RESEARCH EXPERIENCES FOR UNDERGRADUETS IN RESILIENCE AGAINST EXTREME WEATHER EVENTS

RESEARCH TOPICS

Project Title: Acoustic Emission Monitoring for Damage Characterization in Infrastructure Composites

Faculty Mentor: Michael Elwardany

This undergraduate research explores the use of acoustic emission (AE) as a non-destructive technique to detect and characterize damage in infrastructure composites such as fiber-reinforced polymers and polymer-modified concretes. AE sensors capture transient elastic waves generated by microcracking, fiber breakage, and delamination during mechanical loading. By analyzing AE signal features, such as amplitude, energy, and frequency content, the study aims to identify distinct damage mechanisms and monitor their evolution in real time. The project provides students with hands-on experience in experimental testing, signal analysis, and data interpretation, while contributing to the development of smarter, more durable, and sustainable infrastructure materials.

Project Title: ML-Assisted Structural Optimization of Flexible Structures Under Hurricane-Induced Wind and Wave Loading

Faculty Mentor: Pedro Hernández-Caban

During hurricane events, extreme wind and storm surge conditions are the primary environmental forces that impact civil infrastructure in coastal regions. In particular, lightweight, tall, and slender structures (e.g., transmission/communication towers, light poles, etc.) are sensitive to the variable and dynamic nature of wind and wave action. Therefore, predicting and optimizing the performance of such structures remains an ongoing challenge. This REU project aims to investigate the application of novel ML assisted optimization strategies that consider the combined effects of wind and waves, and how these interactions affect coastal structures. Research tasks that the REU will participate in include: 1) Perform a literature review to identify state-of-the-art numerical modeling approaches to recreate the combined effects of wind and wave loads; 2) Build computational finite element models (FEM) of different types of flexible structural systems to simulate their dynamic response under extreme wind and wave loads; 3) Conduct a preliminary series of FEM experiments using high-performance computing; 4) Assess the dynamic response of the different structural systems and optimize their aerodynamic and hydrodynamic performance.

 

Project Title: Detection and Assessment of Vertical Roadside Safety Infrastructure from MLS Pointcloud

Faculty Mentor: Eren Erman Ozguven

Transportation safety depends greatly on the visibility and integrity of roadside safety assets such as signs, delineators, and guardrails. While robust traffic design standards and fail-safe signages mitigate crash likelihood, incidents involving damage to these vertical assets often go unreported, leaving potential hazards for unfamiliar drivers. In the southeastern United States, frequent severe weather events, including hurricanes and tornadoes, exacerbate this problem by displacing or damaging road signs, further compromising roadway safety.

Traditionally, the inspection and maintenance of these assets rely on manual surveys and ground-based scouting, which are time-consuming, costly, and disruptive due to the establishment of temporary work zones and road closures. Advanced three-dimensional (3D) technologies such as mobile laser scanning (MLS) offer a promising contactless solution for cataloging and updating roadway safety infrastructure. Pilot studies reveal that classified point clouds derived from corridor mapping can provide detailed insights into the physical condition and visibility of vertical assets, identifying those obstructed by vegetation, damaged, or fully functional. Furthermore, recent developments in deep learning and semantic segmentation automate this classification process, significantly accelerating inspection workflows while maintaining high accuracy. Students will be actively involved in both MLS data collection and computer analyses tasks and will gain knowledge on cutting-edge research in the areas of Corridor Mapping.

This project advances the state of roadway asset management by integrating mobile LiDAR-based remote sensing with artificial intelligence-driven feature recognition. The proposed framework enables the automated detection and condition assessment of roadside safety elements using dense point cloud data captured from mobile mapping systems. Through the combined use of deep learning classifiers and 3D spatial analysis, the study will automate the extraction, identification, and condition evaluation of vertical roadway assets such as traffic signs and guardrails. The research also contributes to the field of transportation engineering by demonstrating how automated 3D inspection pipelines can reduce reliance on manual data collection, improve update frequency for roadway inventories, and enhance safety through proactive asset monitoring. The fusion of MLS data with auxiliary datasets such as vegetation indices or weather records will further enable predictive modeling of asset vulnerability, offering insights into how environmental stressors accelerate deterioration or visibility loss.

