Recently, the Space-Air-Ground-Sea Integrated Network (SAGSIN) attracts great attention due to its ability to provide high-speed communication services to aviation users (AUs) and marine users (MUs), with Low Earth Orbit (LEO) satellites play an essential role. However, available LEO space is nearly saturated and full of massive space junks, which, combined with the ultra-low latency requirements for future 6 G, presents a significant challenge. To address this issue, we propose designing a Very Low Earth Orbit (VLEO)-based satellite network that efficiently serves AUs and MUs. We first create heat maps based on the actual collected data and generate a benchmark observation point model with grid point method. Then we propose an implicit multi-objective continuous multivariate optimization problem to achieve the maximum coverage with minimum VLEO satellites. To solve this problem, we build a constellation simulation system, using the idea of decomposition and polymerization combined with the elite strategic genetic algorithm (ESGA) of swarm intelligence optimization algorithm. Many simulation results are obtained, including the indication that the optimal VLEO constellation has the deployment features of large altitude and low inclination, and has better coverage performance for longitudinal distributed business. The constellation design process is also highly migratory for other non-terrestrial networks.

Deep space communication systems play a key role in human endeavors for lunar basing, Mars, and further cosmic exploration. In pursuit of establishing the Moon as a deep space Internet portal for future human exploration, a lunar communication and navigation integrated (CNI) satellite constellation design is intended into consideration. Based on Fibonacci lattice virtual observation point model, featuring better uniformity and stochasticity, we derive objectives related to access coverage, power, quadruple coverage, and geometric dilution of precision (GDOP), together with multiple multi-objective optimization problems (MOPs) combined with Earth-Moon difference analysis to explore better utilization of overall resources in the lunar integrated constellation. The pareto model, non-dominated sorting genetic algorithm-II (NSGA-II), and the construction of the constellation system are incorporated into the solution to pursue a higher guiding value. In particular, inspired by the lunar frozen orbit (LFO), the hybrid inclination (HyInc) Walker configuration is proposed with theoretical validation and simulation evaluation, showing some superiority over traditional Walker and remains generalizable. Extensive simulation and comprehensive analysis are performed, including the pareto-optimal integrated constellations and the revelation of HyInc Walker’s coverage equalization capability, with the latter being less studied. The entire constellation design process of this work is highly migratory and the proposed perspective of the configuration is enlightening.

The alternative strategy for future space-based satellite (SAT) internet development involves a potential transition from single-layer low Earth orbit (LEO) system to sophisticated heterogeneous multi-layer constellations (HMLCs). HMLC integrates SATs from different orbital layers, including very low Earth orbit, LEO, medium Earth orbit, and geostationary orbit, potentially incorporating communication, navigation, and remote sensing (CNR) capabilities, with the ability to enhance coverage, communication capacity, geometric dilution of precision, resilience, resolution, etc., then offering the diversification of service functions. Realizing this potential hinges on multi-SAT collaboration driven by key technologies including inter-satellite link, advanced antenna technology, on-board processing, topology-aware routing, and artificial intelligence. Designing these sophisticated systems inevitably involves multi-objective optimization problem (MOP) to trade-off conflicting performance metrics, especially in optimizing integrated CNR services. In addition, HMLC’s structural flexibility and integration capabilities also provide an important foundation for future mission configurations extending into deep space. This article explores the HMLC framework, enabling technologies, and the MOP paradigm. Furthermore, it outlines a vision towards intelligent and agile HMLCs, emphasizing future directions in rapid constellation assessment using surrogate models and dynamic in-orbit adaptation to meet evolving demands, paving the way for building adaptive and efficient intelligent space-based networks to support future deep space missions such as Earth-Moon collaboration.