Mixed Consistency Distributed State for Space Situational Awareness in Satellite Swarms
- Paper ID
105711
- authors
- company
Centre National d'Etudes Spatiales (CNES); ENAC; Institut de Recherche en Informatique de Toulouse (IRIT)
- country
France
- year
2026
- abstract
Space Situational Awareness (SSA), the detection, tracking, and cataloging of objects in orbit, is critical for satellite swarm operations in increasingly crowded orbital environments. Effective SSA requires swarm entities to share observation plans, measurements, and catalog updates while managing both cooperative and uncooperative objects. However, different SSA functions have fundamentally different requirements for state consistency and timeliness. A mixed consistency distributed state system tailored for distributed SSA applications is introduced. The system combines consensus mechanisms for strong consistency, Conflict-free Replicated Data Types for eventual consistency, and best-effort gossip dissemination for non-critical information sharing. By allowing SSA algorithms to specify only the consistency guarantees they require, the system reduces algorithm complexity and network load compared to one-size-fits-all approaches. This paper demonstrates how collision avoidance, trajectory prediction, observation planning, and catalog maintenance each benefit from different consistency models. The system is compared against the state-of-the-art implementation from NASA's Starling mission, showing how mixed consistency enables more efficient resource utilization across resource-constrained nodes while maintaining safety and robustness in sparse communication environments.