No abstract

To guarantee a safe human-robot collaboration, a collaborative system development requires a significant amount of real world experiments. Yet, it is critical to avoid injury risks for participants of such experiments. The risks could be reduced by introducing a virtual experiments' stage to detect mistakes in a robot behavior prior to the real world experiments. This paper presents a virtual model of a humanoid robot AR601M in the Gazebo simulator. Unlike the standard human models in the Gazebo, this model allows to simulate the gross and fine motor skills of a human and could be used when performing various human actions in collaborative robotic cells.