Digital Commons Network^™

Evaluating The Use Of Machine Learning For Road Maintenance Cost Estimation, Shahla Shirinzad, David Enke

Engineering Management and Systems Engineering Faculty Research & Creative Works

Accurate estimation of highway maintenance costs is essential for efficient resource allocation and long-term infrastructure sustainability. This study evaluates the effectiveness of advanced machine learning models, including ResNet, Transformer, and other neural network architectures, for forecasting maintenance costs using historical data from the Highway Maintenance Improvement Program (HMIP) provided by the North Carolina Department of Transportation (NC DOT). A comprehensive comparative analysis is conducted across multiple models using standard performance metrics, including R², MAE, RMSE, and MSE. The results demonstrate that advanced architectures, particularly ResNet and Transformer, consistently outperform traditional statistical approaches and baseline machine learning models, achieving …

Solving High-Dimensional Differential Equations Using Recurrent And Residual Neural Network Architectures, Hind Khaled Kolaib

Knowledge Engineering and Data Science

High-dimensional Partial Differential Equations (PDEs) form the foundation of complex process modeling in various scientific and engineering applications, including finance, physics, and optimal control. However, classical numerical methods are adversely affected by the curse of dimensionality, making them inapplicable for large-scale problems. Recently, however, deep learning-based approaches have provided a new toolbox for these high-dimensional PDEs, including methods such as the Deep Backward Stochastic Differential Equation (Deep BSDE) method. Our approach draws on a more sophisticated deep learning backbone, using neural networks (in our case, a Residual Neural Network and a Long Short-Term Memory network (LSTM) integrated into the Deep …

Deep Learning Based High-Resolution Electromagnetic Inversion Imaging Using Deep Convolutional Double-Module Structure, He Ming Yao, Shiji Song, Lijun Jiang, Michael Ng

Electrical and Computer Engineering Faculty Research & Creative Works

In this paper, a novel deep learning (DL) approach has been proposed to realize high-resolution electromagnetic (EM) inversion imaging. The newly proposed approach is based on the deep convolutional double-module structure (DCDMS), consisting of the pixel-interpolating module and the corresponding quality-improving module. While the pixel-interpolating module roughly increases the 'resolution' of the initial input, the following quality-improving module realizes quantitative EM imaging in high resolution. The input of the proposed DCDMS adopts the mixed input scheme, consisting of the received EM scattered field and the initial reconstruction in much low resolution computed from Gauss-Newton method. The output of the proposed …

Incremental Cluster Validity Indices And Their Role In Interpreting Lifelong Learning Systems, Niklas Max Melton

Doctoral Dissertations

Clustering and supervised learning are often treated as distinct paradigms, yet both rely on structure in feature space. This dissertation investigates the relationship between cluster validity indices (CVIs) and supervised learning in real-time and lifelong learning settings where data arrive incrementally and cannot be revisited. Across four studies, it develops methods for online cluster validation, uses supervised learning to improve their interpretability, and applies these ideas to evaluating performance degradation in continual learning.

The first study extends incremental cluster validity indices (iCVIs), enabling widely used validation metrics to operate in streaming environments. Experiments on synthetic and real-world datasets show systematic …

Probability Of Detection For Corrosion-Induced Mass Loss And Yield Strain In Metal Structures Using Fe-C Coated Long Period Fiber Gratings Sensors And Electroresistive Strain Gauges, Ying Zhuo

Doctoral Dissertations

"Graphene-based, Fe-C coated long-period fiber grating (LPFG) sensors individually reveal a causal relation between LPFG wavelength change and Fe-C mass loss. Their correlation parameters vary significantly between the sensors mainly because hydrophobic multilayer graphene and graphene oxide make sensor fabrication inconsistent. This study explores MXene-based, Fe-C coated LPFG sensors due to MXene’s inherent hydrophilic property in the presence of surface functional groups such as hydroxyl and oxygen and the probability of detection (POD) for mass loss from Fe-C coated LPFG sensors when deployed at a fixed location. A novel fabrication process incorporating Piranha solution pretreatment is proposed to ensure uniform …

