Constrained Nonlinear Control for Safe Human-Robot Interaction

Abstract

Robotic systems sharing the workspace with humans must fulfill physical and psychological safety conditions. Therefore, this thesis introduces a flexible control framework for nonlinear systems designed to meet the major challenges in human-robot interaction such as satisfying constraints under real-time conditions or adjusting the interaction behavior to the human’s expectation. While formally guaranteeing adherence to known constraints, the approach achieves robust or probabilistic satisfaction of uncertain constraints and is also applicable to multiple interacting robots.