Drift-Aware Autonomous Collision Avoidance Architecture for Future Satellite Systems

Abstract
The rapid expansion of satellite constellations and the continued accumulation of orbital
debris have significantly increased the risk of satellite–satellite and satellite–debris
collisions. While modern Space Situational Awareness (SSA) systems provide increasingly
accurate tracking and conjunction warnings, collision prevention remains largely advisory
and dependent on human-in-the-loop decision processes. This approach introduces
latency, inconsistency, and poor scalability in dense orbital environments.
This paper proposes a drift-aware, autonomous collision avoidance architecture designed
to operate directly onboard future satellite systems. The framework integrates dynamic
proximity safety envelopes, layered detection and sensing, corrected drift modeling for
improved orbital prediction, autonomous maneuver decision authority, and cooperative
behavior between maneuverable satellites. By explicitly accounting for environmental drift
and accumulated prediction error, the system enables earlier, lower-energy avoidance
maneuvers while reducing uncertainty-driven false alarms.
The proposed architecture shifts collision prevention from passive warning to active,
autonomous execution, providing a scalable foundation for safe orbital operations as
satellite populations continue to grow. This approach offers a path toward sustainable
space traffic management before collision cascades impose irreversible consequences on
the near-Earth orbital