Bridging Language And Game Worlds: Semantic Representations And Text-Driven Terrain Generation For Procedural Content, Zhongyu Xie

Electronic Theses and Dissertations

Procedural Content Generation (PCG) systems produce vast quantities of game levels, terrain, and environments, but lack semantic interfaces: no shared vocabulary exists between natural language, designer intent, and the structured representations generators operate on. This thesis addresses the language-content grounding gap in PCG through two complementary studies spanning semantic analysis and semantic synthesis. The first study introduces a group-supervised contrastive learning framework for semantic representation of symbolic PCG maps under many-to-one semantics, where visually distinct maps may share the same design intent. The framework combines parameter-guided semantic grouping, LLM-based caption augmentation, and a multi-positive contrastive objective that aligns language with …

Texels: A Programmable Textile Interface For Replicating Textures, Maya E. Eusebio

Honors Undergraduate Theses

Self-moving fabric interfaces have massive potential for applications in fields ranging from art to haptic feedback to deployable space structures. However, current systems of implementation face the impracticalities of bulkiness, burnout, and energy consumption on top of limiting designs that can only contract uniformly or create one pre-programmed shape. For this technology to bring the change that it promises, we must break the barriers of usability and sustainability to make it a practical choice. This thesis aims to develop a scalable model of fabric that designers, programmers, and anyone else can acquire or create with accessible materials, integrate into design …

Performance Analysis Of Sparse Neural Networks In Brain Abnormality Detection, Megan Danh

Honors Undergraduate Theses

Neuroimages have held the capability of revealing to medical professionals patterns for brain abnormalities since their development. However, more recently, these professionals and researchers are looking to use neural networks to identify these brain abnormalities through neuroimages for early detection that would allow more effective treatment. Neuroimage datasets, specifically functional magnetic resonance imaging (fMRI), are extremely large in size. This would result in their processing and training to be computationally expensive, even with smaller neural networks. Fortunately, recent pruning methods have recently emerged, where network weights and neurons are pruned to reduce computational cost without compromising too much accuracy. By …

Temperature Compensation In Loop And Patch Fss Strain Sensors: Analysis And Experimental Validation, Swathi Muthyala Ramesh, Kristen M. Donnell

Electrical and Computer Engineering Faculty Research & Creative Works

Frequency selective surfaces (FSSs) are arrays of conductive elements or apertures that exhibit frequency-dependent reflection and transmission properties. Their electromagnetic response is influenced by geometry and environmental conditions, making them attractive for wireless strain-sensing applications. However, temperature variations can produce frequency shifts similar to those caused by strain, reducing measurement accuracy. This work investigates the effects of intrinsic temperature compensation on two common FSS unit cell geometries—loop and patch—through comprehensive simulation analysis. The results show that loop-based cells offer superior thermal stability, while patch-based cells provide greater strain sensitivity, illustrating the trade-off between thermal robustness and mechanical responsiveness. A patch-type …

Mitigating Hysteresis In Metal-Coated Fibers Via Optimized Thermal Treatment For Advanced Distributed High-Temperature Sensing Applications, Koustav Dey, Rony Kumer Saha, Bohong Zhang, S. Narasimman, Farhan Mumtaz, Jeffrey D. Smith, Rex E. Gerald, Ronald J. O'Malley, Jie Huang

Electrical and Computer Engineering Faculty Research & Creative Works

Metal-coated optical fibers are widely employed in sensing applications owing to their superior mechanical strength and corrosion resistance. However, their calibration at elevated temperatures is hindered by hysteresis, manifested as discrepancies between heating and cooling cycles, primarily caused by residual strain from mismatched thermal expansion coefficients (TECs) between the metal coating and silica cladding. This research introduces an optimal heat treatment procedure aimed at minimizing the impact of the mismatch in TECs between the cladding and the coating materials that causes the residual strain in gold (Au) and copper (Cu) coated fibers for achieving reliable distributed high temperature sensing up …

