Door and Doorway Etiquette for Virtual Humans

Area of Science:

• Computer Science
• Artificial Intelligence
• Human-Computer Interaction

Background:

• Autonomous multi-human animation literature often overlooks social behaviors like door etiquette.
• Simulating pedestrian interactions at doorways requires integrating perception, navigation, and social decision-making.

Purpose of the Study:

• To introduce a computational framework for simulating socially motivated pedestrian behaviors at doorways.
• To model the common courtesy of holding doors open for others in autonomous animation.

Main Methods:

• An artificial life approach using a layered architecture (mental, behavioral, motor) for autonomous pedestrians.
• A decentralized, agent-based strategy for ordering pedestrians and a state-based model for door interaction.
• A Bayesian network decision model in the mental layer for dynamic, socially-aware behavior selection.

Main Results:

• The framework successfully simulates orderly pedestrian passage through doorways, incorporating social etiquette.
• It handles various door types and diverse doorway etiquette scenarios.
• The simulation achieves efficient, real-time performance.

Conclusions:

• The proposed framework effectively models complex, socially motivated human behaviors in constrained environments.
• This work advances autonomous multi-human animation by incorporating nuanced social interactions like door etiquette.