Gaussian-based Proxemic Zone Definition of Groups for Social Robot Navigation

Chipana P.; Broyer N.; Barrios-Aranibar D.; Diaz-Amado J.; Cardinale Y. · Proceedings of the 2024 Latin American Robotics Symposium, LARS 2024 (2024) · DOI: 10.1109/lars64411.2024.10786418

Resumen

The increasing presence of service robots in human-populated environments has driven the need for these robots to understand and adhere to social norms. In such environments, social robots must consider the presence of individuals or groups of people, whose behaviours are different and condition robots, navigation. In the literature, human safety regarding service robot navigation is classified into physical and psychological safety. While the former has been extensively addressed, the latter is still emerging. One effective method for ensuring psychological safety in social robot navigation is respecting social constraints, such as the proxemics interactions, according to the Hall's proxemic theory. Most initiatives represent the proxemic zones (intimate, personal, social, public) of individuals and group of people with Gaussian functions, enabling social navigation. Although Gaussian functions have proven effective for individual proxemic zones, their efficacy for groups remains unclear. This paper proposes a Gaussian function to model the proxemic zones of a group of people, aiming to minimize the area while respecting personal spaces like the O-space, thereby achieving more socially acceptable social navigation. The proposed function is validated through comparative experiments, demonstrating significant improvements in navigation efficiency and psychological safety.

Instituciones: Universidad Católica San Pablo