Large Language Models For Neurology: A Mini Review, Donald C. Wunsch, Daniel B. Hier

Electrical and Computer Engineering Faculty Research & Creative Works

Large language models have the potential to transform neurology by augmenting diagnostic reasoning, streamlining documentation, and improving workflow efficiency. This Mini Review surveys emerging applications of large language models in Alzheimer's disease, Parkinson's disease, multiple sclerosis, and epilepsy, with emphasis on ambient documentation, multimodal data integration, and clinical decision support. Key barriers to adoption include bias, privacy, reliability, and regulatory alignment. Looking ahead, neurology-focused language models may develop greater fluency in biomedical ontologies and FHIR standards, improving data interoperability and supporting more seamless collaboration between clinicians and AI systems. Two future developments have the potential to be particularly impactful: (1) …

Evaluation Of Multiple Generative Large Language Models On Neurology Board-Style Questions, Mohammad Almomani, Vijaya Valaparla, James Weatherhead, Xiang Fang, Alok Dabi, Chih Ying Li, Peter Mccaffrey, Dan Hier, Jorge Mario Rodríguez-Fernández

Electrical and Computer Engineering Faculty Research & Creative Works

Objective: To compare the performance of eight large language models (LLMs) with neurology residents on board-style multiple-choice questions across seven subspecialties and two cognitive levels. Methods: In a cross-sectional benchmarking study, we evaluated Bard, Claude, Gemini v1, Gemini 2.5, ChatGPT-3.5, ChatGPT-4, ChatGPT-4o, and ChatGPT-5 using 107 text-only items spanning movement disorders, vascular neurology, neuroanatomy, neuroimmunology, epilepsy, neuromuscular disease, and neuro-infectious disease. Items were labeled as lower- or higher-order per Bloom's taxonomy by two neurologists. Models answered each item in a fresh session and reported confidence and Bloom classification. Residents completed the same set under exam-like conditions. Outcomes included overall and …

Safe Optimal Control Framework For Cooperative Manipulation Of Objects In Human–Robot Teams, Irfan Ganie, Sarangapani Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

This article introduces a distributed deep neural network (NN)-based adaptive control framework for cooperative object manipulation in human–robot teams with unknown agent dynamics by using three distinct multilayer NN observers (MNNOs). The first observer, termed the reference point estimator, enables each robotic agent to estimate the object's reference center using consensus-based learning, even without direct access to global reference trajectories. The second observer, referred to as the human force-to-trajectory estimator, uses human-applied forces to infer the intended position, velocity, and acceleration of the object, enabling real-time estimation of human intent. Together, these two observers allow distributed estimation of human-intended motion. …

Self-Calibrating Uav Navigation: Reinforcement Learning Approaches For Horizontal Trajectory Estimation, Shirin Nasr-Esfahani, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

Accurate unmanned aerial vehicle (UAV) trajectory estimation is essential for autonomous navigation, particularly in GPS-denied environments. Visualodometry and simultaneous localization and mapping (SLAM) approaches require precise camera intrinsic parameters, which are typically obtained through predefined or offline calibration. Instead, in this work, we propose a reinforcement learning (RL)-based self-calibration framework that estimates camera intrinsic parameters directly from monocular video sequences, without requiring prior knowledge of the camera, environment, or calibration targets. This intrinsic parameter estimation is then leveraged to achieve robust UAV trajectory estimation using only video data. We formulate the problem as a sequential decision-making task, where an RL …

Integrating Optical And Radiofrequency Interferometry For Enhanced Phase Sensing, Ruimin Jie, Zhaopeng Zhang, Chen Zhu, Jie Huang

Electrical and Computer Engineering Faculty Research & Creative Works

Interferometry is a crucial investigative technique used across diverse fields to achieve high-precision measurements. It works by analyzing the phase difference between two interfering waves, which results from variations in optical path lengths within an interferometer. We introduce a novel method for directly measuring changes in the phase difference within an optical interferometer, importantly, with the added advantage of a controllable enhancement factor. This approach is achieved through a two-step process: first, the optical phase difference is encoded into a sub-GHz radiofrequency (RF) signal using microwave-photonic manipulation; then, RF interferometry-assisted phase amplification is implemented at the destructive interference point. In …

