Digital Commons Network^™

Experimental Thermal Characterization And Comparison Of Direct Liquid Cooling And Air Cooling For A High-Density Gpu Server Thermal Test Vehicle, Oluwagbolahan Esan

Mechanical and Aerospace Engineering Theses

As the power density of artificial intelligence (AI) and high-performance computing (HPC) servers escalates, air cooling is approaching fundamental scaling limits set by air's low convective heat transfer coefficient and rising parasitic fan power. This thesis presents an experimental comparison of direct-to-chip liquid cooling and air cooling for a high-density GPU server, using a Thermal Test Vehicle (TTV) that emulates the device layout and power-density regime of an 8-GPU SXM5-class baseboard (eight emulated GPU blocks and four NVLink Switch modules). The liquid configuration was tested at 8.8 kW (8.0 kW GPU, 800 W switch) across three Coolant Distribution Unit (CDU) …

End-To-End Development And Experimental Validation Of A 1/10-Scale Autonomous Vehicle, Rikkin Pankaj Panchal

Electrical Engineering Theses

Autonomous vehicle development demands vast resources, making scaled down platforms a critical alternative for solving core algorithmic challenges. The primary contribution of this thesis is the end to end development and validation of a complete real time autonomous driving pipeline deployed on a one tenth scale vehicle. To streamline platform development, an AI assisted annotation framework automates dataset generation, significantly reducing manual labor while improving training data quality. The system perception stack features a reinforcement learning guided online multi camera calibration framework that enables adaptive surround view stitching without the need for offline recalibration. This is paired with robust lane …

Heterogeneous Robots Cooperation Via Multi-Agent Reinforcement Learning, Rehab Uddin Shawon

Graduate Theses/Dissertations

Enabling heterogeneous robots with diverse capabilities, roles, and task responsibilities to coordinate effectively in complex, dynamic environments remains a fundamental challenge in autonomous multi-robot systems. Multi-Agent Reinforcement Learning (MARL) provides a promising framework for decentralized cooperation. However, most existing MARL approaches assume homogeneous agents or fixed single-task settings and suffer significant performance degradation as the number of robot and task types increases. This thesis presents a scalable shared-policy MARL framework that allows heterogeneous robots to learn specialized behaviors for individual, sequential, and collaborative tasks involving temporal and spatial dependencies through a single neural policy. I first embed robot identity directly …

A Taxonomy-Driven Modular Defense Against Non-Canonical Language In Vision-Language-Action Models, Viraj Samson

Electronic Theses and Dissertations

Vision-Language-Action (VLA) models have recently achieved strong performance across manipulation benchmarks, but these benchmarks rely on highly templated instructions on which the models are typically fine-tuned, leaving their behavior under realistic language-side perturbation unclear. It remains an open question whether the linguistic flexibility inherited from vision-language pretraining survives this fine-tuning, or whether the resulting policies become narrowly tuned to benchmark phrasing and brittle to the intent-preserving language variation that real users naturally produce. We address this gap with a systematic study of VLA robustness under non-canonical instructions, comprising three components: a structured taxonomy of intent-preserving variations spanning linguistic, orthographic, and …

Computational Modeling Of Biophysical Transport: Respiratory Drug Delivery And Tissue-Level Physics, Md Tariqul Hossain

Electronic Theses and Dissertations

The nasopharynx, the upper part of the pharynx located at the back of the nose serves as a critical hotspot for initial respiratory infections via inhaled transmission, largely due to the presence of specific surface receptors that pathogens can exploit for cell invasion, combined with a relatively sparse local mucociliary substrate. To enhance the therapeutic efficacy against certain pathogens, such as the SARS and Influenza viruses, it is therefore essential to improve the targeted delivery of drugs to the nasopharynx. This study explores the use of intranasal sprays as a method for drug administration and models the transport of sprayed …

Wearable Biosensors For Continuous Monitoring Of Chronic Kidney Disease: Materials, Biofluids, And Digital Health Integration, Anupamaa Sivasubramanian, Shankara Narayanan, Gymama Slaughter

