Human Behavioral Simulation Using Affordance-Based Agent Model

Abstract

In this paper, we propose a novel agent-based simulation modeling of human behaviors. A conceptual framework of human behavioral simulation is suggested using the ecological definition of affordances in order to mimic perception-based human actions interacting with the environment. A simulation example of a 'warehouse fire evacuation' is illustrated to demonstrate the applicability of the proposed framework. The perception-based human behaviors and planning algorithms are adapted and embedded within human agent models using the Static and Dynamic Floor Field Indicators, which represent the evacuee's prior knowledge of the floor layout and perceivable information of dynamic environmental changes, respectively. The proposed framework is expected to capture the natural manners in which humans participate in systems and enhance the simulation fidelity by incorporating cognitive intent into human behavior simulations.