Investigating pedestrian-robot interaction in a context manipulation experiment

Publication date

DOI

Document Type

Master Thesis

Collections

Open Access logo

License

CC-BY-NC-ND

Abstract

In recent years a growing number of companies started pilot projects with delivery robots. Previous research on pedestrian-robot interaction has looked at, for instance, the effect of robot velocity and anthropomorphism on pedestrian behavior towards a robot, as well as people’s ethical concerns towards these robots. However, little is known about what effect a clearly conveyed function of a robot has on behavior towards this robot. Therefore, a field study was done where observations of pedestrians encountering a stationary mobile robot without a clear function were compared to observations of pedestrians encountering a stationary Garbage cleaning robot. We found that, for pedestrians that encountered the Garbage bot, the chance of ignoring that robot was higher compared to pedestrians encountering the robot with no clear function. An additional finding was that women are more likely to ignore a mobile robot compared to men. The findings have emphasized that companies would benefit from ensuring that their robot conveys a clear function. The clear conveyance of the robot’s function decreases the likelihood of pedestrians interfering with the robot’s tasks. In addition, it decreases the chance of pedestrians getting distracted by the mobile robot, therewith increasing overall road safety. Future studies could conduct experiments with an autonomously driving robot, rather than a stationary one. These studies could aim to investigate the behavior of the demographic groups for which the current study did not provide enough data, as well as investigate how behavior towards mobile robots changes over time.

Keywords

Human-robot Interaction; AI; delivery robot; Bayesian statistics

Citation