In context
As service robots entered hotels, museums, and hospitals, human-robot integration demanded not only physical safety but also psychological comfort. Social navigation research typically modeled personal space or learned social behaviors, yet most approaches overlooked human gaze—a key cue for attention and psychological safety. This gap motivated the need for navigation that respects where people are looking.
What was reported
Researchers from Nankai University proposed a socially-aware navigation method that explicitly models human gaze-related area constraints. The system first determines gaze state and builds a horizon-interaction model based on perceived gaze direction and target, using RGB-D camera data. This model defines three zones: the attention area (front 120°), perception area (60°–94° on each side), and blind zone behind the person.
To integrate gaze constraints into navigation, the authors extended the social force model by introducing three novel forces inspired by electromagnetic fields and limit cycle theory: a sight line electrostatic force that repels the robot from the gaze axis, a sight line magnetic deflection force that steers the robot away from the attention area, and a potential field force based on the horizon-interaction model. These forces are combined with the original social forces to drive the robot's motion controller.
Simulation experiments with qualitative and quantitative analyses, plus questionnaire evaluations, demonstrated that considering gaze constraints improves the robot's ability to avoid or approach regions of interest according to task type, enhancing social attributes while ensuring human psychological comfort.
Why it mattered
This work highlighted the importance of gaze as a social cue in human-robot interaction, offering a systematic way to encode it into navigation. It provided a foundation for more natural and acceptable robot behavior in shared spaces, potentially influencing future service robot design in crowded environments.
“This design leads the robot to avoid or approach regions of interest to humans under social norms according to the task type, which enhances the social attributes of the service robot while ensuring the psychological comfort of human.”
Source: 《机器人》期刊 (robot.sia.cn) · Published 2023-11-10 · “考虑人类视线区域约束的机器人社交导航”