Research on Hybrid Constellation Design Method for Key Target Collaborative Observation Based on Adaptive-Reference-Point-Adjusted NSGA-III

The rapid deployment of low-orbit mega-constellations has intensified the need for space situational awareness systems to transition from wide-area surveillance to persistent collaborative tracking of maneuvering targets. However, existing constellation design methods face three fundamental challenges: insufficient coverage overlap for multi-satellite collaborative observation, poor temporal continuity for long-arc tracking, and weak resilience against satellite failures. To address these bottlenecks, this paper proposes a hybrid constellation multi-objective optimization design method based on adaptive-reference-point-adjusted NSGA-III. The configuration adopts a Walker–Flower hybrid architecture, where the Walker constellation provides wide-area target search and the Flower constellation enables regional persistent tracking, forming a “search-track” cascaded collaborative mechanism. The optimization model integrates four competing objectives—observation duration, constellation cost, positioning accuracy, and performance decay slope—into a unified framework. To overcome the limitations of classical NSGA-III, three improvement strategies are introduced: adaptive reference point adjustment, constraint-dominated sorting, and mixed-integer encoding. Simulation results over a 24 h LEO target observation scenario with 168 background satellites demonstrate that the proposed algorithm achieves an average observation duration of 33.7 h across four maneuvering targets, outperforming classical NSGA-III by 54.8%, while improving positioning accuracy by 97.7% and resilience by 45.5%. All recommended solutions strictly satisfy engineering constraints, whereas comparison algorithms retain infeasible solutions due to the penalty function approach. The proposed method provides a practical and robust engineering solution for LEO key target collaborative observation constellation design, with demonstrated advantages in tracking continuity, positioning accuracy, and system resilience.