From learning to safety: A Direct Data-Driven Framework for Constrained Control

TL;DR

This paper introduces a novel data-driven control framework that ensures safety in model-free learning control by directly deriving safe control inputs from data, using a new safety certificate called SACBF.

Contribution

It proposes the first direct data-driven safety filters and a new safety certificate, SACBF, enabling safe, optimal control without relying on predictive models.

Findings

Framework guarantees safety and recursive feasibility.

Simulation shows superior safety and task performance.

Decouples safety enforcement from performance optimization.

Abstract

Ensuring safety in the sense of constraint satisfaction for learning-based control is a critical challenge, especially in the model-free case. While safety filters address this challenge in the model-based setting by modifying unsafe control inputs, they typically rely on predictive models derived from physics or data. This reliance limits their applicability for advanced model-free learning control methods. To address this gap, we propose a new optimization-based control framework that determines safe control inputs directly from data. The benefit of the framework is that it can be updated through arbitrary model-free learning algorithms to pursue optimal performance. As a key component, the concept of direct data-driven safety filters (3DSF) is first proposed. The framework employs a novel safety certificate, called the state-action control barrier function (SACBF). We present three different schemes to learn the SACBF. Furthermore, based on input-to-state safety analysis, we present the error-to-state safety analysis framework, which provides formal guarantees on safety and recursive feasibility even in the presence of learning inaccuracies. The proposed control framework bridges the gap between model-free learning-based control and constrained control, by decoupling performance optimization from safety enforcement. Simulations on vehicle control illustrate the superior performance regarding constraint satisfaction and task achievement compared to model-based methods and reward shaping.