Enhancing Measurement Accuracy: The Impact Of Missing Data On Parameter Estimation In Mass-Spring-Damper Systems, Michkath Omanda Bouraima, Steven Thompson, Maciej J. Zawodniok

Electrical and Computer Engineering Faculty Research & Creative Works

In sensor-driven dynamic systems, missing data can severely degrade parameter estimation accuracy. This article investigates the impact of missing data on phase estimation in a mass-spring-damper system using an information-theoretic framework based on the Cramér-Rao Lower Bound (CRLB). Closed-form CRLB expressions are derived for four scenarios: complete data, missing completely at random (MCAR) deletion, MCAR-based imputation, and missing at random (MAR) missingness via a selection-weighted formulation. These bounds are used as theoretical benchmarks to evaluate classical imputation methods (last observation carried forward (LOCF), linear interpolation) and advanced approaches (Kalman filtering, Rauch-Tung-Striebel (RTS) smoothing, Bayesian inference, and transformer-based imputation) through Monte …

Rf-Attennet: A Hybrid Attention-Enhanced Network For Mixed Signal Classification In Uav Swarm Detection, Prajoy Podder, Mohammad Atikur Rahman, Maciej Zawodniok, Sanjay Madria

Electrical and Computer Engineering Faculty Research & Creative Works

The continuous increase of UAVs, particularly in swarms, creates significant challenges for security and airspace regulation. Traditional RF fingerprinting methods struggle to detect and classify UAV swarms due to overlapping signals and interference. This study introduces RF-AttenNet, a hybrid deep learning model designed to classify mixed UAV signals by analyzing composite RF spectrograms. RF-AttenNet uses dual attention mechanisms, channel and spatial attention to focus on critical spectral features, enabling the model to effectively separate and identify overlapping UAV signals. We have developed custom composite UAV datasets that simulate real-world swarm interference, incorporating both single and mixed UAV classes. RF-AttenNet achieves …

Transformer-Customer Relationship Identification Based On Load-Switching Fluctuation Characteristics Considering Same-Feeder-Adjacent-Transformer Condition, Yanan Zhang, Gan Zhou, Yuyuan Liu, Wei Gu, Yanjun Feng, Yujue Wang, Rui Bo

Electrical and Computer Engineering Faculty Research & Creative Works

Accurately identifying the connectivity between transformers and downstream three-phase customers in low-voltage distribution networks is challenging, because voltage curves of different phases and nearby nodes can be weakly distinguishable, especially when adjacent transformers on the same feeder serve geographically close customers with highly similar voltage curves. This paper proposes a novel method based on load-switching fluctuation characteristics recorded by smart meters. By extracting localized current and voltage fluctuations and establishing correlation matching, the method overcomes the limited discriminability using steady-state measurements. The method operates in two stages: first, switching-induced fluctuation characteristics are extracted and matched to cluster customers by the …

Corrections To: Enhancing Measurement Accuracy: The Impact Of Missing Data On Parameter Estimation In Mass-Spring-Damper Systems (Ieee Transactions On Instrumentation And Measurement (2026) 75 (1–12) Doi: 10.1109/Tim.2026.3676091), Michkath Omanda Bouraima, Steven Thompson, Maciej J. Zawodniok

Electrical and Computer Engineering Faculty Research & Creative Works

In the above article [1], a wording ambiguity appears in Proposition 4 regarding the description of the missing at random (MAR) mechanism. The published sentence states that the probability of observing the kth sample depends on the realized measurement value. This wording may be interpreted as dependence on the current unobserved value y[tk], which could suggest a missing not at random (MNAR) mechanism. The intended MAR mechanism is that the observation probability for the kth sample depends only on previously observed measurement information, such as y[tk-1], and not on the current unobserved value y[tk]. Therefore, the corrected wording clarifies that …