Center for Bioelectronics Publications

Chronic kidney disease (CKD) is a progressive and irreversible disorder affecting over 850 million individuals globally and is associated with significant morbidity, mortality, and healthcare burden. Conventional diagnostic approaches rely on intermittent laboratory measurements, including serum creatinine, estimated glomerular filtration rate (eGFR), and urinary albumin, which provide limited temporal resolution and fail to capture dynamic physiological changes. Recent advances in wearable biosensing technologies offer new opportunities for continuous, non-invasive monitoring of biochemical and physiological markers relevant to renal function. This review provides a comprehensive analysis of wearable biosensors for CKD monitoring, focusing on sensing mechanisms (electrochemical, optical, and field-effect transistor), …

Cross-Linking Peg Microgels With Mesoporous Organosilica Nanoparticles To Engineer Microporous Annealed Particle Scaffold Properties, Jonathan M. Zuidema, Alessandro Ajo, Marco Carofiglio, Luisa De Cola

Mechanical Engineering Faculty Publications

Microporous annealed particle (MAP) scaffolds composed of cross-linked microgels are emerging as important biomaterials due to their modular design, high porosity, and large surface area, which allow for highly tailored properties, cellular infiltration, nutrient and waste exchange, and improved drug release characteristics. To further expand on these technologies and enhance their modular design, thiolated mesoporous organosilica nanoparticles (ssMSN-SH), which degrade in response to glutathione, were first cross-linked inside of 4-arm poly(ethylene glycol) maleimide microgels. The microgel size, rheological properties, and payload release characteristics were determined. Hierarchical MAP scaffolds were then created by cross-linking 4-arm poly(ethylene glycol) maleimide microgels with ssMSN-SH. …

Effects Of Matched Freespace And Wall Damping On Stiffness Perception, Madeleine Toland, M Bourgeois, Emma Treadway

Engineering Faculty Research

Stiffness perception has previously been shown to be influenced by damping. This work addresses open questions about what damping conditions are best for enabling classification of virtual walls by their stiffness. Through a forced-choice classification experiment with 12 participants, we found that the highest mean classification performance (in terms of both information transfer and percent correct responses) occurred when freespace damping matched wall damping, rather than being higher or lower. However, statistically significant differences did not appear between the closest damping values employed, suggesting that the effects do not become pronounced until the damping is quite mismatched.

Ai For Life Sciences: From Geometric Protein Modeling To Multimodal Drug Design, Feng Jiang

Computer Science and Engineering Dissertations

Predicting biomolecular interactions, from immune recognition to drug–target binding, is a central problem in the life sciences and computational drug discovery. Deep learning has advanced this area, yet three challenges persist: the topology of large, highly imbalanced interaction networks; structural noise in computationally predicted protein models; and the integration of multimodal information such as functional text and taxonomic annotations. This dissertation develops a coherent set of models spanning immune complex prediction and small-molecule drug design: graph learning that addresses network topology and severe class imbalance; a noise-tolerant method that fuses predicted structures with evolutionary sequence features; and multimodal representation learning …

Ransomware As Organization: A Comparative Analysis Of Corporate And Criminal Structures In Conti, George Urling

Theses, Dissertations and Capstones

Cybercriminal groups continue to pose major threats to global cybersecurity. One of the most common types of cybercriminal groups are, “Ransomware-as-a-Service (RaaS)" groups, who create and sell ransomware. While research is conducted into the development of ransomware, there is limited reporting on the organizational structure and habits of RaaS groups. In 2022, prominent RaaS group Conti had their chat logs leaked, with the logs ranging from 2020 to 2022. This study seeks to provide a deeper understanding of RaaS group structures by utilizing the Conti leaked logs as a case study. The study, entitled “Ransomware as Organization: A Comparative Analysis …

A Macrocognitive Design Taxonomy For Simulation-Based Training Systems: Bridging Cognitive Theory And Human-Computer Interaction, Jessica M. Johnson

