Human-inspired Route Selection for Efficient Social Robot Navigation in Narrow Spaces

Narrow human environments are challenging for social robot navigation, where distance-optimal shortcut routes to destination are frequently blocked by people in the aisles. People in these situations select the appropriate route between detour and shortcut in an efficient and socially-aware way, by balancing their extra effort (if choosing detour) and the required effort from the person on the way to make space to pass. Similar balance strategy is utilized when people decide the politeness of the request utterance to compensate for the way yielding effort from others. Inspired by this human balance strategy, we developed a decision-making methodological framework for socially-aware high-level path planning and interaction utterance when the robot navigates in narrow shared spaces and implemented it onto a humanoid mobile robot. In order to demonstrate the transparency and utility of this framework, we present a specific instantiation by conducting a set of human-robot experiments in a real store. Data from 30 people were collected via two field experiments to identify socially-acceptable parameters. Results of a within-participants study with 28 human customers interestingly demonstrated that the proposed socially-aware method was more time-efficient compared to the detour-only and shortcut-only strategies.