Digital Commons Network^™

A Deep Learning Framework For Last-Mile Delivery Enhancement Using Social Media, Valeria Laynes Fiascunari

Graduate Thesis and Dissertation post-2024

Over the past decade, people have been spending more time online. Almost anything can be done from a laptop or cellphone. This is one of the reasons why e-commerce has been in a constant boom, as it is easier to buy something online and have it delivered to the front door than to go to the store. As more people engage in this activity, e-commerce platforms' challenges are more complicated and need to be addressed faster. However, these challenges escape the company's scope when external factors influence the objective of optimized deliveries, for example, traffic issues or bad weather during …

Modeling Frameworks To Understand And Predict Land Use Type And Land Value Patterns In Florida, Lauren Hoover

Graduate Thesis and Dissertation post-2024

Florida’s population has been rapidly growing, with a 14.6% growth in population from 2010 to 2020. This growth in population has led to an associated increase in demand for housing and commercial establishments, which have resulted in changes in land use patterns across the state. For this dissertation, we focus on methods to qualitatively and quantitatively examine how land use patterns are likely to evolve under the expected scenario of population growth in Florida while explicitly allowing for interactions between demographics, transportation and land use patterns. In our first objective we build a simulation framework to predict land use evolution …

Examining The Feasibility And Influencing Factors Of Large Language Models In Source Code Security Analysis, Jie Lin

Graduate Thesis and Dissertation post-2024

Large Language Models (LLMs) have recently emerged as promising tools for analyzing source code; however, their capabilities for automated vulnerability detection and localization remain underexplored, particularly given the widespread reliance on closed-source models which introduce significant privacy, security, and transparency risks. This dissertation rigorously examines the feasibility and key influencing factors determining the efficacy of open-source LLMs in source code security analysis. Initially, we investigate the impact of tokenized input length on vulnerability detection accuracy and explicitness across Java vulnerability datasets using ten distinct LLM architectures, identifying robust performance in specific models while others demonstrate considerable accuracy deterioration. Subsequently, we …

The Impact Of Visualization Styles On Movement Imitation Accuracy In Virtual Reality, Gabriela R. Shamblin

Graduate Thesis and Dissertation post-2024

Virtual reality (VR) has become a powerful tool for motor learning and skill acquisition, offering immersive environments for users to practice and refine movements. This thesis investigates how different visualization styles in VR affect movement imitation accuracy, specifically focusing on hand movements. While prior research has explored precise alignment and visualization individually, few studies have examined their combined impact. This study addresses that gap by evaluating the effectiveness of various visualization methods in relation to offset, animation, and manual type.

We developed an application to ensure all participants experienced each visualization factor as 12 combinations in varied sequences. The user …

Multiscale Modeling And Experimental Evaluation Of Droplet Evaporation In High-Pressure Pulsed Spray Cooling For Thermal Management Applications, Fernando Soria Proano

Graduate Thesis and Dissertation post-2024

Droplet evaporation plays a vital role in spray cooling applications, helping to maintain safe operating temperatures for high-powered devices such as lasers and supercomputers. There exists an ongoing debate regarding the appropriateness of diffusion-limited versus kinetically limited models for describing this complex process. This work seeks to bridge the macro-scale evaporation, governed primarily by capillary forces, with the nano-scale interactions among the vapor, liquid, and solid phases through a comprehensive description of disjoining pressure.

By integrating principles from lubrication theory, heat conduction, diffusion, and statistical methods, a detailed model for evaporative mass flux is developed. This model incorporates various interactions, …

Atomic Layer Deposition As A Method Of Modifying Sintering Behavior Of Nanopowders, Eric Bissell

Graduate Thesis and Dissertation post-2024

Nanocrystalline ceramics promise superior mechanical, electrical and optical properties, yet their fabrication is impeded by a lack of scalable manufacturing methods. This thesis explores atomic layer deposition (ALD) as an enabler to manufacturing nanocrystalline ceramics. Chapter 1 frames the scientific and industrial motivation for combining traditional sintering methods with ALD, highlighting the challenges involved in nanocrystalline ceramics manufacturing, and the opportunities that lie in the combination of the two techniques. Chapter 2 surveys the state-of-the-art in nanocrystalline ceramics manufacturing, powder ALD chemistry and reactor design. Three challenges of powder ALD are highlighted and a critical comparison of fluidized-bed versus rotary-bed …