Virginia Digital Maritime Center (VDMC) Faculty Publications

Simulation-based training systems are increasingly deployed to prepare learners for complex, safety-critical, and dynamic work environments. While advances in computing have enabled immersive and data-rich simulations, many systems remain optimized for procedural accuracy and surface-level task performance rather than the macrocognitive processes that underpin adaptive expertise. Macrocognition encompasses higher-order cognitive processes that are essential for performance transfer beyond controlled training conditions. When these processes are insufficiently supported, training systems risk fostering brittle strategies and negative training effects. This paper introduces a macrocognitive design taxonomy for simulation-based training systems derived from a large-scale meta-analysis examining the transfer of macrocognitive skills from …

Goal-Driven Shared Control In Eeg-Based Brain Machine Interface For Freewill Reaching And Grasping With Movement Intention Detection And Goal Position Decoding, Bhoj Raj Thapa

Theses and Dissertations--Electrical and Computer Engineering

Upper limb motor impairments can severely limit a person’s ability to perform everyday reaching and grasping tasks. Electroencephalogram (EEG)-based brain machine interfaces (BMIs) offer a non-invasive approach for translating neural activity into control signals for assistive devices such as robotic arms. However, traditional EEG-based BMI studies have generally focused on externally cued paradigms, where both movement timing and target selection are specified by the experimenter rather than freely chosen by the user. In addition, shared control offers a practical framework for assistive BMI operation by dividing responsibility between the user and the intelligent robotic system. However, in many EEG-based shared …

Ai-Driven Network Orchestration: Adaptive Routing For Federated Learning Over Software-Defined Hybrid Wans, Osama Abu Hamdan

Computer Science and Engineering Dissertations - Archive

Modern wide-area networks increasingly adopt hybrid architectures that combine high-capacity wired backbones with flexible wireless links to extend connectivity to remote and underserved locations. However, the bandwidth variability inherent in wireless segments creates routing challenges that traditional protocols, designed for static link capacities, cannot adequately address. Simultaneously, Federated Learning (FL) has emerged as a privacy-preserving distributed machine learning paradigm in which geographically dispersed clients collaboratively train shared models without exchanging raw data. When deployed over wide-area networks, FL training is severely bottlenecked by communication overhead, particularly in cross-silo settings where model payloads reach hundreds of megabytes and synchronous aggregation protocols …

Rope Cutting Task Module, Rebecca Usher

Engineering Theses

This project focuses on the design and evaluation of a rope-cutting module for use with robotic platforms developed by New England Robotics Validation and Experimentation Center (NERVE) and the University of Massachusetts Lowell. The system is designed to cut nylon rope using a specialized robotic cutting arm that can be attached to the robot.

The system consists of a module task box that has two pegs that hold two arms to each one. One arm where the spring- mechanism inside of the task module and the second arm the one on top of the box hold the string in place …

Moisture Performance Of Coated Cross-Laminated Timber Under Coastal Conditions In Vertical And Horizontal Orientations With And Without Weather And Vapor Resistive Barriers, Mia Raley

Engineering Theses

Rhode Island is experiencing sea level rise, and more flooding and erosion during storm events due to climate change. The Rhode Island construction industry wants to adopt mass timber, but have concerns with the changing coastal environment on cross-laminated timber (CLT) wall and slab panels. This research aims to understand how coastal conditions affect CLT under different protection conditions, panel orientations, and structure type, by collecting moisture and temperature data on two structures, one commercially-fabricated and one laboratory-fabricated, located on Shell Path from early-July to early-April. Results found that all layers of horizontally-oriented panels benefitted from an additional protective weather-resistant-barrier …

Improving Ground Penetrating Radar Calibration For Chloride Assessment In Reinforced Concrete Bridge Decks, Camila Silva Wendland

