Staggered Integral Online Conformal Prediction for Safe Dynamics Adaptation with Multi-Step Coverage Guarantees

2026-04-07

AI Summary

This work addresses the challenge of quantifying model uncertainty in adaptive dynamical systems when state derivative measurements are unavailable, a setting where conventional online conformal prediction often yields overly conservative safety guarantees. To overcome this limitation, the authors propose Staggered Integral Online Conformal Prediction (SI-OCP), which introduces, for the first time, an integral-based scoring mechanism into online conformal prediction. By explicitly accounting for the cumulative effects of disturbances and learning errors, SI-OCP achieves valid multi-step coverage guarantees without requiring state derivatives. Integrated with robust tube-based model predictive control, the method ensures long-term safety of the closed-loop system. Empirical validation on a quadrotor simulation driven by a fully connected deep neural network demonstrates that SI-OCP effectively balances high safety assurance with strong control performance under complex learned dynamics.

Abstract

Safety-critical control of uncertain, adaptive systems often relies on conservative, worst-case uncertainty bounds that limit closed-loop performance. Online conformal prediction is a powerful data-driven method for quantifying uncertainty when truth values of predicted outputs are revealed online; however, for systems that adapt the dynamics without measurements of the state derivatives, standard online conformal prediction is insufficient to quantify the model uncertainty. We propose Staggered Integral Online Conformal Prediction (SI-OCP), an algorithm utilizing an integral score function to quantify the lumped effect of disturbance and learning error. This approach provides long-run coverage guarantees, resulting in long-run safety when synthesized with safety-critical controllers, including robust tube model predictive control. Finally, we validate the proposed approach through a numerical simulation of an all-layer deep neural network (DNN) adaptive quadcopter using robust tube MPC, highlighting the applicability of our method to complex learning parameterizations and control strategies.

Problem

• online conformal prediction
• model uncertainty
• safe dynamics adaptation
• coverage guarantees
• adaptive control

Innovation

• Online Conformal Prediction
• Integral Score Function
• Safety-Critical Control
• Robust Tube MPC
• Dynamics Adaptation