Computational Methods For Population Genetic Inference Using Identity By Descent And Local Ancestry, Yuan Wei

Graduate Thesis and Dissertation post-2024

Identity by descent (IBD) and local ancestry are essential for population genetic inference, such as understanding genealogical relationships and demographic history through genomic data. With the availability of large and high-resolution datasets, new opportunities and computational challenges are brought for efficient and accurate methods to infer IBD segments and local ancestry. This dissertation presents three computational methods designed to assist in population genetic inference. First, RaPID-Query is introduced to efficiently query IBD segments for individual haplotypes over a large genotype dataset. This method is based on the positional Burrows-Wheeler transform (PBWT) algorithm and utilizes a random projection approach. RaPID-Query is …

Algorithms And Benchmarking For Parallel Identity-By-Descent Segment Detection, Kecong Tang

Graduate Thesis and Dissertation post-2024

As genomic biobank initiatives continue to grow, the availability of large-scale genotype datasets, encompassing hundreds of thousands to millions of individuals, has transformed genetic research and biomedical discovery. However, the sheer volume of this data presents major computational barriers. Efficient and scalable methods are urgently needed to process and extract meaningful signals from biobank-scale data using modern multi-core architectures. One central task in this domain is the detection of identity-by-descent (IBD) segments, which underpins a range of applications including genealogical inference, disease mapping, phasing, and population structure analysis.

This dissertation addresses these challenges by presenting a sequence of contributions that …

Beam-Clip: Multimodal Alignment For Mmwave Beam Pattern Learning, Andrew P. El Kommos

Graduate Thesis and Dissertation post-2024

This thesis presents a novel approach to beam prediction in wireless communication systems using a multimodal masked CLIP (Contrastive Language-Image Pre-training). We introduce a two-phase training methodology that first aligns representations across multiple sensor modalities—GPS, Radar, LiDAR, and RGB images—through masked contrastive learning, followed by task-specific fine-tuning for channel power reconstruction. Our approach adapts CLIP’s pre-training strategy to the domain of wireless signal modeling, enabling the model to learn rich, transferable features that capture the spatial and contextual dependencies of the beam distribution. Notably, the pre-training stage provides a substantial boost to overall performance, significantly improving the model's ability to …

Stress-Strain-Strength Behavior Of Florida Sands And Silty Sands, Sergio A. Marin Savatier

Graduate Thesis and Dissertation post-2024

Understanding the mechanical response of granular materials is fundamental to geotechnical engineering design. The presence of fines significantly influences the stress-strain-strength behavior, particularly when the material response is controlled by fines content. In this study, Florida sands were initially characterized by their grain distribution and particle morphology, while natural fines were evaluated according to their plasticity characteristics. Two materials were selected for detailed investigation: clean sands and a sand-fines mixture containing 20% fines, representing the threshold between coarse- and fine-dominated behavior. Constant rate of strain (CRS) consolidation tests were performed in both materials to study compressibility and effect of strain …

A Numerical Assessment Of Shock-Raindrop Interaction: An Investigation Of Cavitation Dynamics Driven By Internal Pressure Wave Focusing, Reed Forehand

Graduate Thesis and Dissertation post-2024

This dissertation investigates the formation and evolution of cavitation within liquid droplets subjected to shock wave interactions, with emphasis on internal pressure wave focusing as a fragmentation mechanism. A numerical framework is developed in which high-resolution Volume-of-Fluid (VoF) simulations resolve shock transmission into spherical, cylindrical, and cubic water droplets across a range of flow conditions. The resulting pressure histories at droplet centers are extracted and post-processed using the Rayleigh–Plesset equation to model the dynamics of spherical vapor bubbles.

The modeling framework is validated against established benchmarks, including canonical shock tube behavior, shock-droplet interaction studies, and experimental cavitation observations. A series …

Improving Vulnerable Road Users' Safety Through Computer Vision-Based Crossing Direction Prediction And Multimodal Large Language Model-Based Accident Scene Description, Younggun Kim

Graduate Thesis and Dissertation post-2024

Vulnerable Road Users (VRUs), such as pedestrians and cyclists, are among the most at-risk participants in traffic, making their safety a key priority for intelligent transportation systems (ITS). Accurate perception and understanding of VRU behavior are essential for proactive accident prevention and the design of human-centric mobility systems. Computer vision has become a core component of ITS, enhancing VRU safety. This thesis introduces a human behavior-aware transformer-based framework for VRU intention prediction and presents VRU-Accident, a large-scale benchmark enabling systematic assessment of multimodal large language models (MLLMs) in understanding accident scenarios involving VRUs.

