Research Areas
Our research focuses on making existing and historic structures intelligent, resilient, and sustainable through cutting-edge technology and interdisciplinary approaches.
Digital Documentation & Damage Mapping
We capture the physical reality of heritage structures through photogrammetry, laser scanning, and UAV-based imaging, creating high-fidelity 3D models that serve as diagnostic baselines. Our computer vision pipelines automatically identify and quantify damage patterns in masonry—going beyond simple crack detection to understand the complex behavior of joints, materials, and structural systems. These digital twins become living records that track deterioration over time and inform intervention priorities.
Structural Diagnostics
Using non-destructive evaluation techniques like Ground Penetrating Radar and thermal imaging, we see beneath the surface of historic buildings without compromising their integrity. Our data-driven models automate the interpretation of GPR scans to reveal hidden structural elements, voids, and material properties.
Post-Disaster Assessment
When disasters strike historic communities, rapid and accurate damage assessment is important for recovery. We develop frameworks and tools for post-event reconnaissance that operate at both building and community scales. From tornado-damaged structures in the Midwest to flood-impacted masonry after dam breaches in Ukraine, our work identifies which building attributes drive vulnerability. This evidence base helps communities understand their risk and prioritize recovery efforts where they matter most.
Computer Vision & AI
We design and train convolutional neural networks specifically for the challenges of historic structures—complex geometries, heterogeneous materials, and damage patterns that defy conventional classification. Our deep learning architectures perform semantic segmentation of point clouds, automated damage detection, and visual priority mapping. By capturing how human experts evaluate buildings through eye-tracking studies, we encode tacit knowledge into AI systems that can scale inspection and assessment workflows.
Multi-Modal Data Fusion
Heritage structures generate data from diverse sources: laser scans, thermal imaging, moisture sensors, historical records, and community surveys. We use factor analysis and machine learning to identify underlying patterns across these heterogeneous datasets, revealing relationships invisible in any single data stream. This synthesis enables efficient monitoring strategies optimized for specific site conditions and connects physical deterioration to social and environmental factors.
Machine Learning-Informed Physics
We use machine learning for dimensionality reduction on finite element model datasets to identify which features actually drive structural vulnerability. By letting real disaster data reveal what matters, we focus our physics-based simulations on critical details—like roof-to-wall connections and other failure mechanisms observed in the field. This data-driven approach to simulation ensures our computational models investigate the right questions, making our predictions both efficient and grounded in actual building performance.
Adaptive Reuse & Optimization
Historic buildings must evolve to meet contemporary needs while preserving their character and integrity. We integrate Shape Grammar with artificial intelligence to explore design solution spaces for adaptive reuse—determining optimal structural wall layouts, spatial configurations, and intervention strategies. Our computational design tools balance competing objectives: structural performance, spatial functionality, heritage significance, and sustainability. The result is evidence-based pathways for breathing new life into existing structures.
Resilience Frameworks
We create actionable frameworks that help historic communities prepare for, withstand, and recover from disasters and climate change. Our work connects social science insights about community agency with engineering tools for vulnerability reduction and adaptation planning. From flood-prone Vermont villages to climate-vulnerable urban districts, we develop strategies that preserve not just buildings but the social fabric and cultural continuity they support. Our goal is proactive resilience—preventing loss before it happens.
Knowledge Transfer & Decision Support
Technical research only creates impact when it reaches the people who make decisions about heritage. We translate complex data science and engineering findings into guidance that architects, preservation planners, building owners, and policymakers can act on. Through visualization tools, decision frameworks, and stakeholder engagement, we bridge the knowledge-action gap. Our work ensures that cutting-edge research becomes practical wisdom for those stewarding historic places.
