Locally Semiconcave Control Barrier Function Design for Human Assist Control

Abstract

A control barrier function (CBF), has become the basis for control methods in many studies in recent years. In particular, a human assist control by using a CBF has attracted much attention since theoretically guarantees that state constraints will hold for all human operators' input. In this paper, we propose a locally semiconcave CBF (LS-CBF) using a minimum projection method for multiple CBFs. An LS-CBF enables us to design human assist controls taking decision making into account. The effectiveness of the proposed method is confirmed by using the steering assist control problem of a two-wheeled mobile robot as an example.