This project develops and validates a multimodal artificial intelligence (AI) framework for occupant-centric thermal comfort control in office environments by integrating non-intrusive behavioral sensing, physiological measurements, and user fan-adjustment actions. The research leverages data collected from wearable devices, environmental sensors, and occupant interactions with personal comfort systems to better understand how individuals perceive and respond to thermal conditions in real time. By combining these diverse data streams, the study seeks to capture both physiological and behavioral indicators of comfort and discomfort that are often overlooked in conventional building control approaches. Using advanced machine learning and predictive modeling techniques, the project aims to identify early signs of thermal discomfort before occupants take corrective actions, such as adjusting personal fans or modifying their workspace environment. 

This project develops an interdisciplinary framework to evaluate how the visual condition and physical morphology of buildings influence public safety and urban resilience. By combining architectural assessment, computer vision, and spatial crime analysis, the study aims to classify neighborhood “visual health,” identify design patterns associated with crime hotspots and coldspots, and provide evidence for proactive, design-centered urban investment strategies.

This project redesigns CMGT 608 to move beyond static LEED theory by creating an AI-assisted, simulation-based learning environment where students can explore sustainability decisions through a Digital Twin framework. By developing a Revit-to-Python interface and deploying it in the classroom, the project aims to improve student synthesis of LEED categories, support AI-integrated pedagogy, and produce a redesigned course module, peer-reviewed publication, and documented AI framework.

This research develops data-driven models to assess residential energy performance in the United States, with a particular focus on low-income households facing high energy burdens. By integrating national housing, energy, and time-use datasets, including the American Housing Survey, Residential Energy Consumption Survey, and American Time Use Survey, the study creates a large-scale framework for analyzing household characteristics, occupancy patterns, and end-use energy consumption. The research also evaluates the impact of different Energy Efficiency Measures across household types and develops decision-support models to guide building- and community-level energy improvements. Overall, the project aims to advance sustainable and equitable energy use in the built environment by supporting policies and retrofit strategies that reduce energy burdens and promote energy justice.