Digital Commons Network^™

Next-Generation Computing Hardware: Advancements In Tantalum Oxide Reram For Ai And Neuromorphic Applications, Rajas Ravindra Mathkari

Electronic Theses & Dissertations (2024 - present)

The rapid development of artificial intelligence, machine learning, and data-intensive computing has exposed the fundamental limitations of conventional von Neumann architectures, in which energy and time are continuously lost transferring data between physically separate memory and processing units. In contrast, the human brain performs complex computations directly at the point of memory storage through billions of parallel synaptic connections, a paradigm known as in-memory computing. Realizing this in hardware requires memory devices that are fast, energy-efficient, non-volatile, and capable of storing multiple resistance levels in an analog manner. Resistive Random Access Memory (ReRAM) based on tantalum oxide (TaOx) is one …

Investigation Of Fine-Grain Cu And Cu Alloys For Low-Temperature Hybrid Bonding Applications, Sarabjot Singh

Electronic Theses & Dissertations (2024 - present)

Hybrid bonding has emerged as a key enabler for next-generation three-dimensional (3D) integration, offering fine-pitch interconnects and improved electrical performance. However, conventional Cu–Cu hybrid bonding typically requires elevated temperatures to achieve sufficient diffusion and interface quality, posing challenges for temperature-sensitive device integration and process compatibility. This work investigates materials engineering approaches to enable low-temperature Cu–Cu bonding through both microstructure design and alloying strategies.

This work begins by examining grain refinement in Cu as a pathway to enhance diffusion through increased grain boundary density, providing efficient atomic transport without introducing additional elements. Three Cu-based systems Cu–Co, Cu–Ag, and Cu–Al were systematically …

Development Of Alternative Plasma Etching Techniques For The Selective Removal Of Tan With Respect To Sioch Dielectric Materials To Enable Future Back-End-Of-The-Line Scaling, Ivo Otto Iv

Electronic Theses & Dissertations (2024 - present)

Transistor scaling has continued according to Moore’s Law for over fifty years. As transistor size decreases, adequate power delivery is required to enable transistor scaling without performance loss. Power delivery is provided by a metal interconnect network with insulating dielectric that connects the transistor level to the power source, the signal speed within this metal line network limiting transistor level switching speeds. Reduction of signal delay has moved from primarily dimension-based improvement towards adoption of conductor and dielectric materials with lower resistivity and a reduced dielectric constant value, respectively: transitioning from Al/SiO₂ to Cu/low-κ SiOCH. With this transition comes …

On Distribution Matching For Probabilistic Constellation Shaping, Dimitrie C. Popescu, Cameron Kowski

Electrical & Computer Engineering Faculty Publications

Distribution matching (DM) transforms a stream of equally likely bits into digital symbols that correspond to a specific digital modulation scheme with probabilities of occurrence following a specified statistical distribution. This technique is referred to as probabilistic constellation shaping (PCS), and using the Maxwell–Boltzmann (MB) distribution improves spectral and energy efficiency of digital modulation schemes by enabling more frequent transmissions of lower energy digital symbols. The background behind PCS is presented in the paper and the two main DM methods widely used in practice are discussed, outlining scenarios in which one DM method may be preferable over the other. For …

Early-Time/High-Frequency Electromagnetic Induction Sensing For Minimal-Metal And None-Metallic Subsurface Targets, Michele Louise Maxson

Dartmouth College Ph.D Dissertations

Conventional electromagnetic induction (EMI) systems operate within the quasi-static regime, where measurements are dominated by conduction currents and are primarily sensitive to highly conductive targets and bulk soil properties. As a result, these systems exhibit limited sensitivity to low-conductivity and layered media, such as permafrost, composite materials, and minimum-metal landmines, where diagnostically relevant information resides in early-time/high-frequency electromagnetic responses, generally above 100 kHz. This limitation reflects a mismatch between conventional EMI system design and the underlying target physics, restricting the ability of standard EMI approaches to resolve fine-scale non-metallic subsurface structure.

