Virtual Inhabitant – A Conceptual Framework and Prototype for a Manikin Simulating Human Metabolic Processes

Abstract

With growing concern over climate change, reducing building energy demand is critical. Building energy use is shaped by many factors, including the persistent gap between predicted and actual consumption. Better understanding human–building interactions can improve building design and modeling by accounting for both energy performance and Indoor Environmental Quality (IEQ). However, despite increasing attention to Occupant Behavior (OB), a consistent and comparable blueprint for modeling presence, movement, and interactions with buildings is still lacking. Traditional simulations also struggle to represent IEQ dynamics that depend on human physiology. This paper introduces the conceptual design of Virtual Inhabitant (VI): a prototype artificial occupant designed to replicate Human Metabolic Processes (HMP) in indoor environments. VI can reproduce occupancy and activity patterns while assessing IEQ, following state-of-the-art approaches for human-related behavior modeling. The system combines two core elements: a body (components and sensors that emulate human heat, moisture, and air exchange) and a brain (a Python-based control model that generates OB patterns). The brain integrates stochastic scheduling, data-driven methods, and AI-based approaches to create realistic presence and activity scenarios. We present the methodological framework, describe the body design, and discuss potential on-site applications. As a conceptual presentation, this work focuses on the integration of stochastic occupant modeling with physical emulation, providing a foundation for future empirical validation. VI aims to reduce uncertainties in Building Energy Modeling (BEM) and support strategies for energy-efficient, resilient, and comfortable buildings within the ConstrucThor research project.