Robotics Governance Architecture (RGA): A Non-Bypassable Control Layer for Safe Autonomous Systems

This paper introduces the Robotics Governance Architecture (RGA), a unified control architecture for autonomous systems that enforces non-bypassable execution constraints through integrated execution control and runtime safety mechanisms. As autonomous systems increase in capability, existing approaches based on policy enforcement and post-execution monitoring reveal structural limitations. These approaches fail to ensure that control is intrinsically bound to execution, leading to gaps between intended and actual system behavior. RGA addresses this challenge by embedding governance directly into system architecture. The framework integrates capability governance, execution control, and runtime safety enforcement into a cohesive model that enables bounded, verifiable, and continuously validated system behavior. The architecture establishes a closed-loop interaction between control and execution, ensuring that all system actions remain aligned with defined operational and safety constraints and cannot bypass validation mechanisms. By shifting control from an external supervisory function to an intrinsic system property, RGA enables more robust, transparent, and regulation-aligned autonomous systems. This work provides a conceptual and architectural foundation for the design, analysis, and deployment of next-generation autonomous systems, particularly in safety-critical and distributed environments. This paper is part of the Robotics Governance Architecture (RGA) research series, covering capability governance, non-bypassable execution control, and runtime safety enforcement across multiple companion papers.