Gesture Cues in Navigational Robots

Abstract

As robots have become better in tasks such as motion planning and obstacle avoidance, they will soon face a new challenge: sharing a physical space with humans. This challenge means that robots and humans need to be able to interpret what the other is doing at the moment, and predict what will happen in the near future. In the current study we tested whether people would learn from a robot’s navigation behavior while playing a navigational game. The robot was either honest or dishonest in showing its navigational intentions. Results showed differences in people’s understanding of the robot’s behavior, the perceived human-likeness of the robot, and performance in the game. People also improved their performance throughout the dishonest rounds. These findings can be used in the design of robots that need to function effectively in mixed human-robot environments.