The outcomes of this study will directly benefit state departments of transportation and local road agencies seeking cost-effective, data-driven methods for maintaining roadway safety infrastructure. Through the reduction of field time and minimizing traffic disruptions, the proposed system can improve operational efficiency, lower maintenance costs, and enhance safety for both workers and road users. The automated 3D asset inventory can also support digital twin initiatives, enabling agencies to visualize, monitor, and plan maintenance in a dynamic virtual environment. Beyond immediate operational benefits, this research aligns with broader societal goals of resilient and intelligent transportation systems. It provides a foundation for real-time infrastructure monitoring and post-disaster assessments, particularly relevant in regions prone to severe weather. Moreover, involving students in LiDAR data collection, deep learning model development, and 3D visualization tasks fosters a multidisciplinary learning environment, preparing the next generation of transportation and geospatial professionals for emerging roles in AI-assisted infrastructure management.

 

Project Title: Automated Post-Disaster Structural Health Monitoring and Damage Assessment in Aquatic Environments Using Non-Intrusive Computer Vision

Faculty Mentor: Juyeong Choi
Post-disaster inspection of aquatic and coastal infrastructure—such as bridge piers and offshore platforms—is often delayed and hazardous due to limited accessibility, poor visibility from water turbidity, and reliance on diver-based surveys. These constraints prolong service outages and increase maintenance costs following hurricanes, floods, and tsunamis. This project develops an automated, non-intrusive structural health monitoring (SHM) and damage-assessment framework tailored to aquatic environments by integrating close-range photogrammetry, computer vision, and sensor fusion (photogrammetry and LiDAR). The goal is to enable rapid, resilience-oriented post-event condition assessment to support maintenance and recovery decisions.

Over a three-month period, the project combines a structured literature review, controlled laboratory experimentation, and quantitative evaluation. The initial phase (Weeks 1–8) reviews approximately ten foundational studies on aquatic SHM, with emphasis on photo- and videogrammetry-based computer vision methods, automation gaps, and sensing capabilities for common damage modes including scour, cracking, corrosion, displacement, and material loss. Concurrently (Weeks 3–6), laboratory experiments place a damaged concrete specimen in a transparent water basin and collect rotating video sequences under varying conditions (clear, turbulent, and particulate-laden water). Python-based photogrammetry workflows are used to reconstruct 3D models from the videos, and CloudCompare software is applied to quantify damage metrics (e.g., volume loss, crack geometry, and spatial location) and to classify damage using a pretrained model. In the final phase (Weeks 7–12), model performance is evaluated across water conditions and compared against manual visual inspection as a baseline, with results synthesized into a final report featuring visualizations and quantitative metrics.

Expected outcomes include a validated sensor-fusion workflow for aquatic SHM, an experimental dataset capturing damage visibility across turbidity levels, a baseline for AI-based damage classification in underwater or near-water settings, and a foundation for future field-scale resilience monitoring proposals. The project is well suited for participants with basic Python programming skills and an interest in learning or applying point-cloud processing, Structure-from-Motion (SfM), and computer vision techniques.

 

Project title: Causes of flooding and water pollution in river and coastal systems

Faculty Mentor: Ebrahim Ahmadisharaf

The goal of this research is to explore the space-time structure of hydroclimatic hazards—flooding and harmful algal blooms (HABs). To achieve this goal, we will develop frameworks that characterize and predict these hazards using historical data (satellite observations and ground measurements) and data-driven methods like machine learning algorithms. The investigation will be done in selected locations across the US Gulf of Mexico and selected terrestrial watersheds in the US.

UGR Activities: UGRs will be involved in the statistical analyses of various hazard-influencing factors such as precipitation, soil moisture, wind, storm surge, Karenina brevis and nutrients (chlorophyll-a). The student will also explore the spatial and temporal dependencies of these factors. Through these analyses, key causes of different hazards will be identified. Specific activities are: (1) use public domain repositories and create a database of factors influencing the hydroclimatic hazards such as precipitation, wind, storm surge, Karenina brevis and chlorophyll-a using satellite observations and ground measurements; and (2) characterize coastal hazards in the region. The UGRs also will develop presentations/posters for symposiums/conferences (i.e., EWRI, AGU Fall Meeting, AMS Annual Meeting and IGARSS) to share the research results to the scientific community. The student will be working with a large group of postdocs, graduate and undergraduate students with the possibility of working with outside organization from other universities, NASA, USGS, ERDC, NOAA and ORNL.

