User-friendly human-robot interaction based on voice commands and visual systems

This paper describes the embryonal development stage of a ROS-based application with the aim of simplify human-robot interaction of collaborative workstations. The idea is to command the robot with voice commands detected by speech recognition algorithms and empower the robot's understanding of its surroundings to perform the tasks using a visual system capable of recognizing simple objects. In this first version the application has been developed virtually using Gazebo simulation software. Some qualitative tests regarding the vocal control and the visual detection have been performed to understand the system's issues and drawbacks that will be solved in future developments.