Engineering Theses

This study investigates how ground penetrating radar (GPR) can be used to estimate chloride content in reinforced concrete bridge decks. Chloride buildup at the rebar level can lead to corrosion and structural damage, but current testing methods are destructive and limited to small sample areas. GPR offers a non-destructive alternative, but interpreting reflection amplitudes is challenging because they are affected by multiple factors, including moisture content and rebar depth. To address this, a simulation-based calibration method was developed. Finite-difference time-domain simulations were used to study how chloride, moisture, and rebar depth influence GPR reflection amplitudes. A depth correction method was …

Aster Remote Sensing Satellite Imagery For Regional Mineral Mapping In The Mcmurdo Dry Valleys, South Victoria Land, Antarctica, Khurram Riaz, Amin Beiranvand Pour, Jabar Habashi, Aidy M. Muslim, Iman Masoumi, Ali Moradi Afrapoli, Mazlan Hashim, Kamyar Mehranzamir, Farshid Sattari

Mining Engineering Faculty Publications

The McMurdo Dry Valleys (DVs) of South Victoria Land, Antarctica, constitute the largest ice-free region on the continent and one of Earth’s most Mars-analog environments. Their hyper-arid polar desert conditions offer a unique setting for investigating surface weathering and mineralogical processes under extreme climates. This study presents the first regional-scale mapping of alteration and crystalline weathering minerals across the McMurdo DVs. It uses Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) multispectral data; visible and near-infrared (VNIR) and shortwave infrared (SWIR) bands were analyzed through a Spectral Hourglass Workflow, endmember extraction, and spectral unmixing with Matched Filtering (MF) and Constrained …

Characterization Of Full Bridge Strain Transducers For Haulage Equipment Payload Distribution Monitoring, Jean-Pierre Strydom, Steven Schafrik, Zacharias Agioutantis, Matthew J. Beck, Joseph Sottile

Mining Engineering Faculty Publications

Creating a dependable approach for identifying both the mass of a shuttle car and how material is distributed in it removes the need for equipment operators to manually engage the flight chain. The quantification of environmental and installation conditions and the extent of influence considering their combined contribution towards inaccurate or exclusive measurements are to that degree limited. This experimental study investigated how two different strain transducers—installed in a force-shunt configuration—respond to thermo-mechanical loads when used to determine load distribution and position. Initial observations indicated that thermal effects at the installation site contributed to measurement inaccuracies or exclusive readings. The …

Characterization Of Tomographic Piv System For Single-Phase Immersion Cooling Applications, Joel Joshua Oommen

Mechanical and Aerospace Engineering Theses

As the rapid expansion of Artificial Intelligence (AI) infrastructure and High-Performance Computing (HPC) pushes power requirements to new heights, traditional air-cooling methods have become inadequate for modern server demands. Consequently, single-phase immersion cooling (SPIC) has emerged as one of the effective and sustainable alternatives, utilizing dielectric fluids to remove heat through direct contact with electronic components. Despite numerous thermal and reliability studies, the complex three-dimensional movement of buoyant thermal plumes around intricate component shapes are not yet fully understood in confined spaces.

This thesis provides a detailed characterization of Tomo-PIV and a component level experimental study of flow patterns in …

Scaling Llm Inference: From Novel Attention Mechanisms To Efficient Kv Cache Management, Weishu Deng

Computer Science and Engineering Dissertations

The rapid scaling of artificial intelligence workloads has shifted the dominant performance bottleneck of modern computing systems from compute to memory. Graph neural networks (GNNs) issue increasingly irregular memory accesses, while large language models (LLMs) issue increasingly large ones; in both cases, the relative scaling of memory bandwidth and capacity continues to lag behind the scaling of compute. Consequently, naively executing these workloads on commodity GPUs results in stalled streaming multiprocessors, exhausted high-bandwidth memory (HBM), and serving stacks that incur PCIe transfers on the critical path. This dissertation argues that the efficient scaling of attention-based AI workloads requires the joint …

Land Use And Land Cover Simulation Via Integrated Modelling With Gis Techniques For Sustainable Land Utilization Development In The Northeast Khong Subwatershed Of Thailand, ฺBanchongsak Faksomboon, Thipphaphone Keoviyavong

