Robotic systems deployed beyond controlled environments encounter abrupt and unobserved dynamics shifts that induce substantial transient performance loss, even when the nominal closed loop remains locally stabilizing. We formulate rapid inference-time recovery for physical AI as a constrained residual learning problem, in which an online correction added to a frozen reinforcement learning policy is regulated to minimize recovery time while remaining inside the robustness margin of the nominal controller. We specifically focus upon the regulation of residual's corrective authority relative to an already competent frozen policy. The learned residual produces bounded action-space corrections online, and a Stability Alignment Gate regulates this corrective authority through magnitude constraints, directional coherence, performance-conditioned regulation, and adaptive gain modulation. The method requires no environment resets, policy retraining, explicit system identification, or privileged shift information. Across mid-episode changes in actuation, mass, and contact conditions, it reduces recovery time relative to frozen SAC by up to \textbf{87\%} on Go1, \textbf{48\%} on Cassie, \textbf{30\%} on H1, and \textbf{20\%} on Scout in simulation, while preserving near-nominal steady-state performance. The framework is also characterized on two \emph{physical} robots, a Trossen ViperX manipulator and an AgileX Scout Mini, cutting recovery time by \textbf{19.5\%} on hardware without modifying onboard controller or accessing shift information.

翻译：暂无翻译