Learning Conservative Neural Control Barrier Functions from Offline Data

TL;DR

This paper introduces an offline learning algorithm for neural control barrier functions, called CCBFs, which enhance safety in dynamical systems by preventing unsafe and out-of-distribution states, outperforming existing methods.

Contribution

The paper proposes a novel offline training algorithm for neural control barrier functions that improve safety and reliability in control systems, inspired by Conservative Q-learning.

Findings

CCBFs outperform existing safety methods in experiments

CCBFs effectively prevent reaching unsafe states

CCBFs maintain task performance with minimal safety compromise

Abstract

Safety filters, particularly those based on control barrier functions, have gained increased interest as effective tools for safe control of dynamical systems. Existing correct-by-construction synthesis algorithms for such filters, however, suffer from the curse-of-dimensionality. Deep learning approaches have been proposed in recent years to address this challenge. In this paper, we add to this set of approaches an algorithm for training neural control barrier functions from offline datasets. Such functions can be used to design constraints for quadratic programs that are then used as safety filters. Our algorithm trains these functions so that the system is not only prevented from reaching unsafe states but is also disincentivized from reaching out-of-distribution ones, at which they would be less reliable. It is inspired by Conservative Q-learning, an offline reinforcement learning algorithm. We call its outputs Conservative Control Barrier Functions (CCBFs). Our empirical results demonstrate that CCBFs outperform existing methods in maintaining safety while minimally affecting task performance. Source code is available at https://github.com/tabz23/CCBF.