Human-Robot Collision Analysis in Gazebo Environment

Abstract

Physical injuries resulting from collisions between people and moving robots are prevalent in automated manufacturing settings. To prevent hazardous situations, it is crucial to use appropriate simulation models of human actors during a design stage of production processes. This paper investigates an issue of human-robot interaction (HRI) safety and considers collision modeling within the Gazebo simulation environment. We evaluate a suitability of three existing human models: the standard Gazebo actor, the Gazebo actor with a collisions plugin, and the AR-601M robot-based human model, for use in scenarios involving collisions with moving physical objects. The study reveals significant differences in performance of these models, highlighting a need for accurate simulation tools in assessing and improving safety in human-robot interactions.