Learning for Safety-Critical Control with Control Barrier Functions

TL;DR

This paper presents a machine learning framework using Control Barrier Functions to address model uncertainty in safe control of nonlinear systems, validated through simulation and real-world experiments on a Segway.

Contribution

It introduces an iterative data collection and controller update method leveraging CBFs to enhance safety despite model uncertainties.

Findings

Successfully reduces model uncertainty impacting safety

Achieves safe control in simulation and real-world experiments

Demonstrates effectiveness on a Segway platform

Abstract

Modern nonlinear control theory seeks to endow systems with properties of stability and safety, and have been deployed successfully in multiple domains. Despite this success, model uncertainty remains a significant challenge in synthesizing safe controllers, leading to degradation in the properties provided by the controllers. This paper develops a machine learning framework utilizing Control Barrier Functions (CBFs) to reduce model uncertainty as it impact the safe behavior of a system. This approach iteratively collects data and updates a controller, ultimately achieving safe behavior. We validate this method in simulation and experimentally on a Segway platform.