Situationsverstehen für die Risikobeurteilung bei der Mensch-Roboter-Kooperation

Abstract

In the proposed system the environment of an industrial robot is captured through algorithms of machine learning. Thus, objects and human actions are determined. Based on semantic analysis situational knowledge is inferred and dynamic risk assessment as well as robotic behaviour are concluded. Consequently, this provides the foundation for a reactive robot system for achieving efficient and safe human-robot-cooperation.