Human-Robot Interfaces for Interactive Receptionist Systems and Wayfinding Applications

Service robots are playing an increasingly relevant role in society. Humanoid robots, especially those equipped with social skills, could be used to address a number of people's daily needs. Knowing how these robots are perceived and potentially accepted by ordinary users when used in common tasks and what the benefits brought are in terms, e.g., of tasks' effectiveness, is becoming of primary importance. This paper specifically focuses on receptionist scenarios, which can be regarded as a good benchmark for social robotics applications given their implications on human-robot interaction. Precisely, the goal of this paper is to investigate how robots used as direction-giving systems can be perceived by human users and can impact on their wayfinding performance. A comparative analysis is performed, considering both solutions from the literature and new implementations which use different types of interfaces to ask for and give directions (voice, in-the-air arm pointing gestures, route tracing) and various embodiments (physical robot, virtual agent, interactive audio-map). Experimental results showed a marked preference for a physical robot-based system showing directions on a map over solutions using gestures, as well as a positive effect of embodiment and social behaviors. Moreover, in the comparison, physical robots were generally preferred to virtual agents.