Natural Hazards Research & Education
Research & Initiatives
Community resilience is the ability of a community to absorb, "bounce back from", and adapt to a disruption. In our research, this disruption is typically a meteorological natural hazard.
Models
Our modeling efforts utilize data-based (machine learning) and physics-based methods. The Hurricane Impact Level (HIL) Model was built using an ensemble of artificial neural networks. Whereas, structural damage and loss modeling from wind hazards for NIST's IN-CORE consists of physics-based fragility curves.
Using the HIL Model to forecast the impact of Hurricane Harvey real-time
Loss curves for a two-story government building found in IN-CORE
Analysis
Our research group utilizes data analytics to further our understanding of socio-physical interactions pertaining to natural hazards research. Dr. Pilkington's initial work in this area involves using artificial neural networks in combination with graph theory to explore such non-linear and complex relationships.
Robotics for Infrastructure Assessment
Our Ghost Robotics V60 legged robot, K-Niner, is currently being tested on construction sites to assess mobility in uncertain terrain conditions, thanks to our industry partners at Clancy & Theys. Our goal is to utilize K-Niner for data collection during surveys of damaged or destroyed infrastructure that may not be safe for human entry. If you are interested in utilizing our ground drone for a project, please contact Dr. Pilkington.
Data
Survey, analysis, and modeling efforts by our group has resulted in uniquely curated datasets with infrastructure, demographic, and meteorological variables, among others. Please contact Dr. Pilkington if you are interested in data sets related to our published work.
Structural Extreme Events Reconnaissance
Dr. Pilkington also serves on the Wind Advisory Board with NSF's StEER and conducts post-event damage assessments, both virtually and in the field.
