SensPS: Sensing Personal Space Comfortable Distance between Human-Human Using Multimodal Sensors

• DOI: 10.48550/arxiv.2502.07441
• Publication Date: 2025-02-11
• Authors: Ko Watanabe; Nico Förster; Shoya Ishimaru
• Keywords: Personal space

Personal space, also known as peripersonal is crucial in human social interaction, influencing comfort, communication, and stress. Estimating respecting personal space essential for enhancing human-computer interaction (HCI) smart environments. preferences vary due to individual traits, cultural background, contextual factors. Advanced multimodal sensing technologies, including eye-tracking wristband sensors, offer opportunities develop adaptive systems that dynamically adjust user comfort levels. Integrating physiological behavioral data enables a deeper understanding of spatial interactions. This study develops sensor-based model estimate comfortable identifies key features preferences. Our findings show particularly data, can effectively predict preferences, with playing more significant role. An experimental involving controlled interactions demonstrates Transformer-based achieves the highest predictive accuracy (F1 score: 0.87) estimating space. Eye-tracking features, such gaze point pupil diameter, emerge most predictors, while signals from sensors contribute marginally. These results highlight potential AI-driven personalization environments, suggesting be leveraged intelligent optimize arrangements workplaces, educational institutions, public settings. Future work should explore larger datasets, real-world applications, additional markers enhance robustness.