Using Intention Recognition in a Simulation Platform to Assess Physical Activity Levels of an Office Building

Designing a building to be conducive to physical activity can significantly contribute to reducing sedentary behavior in a workplace. We aim to develop a game-based simulation platform that provides an intuitive, interactive tool to assess physical activity levels of an office building. The platform simulates workplace physical activity in a 3D virtual building. The assessors can freely manipulate spatial metrics characterizing the building's layouts and investigate their potential associations to physical activity levels. The evaluation may facilitate stakeholders to assess a building plan on public health awareness. One of the key challenges in developing such a platform lies in simulating realistic personal behavior in a complex interaction environment. In this paper, we present our initial development of the simulation platform where we focus on tackling the challenge of incorporating intention recognition in human interactions, demonstrating how this incorporation realistically affects their behaviours. Furthermore, we demonstrate the utility of collective intention recognition techniques in improving recognition reliability as well as efficiency of the simulation platform.