Applied Environmental Research

Land utilization is an important indicator of socioeconomic and environmental changes caused by both natural and man-made factors. Land use and land cover (LULC) simulation is a critical tool for monitoring and predicting LULC and is essential for sustainable development, land resource management and planning. The cellular automata (CA) Markov model is the basis for the current study’s prediction of LULC changes in the Northeast Khong Sub Watershed (NKSW). Landsat data from 2013 to 2023 were used to investigate LULC classification and determine the spatiotemporal distributions of LULC. In addition, LULC data from 2013 and 2023 were used to generate …

Heavy Metals And Organic Carbon In Sediments Of Seagrass Sediments Of Trang Province, Thailand, Siriporn Pradit, Pornthep Wirachwong, Thongchai Nitiratsuwan, Sujaree Bureekul, Supraewpan Lohalaksanadech, Thawanrat Kobkeatthawin, Monticha Jirajaras, Prakrit Noppradit, Sanya Sirivithayapakorn

Applied Environmental Research

In this study, the accumulation of heavy metals (Cd, Cr, Cu, Fe, and Pb) in water, suspended sediments, sediments, and seagrass in Kalase Bay, Trang Province, Thailand, during the 2024 dry season was investigated. These findings indicate that the enrichment factor (EF) for all the metals was less than 1, suggesting that anthropogenic contamination is not a significant concern in the area. The translocation factor (TF) values were less than 1 for all the metals except Cu, whose TF was greater than 1; however, these values were not statistically significant, indicating limited phytoextraction capacity. The bioconcentration factor (BCF) values for …

Engineering Disorder, Defects, And Electrical Transport For High-Performance Thermoelectric Fe2val Heusler Alloys, James Russell Taylor Iv

Dartmouth College Ph.D Dissertations

Fe₂VAl is a promising low-temperature thermoelectric material for waste-heat energy conversion because it is easily processed and comprised of earth-abundant, non-toxic, and low-cost elements. However, viability for applications faces two major challenges: high thermal conductivity and low p-type Seebeck coefficients. This primarily experimental work aims to address these issues through extrinsic and self-doping, point-defect and disorder generation through thermal and electrical processing, disordered and metastable phase composition quantification, and comparison of transport measurements to modeled electronic features.

An accessible phase quantification method was developed for bulk Fe₂VAl using laboratory X-ray diffraction equipment to compare thermoelectric properties …

Towards All-Epitaxial Josephson Junctions With High-Temperature Superconductors, Daniela Amalfi Ojeda

UNF Graduate Theses and Dissertations

Josephson junctions are fundamental components of superconducting electronics. Although junctions based on conventional superconductors are well established, reproducible devices based on high-temperature superconductors remain challenging. This work investigates the growth, fabrication, and electrical characterization of all-epitaxial DyBa₂Cu₃O₇₋ₓ (DBCO) /SrTiO₃ (STO) /DBCO trilayer heterostructures for potential Josephson junction applications. An initial optimization of YBa₂Cu₃O₇₋ₓ (YBCO) growth produced reproducible films with a zero-resistance critical temperature ( ) of K and a maximum  of 89 K. These growth conditions were adapted to DBCO, resulting in reproducible superconducting films with values of approximately 80 K. DBCO/STO/DBCO trilayers were grown by molecular beam epitaxy and …

Blue Carbon Stock Assessment And Climate Resilience Potential Of Mangrove Forest In Hung Thang Commune, Vietnam, Hla Phong Ko, Pham Hong Tinh, Nguyen Thuy Duong

Applied Environmental Research

Mangrove forests are vital blue carbon sinks that contribute to climate change mitigation and coastal resilience. This study assessed the carbon stock and spatial‒temporal dynamics of a mangrove forest in Hung Thang commune, Hai Phong Province, Vietnam, via a combination of field measurements, soil organic carbon (SOC) analysis, and normalized difference vegetation index (NDVI)-based remote sensing models. The results revealed that mangroves in the Hung Thang community store an average of 87.42 tC ha-1, with SOC accounting for approximately 19% of the total carbon. A generalized additive model (GAM) analysis revealed a consistent increase in carbon stocks between 1996 and …

