Adaptive Leader-Follower Behavior in Human-Robot Collaboration

Abstract

As developments in artificial intelligence and robotics progress, more tasks arise in which humans and robots need to collaborate. With changing levels of complementarity in their capabilities, leadership roles will constantly shift. The research presented explores how people adapt their behavior to initiate or accommodate continuous leadership shifts in human-robot collaboration and how this influences trust and understanding. We conducted an experiment in which participants were confronted with seemingly conflicting interests between robot and human in a collaborative task. This was embedded in a physical navigation task with a robot on a leash, inspired by the interaction between guide dogs and blind people. Explicit and implicit feedback factors from the task and the robot partner proved to trigger humans to reconsider when to lead and when to follow, while the outcome of this differed across participants. Overall the participants evaluated the collaboration more positively over time, while participants who took the lead more often valued the collaboration more negatively than other participants.