Current Research Projects
Current Research Projects
Embedded Files
Empowering Smart City Innovations: Digital Twins of the Built Environment at UNC Charlotte
Empowering Smart City Innovations: Digital Twins of the Built Environment at UNC Charlotte
This cutting-edge initiative will create digital twins of the built environment on campus, leveraging advanced IoT and Building Information Modeling (BIM) technologies. The funded facility project closely aligns with Charlotte’s areas of emerging impact in Smart and Sustainable Cities, as well as Transportation and Advanced Mobility. It will foster interdisciplinary collaboration, drive sustainability efforts, and contribute to the smart city ecosystem, while supporting research in sustainable building design, occupant behavior, smart city planning, AI-driven management, and human-building interaction.
Immersive and Interactive Simulations to Evaluate Vulnerable Road Users’ Behavior
Immersive and Interactive Simulations to Evaluate Vulnerable Road Users’ Behavior
Sponsor: College of Engineering Seed Grant
This project will focus on evaluating riders’ behavior of micromobility users in an immersive interactive digital twin with distributed simulations. It will leverage the existing work of the PIs and their external collaborators to remotely connect simulators and conduct a preliminary study of road users’ behavior. The expected results of this project will have a broad impact on developing immersive digital twins for various studies on vulnerable road users and will increase the capacity of research institutions by remotely connecting simulators.
Worker-in-the-loop real-time safety system for short-duration highway work zones
Worker-in-the-loop real-time safety system for short-duration highway work zones
Sponsor: National Science Foundation | Cyber-Physical Systems Program
Highway work zones expose workers to several safety risks. Federal Highway Administration reports that in 2016 a total of 158,000 crashes occurred in work zones in the US resulting in 42,000 injuries and 687 fatalities. This project departs from existing reactive safety systems to a true proactive safety system. It makes fundamental contributions in real-time deep learning algorithm design and processing, edge computing, and assisted reality systems to enable real-time prediction of work zone intrusions and notification of highway workers. The worker-in-the-loop safety system will be co-designed and co-created with the direct help of highway work zone workers, leading industries (DBi and Woolpert), and human factors experts (HumanFIRST Lab) to identify the best feedback mechanisms for alarming workers regarding upcoming safety risks. This project will play a key role in the development of the next-generation cyber-physical systems with powerful edge computing for many emerging safety and security-related applications.
Assessing Environmental Justice & Historical Transportation Impacts in North Carolina
Assessing Environmental Justice & Historical Transportation Impacts in North Carolina
Sponsor: North Carolina Department of Transportation (NCDOT)
The main scope of this project is to provide a definition for transportation inequity, identify North Carolina communities that have experienced inequitable distribution of the benefits and/or burdens of prior inequitable transportation policies and investments, and deduce potential improvement opportunities.
Scalable and extensible machine learning for predictive road maintenance
Scalable and extensible machine learning for predictive road maintenance
Sponsor: Virginia Department of Transportation (VDOT) & Leidos
In this project, in contrast to generative deterministic and probabilistic prediction models, we are developing a discriminative scalable and extensible machine learning algorithm. In the presence of rich data collected by agencies in recent years as well as advanced computing power and by use of state-of-the-art machine learning, we are aiming at minimizing human intervention to improve data modeling and find deep correlations and patterns that might not be apparent to a human. The comprehensive machine learning framework in this study not only considers a wide range of asset classes and fine-grain geographical and environmental variations but also takes into consideration the interrelations between assets. This project is conducted in a cross-collaboration between UNC Charlotte, Leidos and Virginia DOT, to tackle the long-standing problem of data-driven reliable predictive maintenance.
Data-Driven Prediction of Students’ Performance: A case study of online teaching with instructor-created short videos
Data-Driven Prediction of Students’ Performance: A case study of online teaching with instructor-created short videos
Sponsor: Scholarship of Teaching and Learning (SoTL) Program at UNC Charlotte | PROJECT WEBPAGE
The number of students taking online courses continues to grow significantly in the US, and academic leaders are optimistic about the learning process in the online delivery of the courses. In response to this need, advanced versions of learning management systems have become available with improved features that provide intuitive and efficient user interfaces, data tracking capabilities, and compatibility with portable devices. At the same time, they can easily house several types of learning material and keep track of users' engagement. These platforms make several data points on students' behavior readily available to educators. The overall purpose of this study is to transform the data obtained from learning management systems, course video management tools, and the socio-demographic background of students to predict students’ performance to provide valuable early learner-centered feedback. The timely prediction-based feedback is expected to increase students’ engagement and mobilize them upward in performance clusters (e.g., from low-performing to medium-performing cluster). For this purpose, in this study, we evaluate the effectiveness of the newly redesigned Building Information Modeling (BIM) course with the incorporation of short instructor-created videos and develop an algorithm to predict the students’ final performance given the early-in-semester data of their online activities.
Deep TRAC: Deep Transfer-learning for Road Asset Classification
Deep TRAC: Deep Transfer-learning for Road Asset Classification
Sponsor: Virginia Department of Transportation (VDOT) & Leidos
This research sets to create a scalable framework for autonomous detection and defect assessment of broad range of road assets. In collaboration with VDOT and Leidos, we move toward emerging deep learning algorithms. Our initial results show promising accuracy of 90% for training and 82% for inference for classifying eight asset classes. In this study we will cover broader range of asset items with autonomous support for defect detection and condition assessment.
BiG-DATA MANAGEMENT SYSTEM FOR Highway asset management
BiG-DATA MANAGEMENT SYSTEM FOR Highway asset management
Sponsor: Virginia Department of Transportation (VDOT) & Leidos
With recent advances in data sensing and collection technologies, agencies are facing the challenge of managing large volumes of structured, semi-structured and unstructured data. In this study, we are developing a big-data management system that is not only scalable and accessible but also easy to establish and maintain. We are designing a high concurrent structure compatible with legacy systems in VDOT and ready for cloud-computing.
Past Research Projects
Past Research Projects
- Investigating efficiency utilizing data envelopment analysis: A case study of shipyards
Sponsor: North Carolina Department of Transportation (NCDOT) | PI: Dr. Glenda Mayo
- Lean Construction: strategies and plans for regional economic development at Western North Carolina (WNC)
Sponsor: RAPID Center at WCU
- Lean Construction: strategies and plans for regional economic development at Western North Carolina (WNC)
Sponsor: RAPID Center at WCU
- Lean Construction: strategies and plans for regional economic development at Western North Carolina (WNC)
Sponsor: RAPID Center at WCU
Report abuse