AI-Based Modeling of Indoor Air Quality and Thermal Comfort: A Systematic Review of Human-Centric and Adaptive Approaches

Indoor Air Quality (IAQ) and thermal comfort are fundamental determinants of occupant health, cognitive performance, and building energy efficiency. Integrating Artificial Intelligence (AI) and soft computing methods—such as machine learning, fuzzy logic, and hybrid architectures—has enabled more intelligent, adaptive, and predictive control of indoor environments. Since the COVID-19 pandemic, interest in data-driven environmental quality management has intensified, particularly in health-sensitive and energy-constrained settings. This systematic literature review (SLR) analyzes 72 peer-reviewed studies published between 2020 and 2024, synthesizing emerging trends in AI-based modeling of IAQ and thermal comfort. The review is organized across four dimensions: (i) environmental sensing and feature engineering; (ii) AI and soft computing-based modeling strategies; (iii) multi-objective integration of IAQ, thermal comfort, and energy metrics; and (iv) human-in-the-loop, interpretable, and adaptive AI frameworks. Using thematic synthesis and advanced visual analytics—including streamgraphs, bubble plots, Sankey diagrams, UpSet plots, and research gap maps—the review identifies four persistent challenges: (i) the lack of unified, multi-objective models that balance comfort, air quality, and energy use; (ii) underutilization of interpretable and hybrid AI approaches that foster transparency and trust; (iii) limited incorporation of real-time occupant feedback, personalization, and behavioral adaptation; and (iv) scarcity of real-world validation across diverse building types and climatic conditions. The findings highlight the need to develop inclusive, explainable, and field-tested AI systems that integrate subjective and objective data for responsive indoor environmental management. This review offers a structured foundation and roadmap for advancing deployable, human-aware, and sustainable AI solutions in the built environment.