Design Of An Economic Order Quantity–Based Dashboard For Vaccine Packaging Inventory Management, Hadi Susanto, Susmitha Canny, Alexandra Elizabeth Callistha, Shakira Dwi Purwandari, Naila Davina Aurelia

ASEAN Journal on Science and Technology for Development

Background: Timely availability of vaccine packaging is critical to pharmaceutical supply chain reliability, as packaging materials play a key role in maintaining cold chain integrity during distribution. Inadequate availability of packaging materials can disrupt vaccine delivery even when finished vaccines are ready for shipment. The case company faces frequent delivery delays, insufficient safety stock, and irregular replenishment of vaccine packaging materials, which have jeopardized the organization’s target of achieving a 90% on-time delivery rate.

Objective: To design and evaluate an Economic Order Quantity (EOQ)–based inventory dashboard for vaccine packaging that standardizes ordering decisions and strengthens stock visibility in pharmaceutical distribution …

Probing Proficiency-Related Neural Representations With Pca And Ica, Onila R. Narayana Mudalige Don

Graduate Student Theses, Dissertations, & Professional Papers

How proficiency-related information is represented in task fMRI depends not only on the data themselves, but on the representational lens used to summarize them. This thesis examines second-language (L2) proficiency as a problem of representational organization rather than as a simple classification exercise. Using task fMRI from adult language learners performing semantic animacy judgments in their native language (L1) and second language (L2), I derive shared low-dimensional network representations with principal component analysis (PCA) and independent component analysis (ICA), then evaluate those representations under matched leakage-controlled decoding pipelines.

Across analyses, the central comparison is between representational frameworks rather than between …

Learning-Based Graph Shrinking For Quantum Optimization Of Constrained Combinatorial Problems, Monit Sharma, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

Graph shrinking has recently emerged as a powerful preprocessing technique for hybrid classical–quantum optimization, enabling variable and constraint reduction before quantum solving. Conventional approaches rely on Semi-Definite Programming (SDP) relaxations to compute vertex correlations, but these methods suffer from high computational overhead, instance-specific tuning, and limited generalizability. In this work, we replace the handcrafted SDP correlation stage with a reinforcement learning (RL) based correlation estimator, trained to predict merge quality directly from graph structure. We reformulate the graph shrinking process as a Markov Decision Process (MDP), design a Graph Neural Network (GNN) policy to guide vertex merging, and integrate the …

Adaptive Machine Learning Framework For Microstructural Optimization And Mechanical Performance Prediction In Steels, Henry Adekola Haffner

Doctoral Dissertations

Evolution of microstructures, such as polygonal ferrite, acicular ferrite, bainite, and martensite, plays a pivotal role in determining the final microstructural and mechanical properties of steel products. Given the established inter-relationship between processing parameters, microstructure, properties, and performance, precise control of phase transformation is essential to achieve pre-determined properties. To understand transformation routes in different steel grades, time-temperature-transformation (TTT) and continuous-cooling-transformation (CCT) diagrams are necessary and can be described using the Johnson-Mehl-Avrami-Kolmogorov equation and Scheil’s additivity rule. This study presents a comprehensive computational framework for predicting and optimizing microstructure and mechanical properties in advanced high-strength steels (AHSS) using adaptive machine …

Geometric Effects Of Warfighters On Body Armor Fit And Protective Capabilities Against Shock Wave Energy, Melissa Lynn Sutter

Doctoral Dissertations

Body armor is a vital piece of protective equipment for warfighters to defend against threats, necessitating continued development to improve comfort, weight, and protection. However, female warfighters often wear unisex body armor, designed primarily for the male torso. Current research has evaluated the short- and long-term detriments of female warfighters wearing these armors, focusing on comfort and performance. However, these studies do not comprehensively consider how a non-form-fitting armor compromises warfighter safety from battlefield threats. This research examines the geometric effects of female warfighters on armor protection level when defending against shock threats by evaluating the energy distribution on a …

Zirconium Carbide Based Materials For Extreme Aerospace Environments, Nathaniel Hyman Blatt

