Talk to your Cobot: faster and more efficient error-handling in a robotic system with a multi-modal Conversational Agent

Collaborative robot (cobot) systems are expected to be easy to operate. However, reacting to an error message may not be as intuitive as operating the cobot. This paper explores the integration of a multi-modal conversational agent (text, touch and voice) into a robotic system to improve the robot-user interaction during error handling. The conversational agent enables text- and voice-based interactions to improve user experience and efficiency. We evaluate a prototype in a user test with experts and novices in order to assess its effectiveness, usability, and impact on user-robot interaction and show that conversational agents can enhance collaboration between robots and users, particularly benefiting novices.