Managed Autonomy at Runtime: Gear-Based Safety and Governance for Single- and Multi-Agent Cyber-Physical Systems

The paper proposes a discrete-time runtime control system for managed autonomy in software and robotic agents. It combines five execution gears, utility-gated dispatch, and event-driven fallback. For single agents, the authors provide proofs covering monotonic stability, execution safety, eventual stabilization, fallback completeness, and equivalence to a gear-constrained Markov decision process. For multi-agent cyber-physical systems, the design maps runtime evidence to four governance states and adds consensus gating, swarm-level Lyapunov analysis, per-agent gear authority, and rendezvous control. In a three-UR5 robotic assembly cell, evaluated over 10,000 Monte Carlo episodes, it reports 99.6% anomaly detection versus 2.1% for a single-agent baseline and 3.5x lower detection latency.