Safety Beyond the Training Data: Robust Out-of-Distribution MPC via Conformalized System Level Synthesis

2026-02-12

AI Summary

This work proposes a novel approach for safe control under out-of-distribution (OOD) scenarios by integrating weighted conformal prediction with System Level Synthesis (SLS). By modeling state-control-dependent covariance, the method constructs high-confidence, data-density-aware bounds on model error and embeds them within a robust nonlinear Model Predictive Control (MPC) framework. Dynamic constraint tightening is achieved through volume-optimized forward reachable sets. The approach provides, for the first time, theoretical coverage guarantees and robustness analysis under distribution shift, explicitly revealing how data density and trajectory tube size influence predictive coverage. Experiments on nonlinear systems—including a 4D vehicle and a 12D quadrotor—demonstrate that the proposed method significantly enhances safety and robustness in OOD settings compared to baselines using fixed error bounds or lacking robustness guarantees.

Technology Category

• Intelligent Robots: State Estimation
• Natural Language Processing: Safety and Robustness
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Large-scale security measurements
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We present a novel framework for robust out-of-distribution planning and control using conformal prediction (CP) and system level synthesis (SLS), addressing the challenge of ensuring safety and robustness when using learned dynamics models beyond the training data distribution. We first derive high-confidence model error bounds using weighted CP with a learned, state-control-dependent covariance model. These bounds are integrated into an SLS-based robust nonlinear model predictive control (MPC) formulation, which performs constraint tightening over the prediction horizon via volume-optimized forward reachable sets. We provide theoretical guarantees on coverage and robustness under distributional drift, and analyze the impact of data density and trajectory tube size on prediction coverage. Empirically, we demonstrate our method on nonlinear systems of increasing complexity, including a 4D car and a {12D} quadcopter, improving safety and robustness compared to fixed-bound and non-robust baselines, especially outside of the data distribution.

Problem

• out-of-distribution
• safety
• robustness
• model predictive control
• learned dynamics

Innovation

• Conformal Prediction
• System Level Synthesis
• Robust MPC
• Out-of-Distribution Safety
• Reachable Sets