Control Barrier Function based on Point Cloud for Human Assist Control

Abstract

Human assist control has also attracted much attention for a system driven by a human operator in recent years. Human assist controllers based on control barrier functions has been proposed. Control barrier functions enable the actual system input to satisfy safety requirements. In this research, we propose a human assist control method using a control barrier function for collision-free locomotion. Especially, this paper focuses on direct use of point cloud data obtained from a laser range scanner. As a result, the proposed controller achieves both collision avoidance and severe approaching to obstacles. Furthermore, it can also take account of the robot shape.