Redesigning Human-Robot Interaction in Response to Robot Failures: a Participatory Design Methodology

People are increasingly encountering robots in public spaces. To increase the robustness of such in-the-wild robotic applications and to achieve their designed outcomes, existing research focuses on improving the technical reliability of robots and identifying effective strategies to prevent or recover from technical failures. However, in human-robot interaction (HRI), a user’s perception of a robot failure may not necessarily relate to technical issues. We focus on understanding users’ perception of robot behaviours and interactions within the context of a public space. In our exploratory study using a novel participatory design methodology, participants designed robot behaviours for applications in public spaces, and tested their design both in a simulator and on the physical robot. We investigate how participants’ perception and expectations change during this iterative participatory prototyping process, especially when the robot exhibits erratic or unexpected behaviours. Our work provides insights on users’ perception of robot failures, and how users’ design of robot behaviours shifts as they observe the robot within the spatial and social context.