First, the proposed framework leverages multi-modal cues, …

Printability And Phase Transformation Of Selected Alloys Additively Manufactured By Laser Powder Bed Fusion, Thinh Huynh

Graduate Thesis and Dissertation post-2024

The integration of laser powder bed fusion (LPBF) to produce critical engineering components, aimed at achieving reduced weight and enhanced efficiency, has sparked significant interest in the field of additive manufacturing (AM). LPBF is highly desirable due to its rapid production capabilities, minimized material waste, and its ability to facilitate intricate engineering designs. The ongoing challenges associated with processing metals free of cracks while attaining high relative density remain central to advancing AM technology. This has led to a continued interest in alloy development for LPBF given the limited selection of alloys currently available. This study seeks to evaluate the …

The Elliptical Instability In Turbulent Flows: A Mechanistic Framework, David M. Smerina

Graduate Thesis and Dissertation post-2024

The conveyance of energy through the formation, interaction, and destruction of eddies over a wide range of spatial scales, from the largest scale where energy is injected to the smallest scales where the energy is dissipated through viscosity remains one of the most fundamental unsolved problems in fluid mechanics. This dissertation provides a comprehensive investigation spanning multiple flow regimes consisting of high-speed reacting flows and transitional boundary layer turbulence to establish the necessary conditions for turbulent flows to sustain cascades of energy to arbitrarily small scales while demonstrating how the nonlinear development of the elliptical instability leads to the emergence …

Robust Actor-Critic Trajectory Optimization And Autonomous Harvesting Robot Integration, Luis R. Tituana

Graduate Thesis and Dissertation post-2024

This dissertation is organized into two parts. First, we develop a neural network–based constrained trajectory optimization algorithm with stability guarantees, along with a neural network–driven numerical error correction. The framework reformulates the structure of an optimal control problem as a neural network optimization problem, where the solution space lies on a subspace manifold generated by a bio-inspired motion rule that produces open-loop control commands. To address changing conditions, real-world disturbances, and numerical errors, we propose an Actor–Critic–like architecture. In this setup, the Actor network outputs the optimal open-loop control for the optimized trajectory, while the Critic network compensates for disturbances …

Towards Label-Efficient Approaches For Dense Video Tasks, Akash Kumar

Graduate Thesis and Dissertation post-2024

Although deep learning has advanced video understanding, its deployment is limited by two challenges: reliance on manual annotations and restricted ability to operate beyond closed-world settings. Spatio-temporal labeling is costly, while real-world applications demand models that interpret novel, open-ended queries. This dissertation develops methods to improve label efficiency and enable flexible, open-world video analysis.

To alleviate the annotation bottleneck, we establish a semi-supervised framework for video action detection using consistency regularization, where a model learns stable predictions across augmented views of the same video. This is challenging, generic augmentations often affect only static regions, providing limited information about dynamic actions. …

Detonation Morphology For Hypersonic Propulsion, Adam R. Kotler

Graduate Thesis and Dissertation post-2024

Standing detonations are uniquely stabilized in hypersonic flows. The portion of the detonation containing the stabilized shock front is remotely similar to non-reacting normal and oblique shocks and is uniquely differentiable from the latter by its coupling to a reaction front characterized by intense heat-release rates and fast chemical kinetics. The formation of an inert induction region characterized by slow-evolving chemical kinetics, subsequent development of a transition region within which these rates escalate, and terminal coalescence between shock wave and reaction front define the formation sequence of the standing detonation. Observations of this process are reported across multiple research groups. …

Removal Of Long- And Short-Chain Per- And Polyfluoroalkyl Substances From Surface Water Using Green Sorption Media–Nanofiltration Hybrid Processes, Md Touhidul Islam

Graduate Thesis and Dissertation post-2024

Per- and polyfluoroalkyl substances (PFAS) are a class of persistent environmental contaminants that pose significant risks to human health and aquatic ecosystems. This dissertation addresses the removal of both long- and short-chain PFAS from surface water through an integrated approach that combines green sorption media (GSM) with nanofiltration (NF). The research comprises laboratory-scale and field-scale studies designed to evaluate the effectiveness of novel recycled-material-based GSM formulations—namely, CPS and ZIPGEM—as pretreatment options to mitigate membrane fouling and enhance PFAS removal during NF. The study also investigates the adsorption mechanisms under varying water chemistry conditions, including pH, ammonia, and divalent cation concentrations. …

