverage is established within the regions of interest. Subsequently, models are constructed using actual population density and atmospheric condition data to calculate varying channel capacities across different regions. With this foundation, the problem of optimizing the utilization of LEO satellite channel capacity while ensuring communication quality at all times is proposed. To solve this multi-objective and multi-continuous variable optimization problem, the Hybrid Tabu-Weighted Evolutionary (HTWE) algorithm which incorporates tabu and scoring strategies is designed to rapidly obtain the optimal satellite orbit configuration. Compared to traditional algorithms, experimental results demonstrate that the HTWE algorithm improves the convergence speed by over 68% and enhances satellite communication quality in target regions by over 7% in a dynamic SIoT network environment.