Space robotics and sustainability integrating physical, agentic and governance-aware (PAG) AI for autonomous orbital operations
Academic Article in Scopus
The increasing density of orbital debris elevates collision risk and drives the need for autonomous systems capable of operating under strict safety and sustainability constraints. This work introduces the Space Robotics and Sustainability (SRS) architecture, a deterministic and governance-aware decision system that embeds sustainability metrics and regulatory constraints directly within the maneuver-selection loop. Through this formulation, sustainability principles become measurable decision variables that govern maneuver admissibility, prioritization, and execution during delay-tolerant operations. This work also introduces the PAG AI framework, which is composed of Physical AI for constraint-aware interaction with orbital dynamics, Agentic AI for distributed decision-making and traceability, and Governance-Aware AI for embedding regulatory compliance within the control loop. The SRS architecture incorporates two complementary decision mechanisms to operationalize the PAG AI framework. The Sustainability Impact Score (SIS) evaluates candidate maneuvers using multi-criteria indicators, including mission safety, propulsion efficiency, regulatory compliance, and time performance. The Operational Relevance Index (ORI) prioritizes agent activation according to mission context, uncertainty, and risk contribution, enabling resource-aware coordination within the decision pipeline. These mechanisms produce structured and traceable autonomous behavior aligned with operational policies. Validation on the spacecraft near-distance rendezvous problem demonstrates that the SRS architecture generates stable approach trajectories with controlled velocity profiles and consistent adherence to safety constraints. Comparative evaluation against conventional guidance strategies highlights a distinct decision paradigm in which maneuver selection incorporates governance compliance and sustainability-driven evaluation as primary decision criteria. Monte Carlo simulations under uncertain initial conditions confirm high mission success rates and improved sustainability-oriented performance relative to nominal baseline strategies, while preserving full decision traceability through structured logs. The proposed architecture provides a foundation for policy-compliant autonomous systems in orbital operations. Its modular nine-layer structure supports scalability toward more complex debris-removal scenarios, enabling the integration of robotic autonomy, sustainability principles, and governance-aware decision-making in future space missions.
