Work

Optimal multi-objective configuration of a hybrid walker-star/delta navigation satellite constellation in LEO via MOPSO and reinforcement learning

Year: 2026
Type: article
Abstract: Optimal design of satellite constellations in Low Earth Orbit (LEO) for global navigation and positioning applications is of significant importance. This study presents a design framework for LEO satellite constellations based on hybrid Walker-Star/Delta geometric configurations, with the objective of simultaneously minimizing Geometric Dilution of Precision (GDOP) and deployment costs. A multi-objective optimization approach is employed by integrating Multi-Objective Particle Swarm Optimization (MOPSO) with Reinforcement Learning (RL) to adaptively tune key optimization parameters and the resulting RL-MOPSO algorithm is validated against conventional MOPSO and NSGA-II on standard benchmark problems. Satellite manufacturing and launch costs are estimated using a combined SSCM/USCM8 cost model and the design tool is further validated against the operational GLONASS constellation. The proposed hybrid configuration is evaluated against conventional Walker-Delta and Walker-Star configurations across multiple orbital altitudes. Results show that the optimized hybrid configuration achieves a notable improvement in mean GDOP compared with both Walker-Delta and Walker-Star configurations, while maintaining a comparable number of satellites and avoiding any significant increase in deployment cost. These findings demonstrate that the proposed RL-MOPSO based hybrid configuration is an effective approach for designing high-accuracy, cost-efficient LEO navigation constellations for future global navigation and precise positioning applications.
Language: English
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7234264358
FWCI:
Cites: 50
Cited by:
Related to: 10