An Exploration And Evaluation Of Biologically Informed Variational Autoencoders For Scrna-Seq Data Analysis, Marjorie R. Principato

Graduate Thesis and Dissertation post-2024

Deep learning has been widely applied to the analysis of high-dimensional biological omics data, especially single-cell RNA sequencing (scRNA-seq). Incorporating biological information into the architecture of variational autoencoders (VAEs) has been shown to enhance model interpretability, allowing the model’s behavior to reflect the underlying mechanisms of the biological system used for its architecture. In recent years, several studies have employed these biologically informed VAEs to model large-scale single-cell transcriptomics data, with models correctly differentiating between cell states and identifying active pathways. However, systematic benchmarking and comparison of different biologically informed VAE architectures remain limited.

In this study, I evaluated and …

When Sediment Meets Strategy: Linking Shoaling Dynamics, Vessel Behavior, And Dredging Optimization, Matthew P. Davies

Graduate Thesis and Dissertation post-2024

Navigation channels are the economic arteries of U.S. maritime trade, yet their efficiency and safety are continually threatened by sedimentation and the escalating costs of dredging. Traditional tools used by the U.S. Army Corps of Engineers (USACE) to forecast shoaling and estimate dredging needs often fall short of capturing the complex interplay between channel condition, vessel behavior, and maintenance decisions. This dissertation develops and integrates three complementary innovations to address this challenge. First, a Parametric Linear Regression Methodology (PLRM) is introduced to improve shoaling predictions by blending long-term and short-term sedimentation rates with tunable weighting factors. This method achieves forecasting …

Pbwt-Based Methods For Biobank-Scale Haplotype Data Analysis, Pramesh Shakya

Graduate Thesis and Dissertation post-2024

Haplotype matching is the task of finding identical matching segments given a group of aligned genetic sequences. Haplotype matches play an important role as they represent biological phenomena. Long haplotype matching segments could indicate identity-by-descent segments showing some degree of genealogical relatedness between the individuals that the haplotypes belong to. Similarly, long segment matches that are adjacent to each other may indicate recombination events. Advanced genotyping technology has made it feasible for large numbers of individuals to be genotyped resulting in many biobanks across the world. This requires efficient algorithms that can analyze large amounts of genetic data.

In this …

Cross-Layer Optimization In C-V2x Towards Level-5 Vehicular Automation, Mahdi Zaman

Graduate Thesis and Dissertation post-2024

Connected and autonomous vehicles (CAVs) necessitate highly reliable, scalable, and responsive Vehicle-to-Everything (V2X) communication systems. Despite recent advancements, Cellular-V2X (C-V2X) in its state-of-the-art form exhibits limitations in managing resource allocation and communication robustness, particularly in dense vehicular scenarios. Additionally, C-V2X standards have been tailored to particularly support periodic broadcasts; whereas true vehicular autonomy demands service-focused communication which differs in transmission behavior compared to periodic safety messages. This document addresses critical bottlenecks within C-V2X by developing cross-layer optimizations across physical, MAC, and application layers. This thesis proposes enhancements on C-V2X resource scheduling protocol with proposing novel adaptive and selective power control …

Examining Information Flows Within Transformer Models Using Transfer Entropy, Clayton Barham

Graduate Thesis and Dissertation post-2024

The rapid growth of artificial intelligence, in terms of both power and prevalence, motivates the need for new ways to study explainability in deep neural networks. This is especially true for transformer architectures and large language models, which are the driving force behind the current surge in artificial intelligence. The complexity of transformer architecture foils many pre-existing methods for studying explainability in deep neural networks. To address this gap in knowledge, this dissertation proposes a new method of studying explainability in transformer architectures that leverages insights from neuroscience, employing transfer entropy to map information flows between components of the transformer …

On Transient Dynamics And Persistent Organization Of Landscape Evolution, Aysan Hassanzadeh Bavojdan

Graduate Thesis and Dissertation post-2024

The evolution and organization of landscapes result from the competition between tectonic uplift, fluvial incision, and hillslope diffusion. Understanding how these processes interact under varying external conditions (i.e., extreme climatic events) is critical for predicting landscape dynamics and long-term geomorphic adjustment. Using a physically based landscape evolution model, we first investigate how variations in the fluvial incision coefficient (K) and soil diffusion coefficient (D) mimic different climatic conditions and control the transient and steady-state organization of landscapes. Results indicate that landscapes with the same non-dimensional index (defined as the ratio of the timescales of advective (fluvial) to diffusive (hillslope) processes) …

