A Service Robot in the Wild: Analysis of Users Intentions, Robot Behaviors, and Their Impact on the Interaction

Abstract

We consider a service robot that offers chocolate treats to people passing in its proximity: it has the capability of predicting in advance a person's intention to interact, and to actuate an "offering" gesture, subtly extending the tray of chocolates towards a given target. We run the system for more than 5 hours across 3 days and two different crowded public locations; the system implements three possible behaviors that are randomly toggled every few minutes: passive (e.g. never performing the offering gesture); or active, triggered by either a naive distance-based rule, or a smart approach that relies on various behavioral cues of the user. We collect a real-world dataset that includes information on 1777 users with several spontaneous human-robot interactions and study the influence of robot actions on people's behavior. Our comprehensive analysis suggests that users are more prone to engage with the robot when it proactively starts the interaction. We release the dataset and provide insights to make our work reproducible for the community. Also, we report qualitative observations collected during the acquisition campaign and identify future challenges and research directions in the domain of social human-robot interaction.

Figures (7)

• Figure 1: Our robot is tested in two different public environments: it must track multiple nearby people, predict their intention to interact, and proactively offer them chocolate treats.
• Figure 2: Experimental setup (top left) with a zoomed view of the robot with the stretched arm offering chocolate treats (bottom left); a view from behind the robot interacting with a user (bottom right), and the same situation captured by the Robot Sensor (top right).
• Figure 3: Trajectories of users who picked a treat (blue) or not (red), as tracked by the Environment Sensor. Blue trajectories concentrate around the robot, whereas red trajectories are uniformly distributed in space.
• Figure 4: Distribution of the users' torso distance from the robot at the instant of Robot Offer. The IID case shows a broader distribution of users' distance at offering compared to the Distance case.
• Figure 5: Median user distance to robot vs time. The intention to interact is detected at time $t=0$ (vertical dotted line). For people interacting with the robot in any active behavior, at $t=0$ the robot starts the offering motion, carried out in a time interval approximately corresponding to the shaded area. For users of the Passive robot (which does not move), $t=0$ is the time when the robot would have triggered if it had been Active.