Human-Aware Social Navigation With Comfort Space for Mobile Robots in Human–Robot Coexistence Environments

Abstract

Human-aware social robot navigation, which achieves safe and comfortable navigation by generating human-friendly paths, is a crucial and challenging task for mobile robots deployed in human–robot coexistence environments. This article proposes a human-aware social navigation solution for mobile robots with the Gaussian mixture model (GMM)-based comfort spaces. Unlike existing methods, our proposed solution comprehensively considers both the movement and social communication states of individuals to achieve safe and comfortable human-aware social robot navigation while ensuring navigation efficiency. Specifically, for dynamic scenarios, we design a novel comfort space based on a six-component GMM to meet the comfort of individuals in motion. In addition, for communication scenarios, our solution can address a variety of interaction scenarios, such as two people without communication, two people face-to-face communication, discussions between two people in different positions and social group communication of more than two people. The validity and efficiency of the presented solution are fully verified in terms of planned paths, navigation efficiency, and comfort satisfaction through comparative experiments in both simulation and real-world environments. Meanwhile, a questionnaire survey with 241 participants also shows that our proposal is reasonable and feasible.