This project aims to fill in the knowledge gap regarding causes of hydroclimatic hazards—flooding and HABs—in the US Gulf. The student will utilize novel statistical and geospatial analyses and satellite observations to fill the knowledge gap. Identifying the areas that will be at the risk of flooding and excess water pollution will enable adaptation planning and decision making related to enhancing the resilience of aquatic ecosystem, local communities and physical infrastructure. In addition to the conferences and peer-reviewed articles, the results can be disseminated to national experts (e.g., USGS, USACE and NOAA) through informal virtual meetings and to the non-scientific community through collaborating with bay estuary programs and Cooperative Extension agents/specialists through public outreach meetings and educational activities.

 

Project Title: Fire-Resilient Geopolymer Concrete for Protecting Infrastructure in Wildland–Urban Interface (WUI) Areas

Faculty Mentor: Qian Zhang
Wildfires near communities—known as wildland–urban interface (WUI) fires—are becoming more frequent and intense, putting roads, bridges, utilities, and protective structures at serious risk. Traditional concrete can crack, weaken, or spall when exposed to extreme heat, leading to costly damage and slow recovery. This project explores the use of geopolymer concrete, a more heat-resistant and environmentally friendly alternative, to better protect infrastructure in wildfire-prone areas.

Over a three-month period, the project combines a brief review of existing research with hands-on laboratory testing. The student will work with a graduate student to design, prepare, and test geopolymer mixes in the lab, and understand their performances under wild-fire induced high temperatures. By designing and testing geopolymer concrete mixes that can better withstand extreme temperatures, this project aims to extend the service life of infrastructure in wildfire-prone regions while lowering maintenance and replacement costs.

The expected impact of this project includes improved resilience of infrastructure in WUI communities, reduced economic losses following wildfires, and safer conditions for emergency response and recovery. In addition, the use of geopolymer concrete supports sustainability goals by reducing carbon emissions associated with traditional cement, offering a dual benefit of climate mitigation and climate adaptation for communities facing increasing wildfire risk.

 

Project Title: Prototyping and Assessment of the Virtual Engineering Lab

Faculty Mentor: Qianwen (Vivian) Guo

In this project, the student will develop and pilot a prototype virtual engineering laboratory (similar to shown in picture below) using virtual reality technology in a lower-division undergraduate course. The prototype will be evaluated for quality, usability, and appropriateness for undergraduate instruction. One course that will be targeted is CEG 2202L Intro to Geomatics Engineering, which focuses on computations and field procedures associated with the measurement of distances and angles using tape, level, transit, EDMs, and total station. A mixed-method assessment approach will be used to evaluate student experiences with the virtual lab. The quantitative method will be used to gather students' evaluations of various aspects of the prototype and the qualitative method will be used to identify areas for improvement in the developed prototype. Through this iterative process, this task aims to develop a comprehensive and user-centered version of the virtual engineering lab, and ultimately deploy it in more classrooms to enhance the learning experience of more students

 

Project Title: Enhancing Urban Resilience through Green Infrastructure Optimization for Stormwater Management

Faculty Mentor: Nasrin Alamdari

This REU project, focuses on improving urban resilience to extreme weather events, particularly flooding caused by heavy rainfall. The project aims to explore how strategic green infrastructure (GI) solutions, such as permeable pavements, bioretention systems, and green roofs, can mitigate stormwater runoff, reduce flooding risks, and protect critical infrastructure. By combining hydrological modeling with community-based research, the project will examine the effectiveness of various GI systems in enhancing flood resilience while addressing the unique needs of vulnerable communities.  The student will collect and analyze hydrological data from flood-prone urban areas and use modeling software like PCSWMM to simulate the effectiveness GI systems, such as permeable pavements and bioretention areas, in reducing stormwater runoff during extreme weather events. They will assess various GI interventions for their ability to enhance urban resilience, focusing on flood mitigation, water quality improvement, and cost-effectiveness. The research will fill critical knowledge gaps about the scalability, cost-effectiveness, and resilience benefits of GI systems in diverse urban settings. Students will develop key metrics to assess resilience outcomes, such as reduced flood risk, improved water quality, and economic savings, while engaging with local communities to ensure equitable solutions. This interdisciplinary approach will equip students with the skills necessary to address the growing challenges posed by climate change and urbanization, fostering the next generation of researchers and practitioners in the field of infrastructure resilience.