Assessing Forest Fragmentation And Edge Expansion In Nan Province, Thailand, Using Landsat Data (2014–2022), Kasidit Rison, Marut Fuangarworn, Chatchawan Chaisuekul

Applied Environmental Research

The increasing pressure from land-use change and agricultural expansion in Nan Province, northern Thailand, has accelerated forest fragmentation and reduced the ecosystem service value (ESV) of the landscape. This study aims to: assess land use and forest fragmentation change in Nan Province between 2014 and 2022, and evaluate the impact of land use change on ESV. Landsat 8 OLI satellite imagery from 2014 and 2022 was used to classify land cover into five types: forest, agriculture, settlements, exposed soil, and others. Forest fragmentation was evaluated using landscape metrics such as patch number, mean patch size, edge length, and core area. …

Environmental Sustainability Performance And Efficiency In Thailand’S Rajabhat Universities : A Cross-Regional Analysis Of The Impact Rankings, Konstantinos Balis

Applied Environmental Research

Higher education institutions (HEIs) are expected to increasingly communicate their contributions to the United Nations’ Sustainable Development Goals (SDGs), with the rise of alternative sustainability rankings serving as prominent reporting instruments for institutional transparency. In Thailand, Rajabhat Universities, a network of government HEIs historically associated with their mission of teacher training and local development, have been actively participating in the Times Higher Education (THE) Impact Rankings, presenting a unique case study to explore cross-regional sustainability efforts in relation to global benchmarks. To evaluate trends, regional disparities, and engagement performance across the sustainability pillars, this study uses a descriptive quantitative design …

Investigation Of Automated 3d Scanning Strategies For Large-Scale Polishing Applications, Casey Egan

Open Access Master's Theses

In automated scanning for industrial refinishing applications, large workpieces often exceed the reach of a single robot base position, and may require the environment to be scanned from multiple locations and stitched to a common reference frame. This thesis compares two robotic scanning workflows for such an object: a full scan that reconstructs the object from a single robot base position, and a segmented scan occurring across multiple simulated robot base positions that are stitched together through fiducial-based localization. The results from both workflows are compared for measurement accuracy of the object's bounding box against known dimensions, position of the …

Security-Oriented Voice Authentication Using Machine Learning, Ifeoluwa Stella Elegbe

College of Graduate Studies: Theses & Dissertations

This study develops and evaluates a machine learning and deep learning-based voice authentication system for secure identity verification. As traditional authentication methods such as passwords, PINs, and security tokens continue to face challenges, including identity theft, forgetting, and unauthorized access, voice biometrics offers a more secure, convenient, and user-friendly alternative, especially for remote, hands-free, and accessibility-focused applications. The study adopts a closed-set speaker identification framework, where the system determines the most likely speaker from a predefined group of enrolled users. A structured methodology is implemented, beginning with audio preprocessing and feature extraction. Key acoustic features, including Mel-Frequency Cepstral Coefficients (MFCCs), …

Data-Driven Prediction Of Compressive Strength In Pofa-Modified Concrete: A Comparative Study Of Lr, Nlr, And Ann Models, Shuaaib A. Mohammed, Hawdin Ismael Ibrahim, Diar Fatah Abdulrahman Askar, Hersh F Mahmood, Mehdi Khodaei, Soran Abdrahman Ahmad

Mansoura Engineering Journal

Production of concrete results in high CO₂ emissions, high energy demand, and consumption of various raw materials; however, it is one of the most widely used materials in construction to date. Therefore, researchers aim to minimize the impact of concrete production through various methods, such as introducing sustainable sources of materials that can be used in concrete. For instance, agricultural waste materials can be used as partial cement replacement. This study presents a comprehensive data-driven analysis of the effects of POFA on the properties of normal-strength concrete. By synthesizing a database of 196 experimental datasets from existing literature, three …