Study Of Injector Performance From An Accumulator-Based Plumbing System, Michael R. Benedict

Graduate Thesis and Dissertation post-2024

Accumulators are an alternative solution to pumps as they do not require energy in flight. This means power generated does not have to be diverted to an accumulator, boosting overall efficiency. Additionally, accumulators do not have pressure oscillations or cavitation risks typically associated with pump-driven systems. Further, they are lightweight systems with lower points of failure than traditional pumps. This makes them an interesting candidate for use in hypersonic flight. However, prior to flight data, accumulator-based systems should be tested statically on the ground to determine its feasibility. In this study, a piston accumulator is used in a plumbing system …

Exploring Segmentation, Detection And Tracking In Videos, Jyoti Kini

Graduate Thesis and Dissertation post-2024

Perception is the cornerstone of autonomy, enabling agents to interpret complex environments through segmentation, detection, and tracking. These perception capabilities are fundamental to intelligent systems, whether guiding self-driving vehicles along urban roads, aerial robots mapping diverse terrains, or embodied AI agents interacting naturally with people and objects. Despite sustained research efforts, current perception models remain constrained by fragmented architectures, heavy annotation dependence, and poor generalization to unseen viewpoints, which hinder scalability and reliability. This dissertation advances scene understanding through innovations in segmentation, detection, and tracking, with contributions spanning four key directions that emphasize scalability through end-to-end formulations in 2D and …

Informative Path Planning Algorithms For Anomaly Detection, Samuel Matloob

Graduate Thesis and Dissertation post-2024

Informative path planning (IPP) algorithms are widely used to control the movement of drones or ground robots when the objective is to efficiently collect information from an environment of interest. For instance, in agriculture, drones might be used for the timely detection of outbreaks of plant diseases. Two distinct classes of IPP algorithms are the system2atic algorithms, where the robot moves in a regular planned pattern (such as lawnmower) aiming for uniform sampling and random algorithms (such as random waypoint) that aims to collect a random sample of the environment. In this dissertation, we present several novel IPP algorithms, that …

Thermal Characterization Of Cmc Materials Subjected To Deflagration And Detonation Process, Luis C. Longas

Graduate Thesis and Dissertation post-2024

Transition to greener fuels such as hydrogen is a process that requires developing several technologies. One of those is being able to develop materials that can withstand the higher temperatures of hydrogen combustion. One of the most promising materials are the ceramic matrix composites which could potentially replace the superalloys with thermal barrier coatings which are used in gas turbine engines. This project aims to characterize a ceramic matrix composite manufactured by Polymer Infiltration and Pyrolysis process. The characterization of the material would be done under deflagration and detonation operating conditions. This study would pave the way for the development …

A Predictive Model Of Adsorption Of Cryogenic Liquids In Aerogel-Based Materials, Julie E. Foroosh

Graduate Thesis and Dissertation post-2024

Commodities that are gaseous at standard temperature and pressure are generally stored in thick-walled pressure vessels, or as cryogenic liquids in vacuum-jacketed tanks. An advantage of cryogenic storage is that more mass can be stored per unit volume. However, cryogenic tanks are often bulky and cannot be made into conformal geometries. Adsorption storage in porous and flexible materials, such as aerogel blankets, is a promising method for storing these commodities. Previous studies have shown that when the commodity is adsorbed from the liquid cryogen, instead of from a gaseous state, the mass adsorbed per unit volume of aerogel blanket is …

Modeling And Predicting The Thermal Behavior Of Si/Cnt Nano Composites Using Pinns, Olamide Osigbemeh

Graduate Thesis and Dissertation post-2024

Silicon/Carbon Nanotube (Si/CNT) composites have shown great performance as Thermal Interface Materials (TIMs) for cooling in advanced electronics. The heat transfer in these materials is highly influenced by the interfacial thermal resistance (ITR) at the interfaces between the CNT filler and the silicon matrix. The ITR poses a critical challenge that hinders the design and performance of the TIM. This research designs a Physics Informed Neural Network (PINN) that predicts the thermal behavior of Si/CNT through a forward and inverse solver since they are known networks in solving complex partial and ordinary differential equations. However, despite the advantages of PINNs, …