Active Sampling for Constrained Simulation-based Verification of Uncertain Nonlinear Systems

TL;DR

This paper introduces an active sampling method for efficient, data-driven verification of uncertain nonlinear systems, significantly reducing the need for exhaustive simulations and improving prediction accuracy.

Contribution

It presents a novel online estimation technique for prediction accuracy and an active sampling algorithm that enhances verification efficiency within limited sample budgets.

Findings

Up to 50% improvement over existing methods

Effective online prediction accuracy estimation

Validated on three case studies

Abstract

Increasingly demanding performance requirements for dynamical systems motivates the adoption of nonlinear and adaptive control techniques. One challenge is the nonlinearity of the resulting closed-loop system complicates verification that the system does satisfy the requirements at all possible operating conditions. This paper presents a data-driven procedure for efficient simulation-based, statistical verification without the reliance upon exhaustive simulations. In contrast to previous work, this approach introduces a method for online estimation of prediction accuracy without the use of external validation sets. This work also develops a novel active sampling algorithm that iteratively selects additional training points in order to maximize the accuracy of the predictions while still limited to a sample budget. Three case studies demonstrate the utility of the new approach and the results show up to a 50% improvement over state-of-the-art techniques.