Doctoral Dissertations

This research focuses on the processing and properties of zirconium carbide-based materials to promote their use in extreme environment aerospace applications, including nuclear thermal propulsion and hyper sonics. Several carbide systems including ZrC, ZrC-Mo cermets, (Zr, Nb)C, and a high entropy carbide were developed. The ZrC-Mo cermet was studied extensively to understand the effect of starting carbide grain size on the final microstructure, composition, elastic moduli, hardness, fracture toughness, room and elevated temperature flexural strength, thermal diffusivity, electrical resistivity, thermal expansion coefficient, and thermal conductivity. It was shown that heat transport in the cermets was dominated by the ZrC phase …

Architecting A Complex Adaptive System Model For Selecting Policies To Reduce Kidney Discard, Lirim Ashiku

Doctoral Dissertations

"The kidney allocation system is a complex, evolving system involving multiple heterogeneous agents. Each agent exhibits emergent behavior that may not fully align with the complex system goals. Therefore, there is a need for a transdisciplinary systems approach to visualize the interdependency among agents and understand the dominant patterns that shape the kidney allocation system.

First, this research presented an incremental hierarchical system engineering approach in identifying the agents’ needs and behaviors toward the complex systems’ goal of maximizing deceased donor kidney utilization and reducing kidney discard. The hierarchical systems approach linked with model-based system engineering aided in eliciting agents’ …

Assessment Of Simulation Software Used For Cubesat Gnc Verification And Validation By University Research Teams, Alexander Taiyo Newett

Masters Theses

As the growth in university satellite teams continues, along with the greater trend in the small satellite market, the need for a guidance, navigation, and control verification and validation pipeline suitable for these young and inexperienced teams becomes evident. Much of the mathematical theory and software implementation of GNC concepts are large hurdles for teams largely composed of undergraduate students.

Many software packages exist that can help these teams achieve GNC verification and validation. If the learning curves of these software packages can be overcome, new satellite teams have the opportunity to better build and test GNC algorithms that are …

Assessment And Comparison Of Selected Carbon Ablation Models In Hypersonic Free-Flight And Arc-Jet Conditions, Andrew Steven Heider

Masters Theses

Accurate prediction of ablative thermal protection system (TPS) performance is critical for hypersonic vehicle design. However, numerical prediction of ablation remains challenging because results are influenced by complex physics and the choice of surface chemistry model and assumptions made in material response calculations. The objective of this work is to evaluate several carbon ablation models and their implementation within modern computational fluid dynamics (CFD) codes. Numerical simulations were performed using NASA’s LAURA flow solver and the commercial CFD code ANSYS Fluent, which coupled Navier-Stokes with surface chemistry models describing carbon oxidation, nitridation, and sublimation reactions. Several reaction sets were considered, …

Investigating The Dynamic Rating Method: Effects Of Flood Wave Geometry On Discharge Estimation, Daniel James Read

Masters Theses

Rating curves are a vital tool to convert from observed stage to discharge in streamflow monitoring. Most gaging sites utilize a simple rating curve, which assumes a monotonic relationship between stage and discharge. In most cases, this assumption is valid; however, dynamic effects of flood waves often cause significant error in discharge estimation for mildly sloped streams. The dynamic rating method utilizes a numerical solution of the St. Venant Equations applied to time series stage data to compute discharge. Within this method, there is a flood wave parameter called the flood wave factor. The researchers designed a formal computational model …

Examination Of Laser Microwelding As A Technique For The Repair Of Steel Castings, Allan John Woldow

Masters Theses

"Metal castings have long been plagued with minor defects caused by various process and design conditions. Commonly, minor defects are not seen until after machining operations have been completed, and by the time they are identified, they cannot be salvaged in a manner that is structurally sound and does not have other detrimental effects on the product. Detrimental effects may include warping from high heat input in traditional welding, warping from reheat treatment, oxidation of machined surfaces, microstructural anomalies, and reduced mechanical properties. Micro laser welding has the potential to provide a solution to these issues by providing a low …