"Guess what I'm doing": Extending legibility to sequential decision tasks

Faria et al. [faria2017iros, faria21roman] expanded to multi-party scenarios the impact of legibility in Human-Robot Interaction (HRI). In faria2017iros, the authors explore the impact of applying legible motions in multi-party scenarios. The authors present a user study with a robot serving cups of water to groups of three human partners, who do not know the order through which the robot is going to serve them. The results of the study, show that using only efficient movements led to worst collaboration between the humans and the robot, than when the robot uses legible movements. When the robot focus only on using efficient movements, the humans interacting with the robot would even sometimes get confused regarding who the robot was going to serve and would get in the way of each other.