A human-centered approach to robot gesture based communication within collaborative working processes

Abstract

The increasing ability of industrial robots to perform complex tasks in collaboration with humans requires more capable ways of communication and interaction. Traditional systems use separate interfaces such as touchscreens or control panels in order to operate the robot, or to communicate its state and prospective actions to the user. Transferring human communication, such as gestures to technical non-humanoid robots, creates various opportunities for more intuitive human-robot-interaction. Interaction shall no longer require a separate interface such as a control panel. Instead, it should take place directly between human and robot. To explore intuitive interaction, we identified gestures that are relevant for co-working tasks from human observations. Based on a decomposition approach we transferred them to robotic systems of increasing abstraction and experimentally evaluated how well these gestures are recognized by humans. We created a human-robot interaction use-case in order to perform the task of handling dangerous liquid. Results indicate that several gestures are well perceived when displayed with context information regarding the task.