In this thesis, I investigate early-time/high-frequency sensitivity as a …

Neural Network Transceiver For Ltv Mimo Channels, Iresha Amarasekara

Electronic Theses & Dissertations (2024 - present)

Eigenfunctions are commonly employed to characterize kernels in various data-driven analyses. In machine learning, eigenfunction decomposition typically relies on Mercer's theorem, which assumes kernel symmetry. However, this condition is often unmet in communication systems, where channel kernels are asymmetric due to differences in downlink and uplink propagation environments. The High-Order Generalized Mercer's Theorem (HOGMT) provides a systematic approach for decomposing multidimensional asymmetric kernels into eigenfunctions. To address the complexity of eigen-decomposition, this work introduces a baseline neural network (NN) framework HNET. The HNET framework is further enhanced by incorporating the augmented Lagrangian method (ALM) to explicitly enforce orthogonality constraints. This …

Towards Supporting Real-Time Estimation Of Vehicle Fuel Consumption And Co2 Emissions In Smart City Applications, Abrar Alali, Stephan Olariu

Computer Science Faculty Publications

This paper evaluates a simplified physics-based energy demand model designed to estimate vehicle fuel consumption and CO₂ emissions—a critical tool for sustainable transportation planning and smart city applications. Unlike data-driven regression models that lack generalizability for user-defined conditions or complex physics-based approaches that rely on extensive, often proprietary data, the simplified model is distinguished by its minimal parameter requirements, depending primarily on a single, overarching powertrain efficiency value. A key contribution is the comprehensive empirical evaluation of the simplified model against official Environmental Protection Agency (EPA) test data across multiple driving cycles and vehicle types, providing a rigorous validation previously …

Estimating Sediment Properties Using A New Source Level Function For Wind-Driven Underwater Sound Derived From Long-Term Archival Data, S Bruce Martin, Martin Siderius

Electrical and Computer Engineering Faculty Publications and Presentations

Wind-driven breaking waves generate the background sound throughout the ocean. An accurate source level for wind-driven breaking waves is needed for estimating the ambient sound levels needed for sound exposure modeling, environmental assessments, and assessing the detection performance of sonars. Previous models applied a constant roll-off of sound levels at -16 dB/decade at all wind speeds, and these models' source levels were flat at frequencies below ∼1000 Hz due to a lack of measurements. Here, we analyzed 16 long-term archival datasets with limited anthropogenic sound sources to estimate the wind-driven source level down to 100 Hz. We estimated the site-specific …

Deep Reinforcement Learning-Enhanced Model Predictive Control Using Dc-Link Energy Modulation For Inter-Area Oscillation Damping In Pmsg-Based Power Systems, Emad M. Elhaji, Shichao Liu, Hicham Chaoui, Osarodion Emmanuel Egbomwan

Electrical & Computer Engineering Faculty Publications

Low-frequency oscillations remain a major concern in modern power systems with increasing penetration of renewable energy sources. Conventional damping methods, such as power system stabilizers and fixed-gain controllers, may not provide sufficient damping for converter-interfaced wind generators under changing operating conditions and disturbances. To address these limitations, this paper proposes a deep reinforcement-learning-enhanced model predictive control (DRL-MPC) framework for damping inter-area oscillations in multi-machine power systems incorporating permanent magnet synchronous generator (PMSG)-based wind turbines. The proposed approach leverages the energy stored in the DC-link capacitor as a fast dynamic buffer to enable supplementary active-power modulation for oscillation damping. A twin-delayed …

Co-Sputtered Cuni Heteroatomic Electrocatalyst For Enhanced 5-Hydroxymethylfurfural Selective Electrochemical Conversion, Moumita Dikshit, Baleeswaraiah Muchharla, Luz Vazquez Rivera, Kapil Kumar, Sunita Sanwaria, Kishor Kumar Sadasivuni, Abdennaceur Karoui, Sandeep Kumar, Adetayo Adedeji, Bijandra Kumar

Civil & Environmental Engineering Faculty Publications

The electrochemical conversion of biomass-derived 5-hydroxymethylfurfural (HMF) represents a promising, economically viable, and environmentally sustainable approach for producing value-added chemicals using renewable energy and in situ hydrogen generated through water electrolysis. However, the electrochemical hydrogenation (ECH) of HMF remains challenging due to the inherently low catalytic activity and selectivity of the electrodes, compounded by competition with the kinetically favored hydrogen evolution reaction (HER) in aqueous electrolytes. In this work, we demonstrate that Cu_xNi_100-x heteroatomic thin films, fabricated via direct current (DC) magnetron co-sputtering, achieve a more than one order of magnitude increase in the HMF to 2,5-Bis-hydroxymethylfuran …

A. Vs. I In Ai: Is There A Threshold To "Engineered" Intelligence?, Joshit Mohanty

Engineering Management & Systems Engineering Faculty Publications

Despite artificial intelligence reshaping the world, its development generates uncertainties regarding future capabilities. AI simultaneously exists as an artifact of engineering design and as autonomous intelligence, creating an observer-participant feedback loop. This paper proposes that embodied AI faces a bandwidth-limited intelligence threshold T_h that it arises from B = min(C_sens,C_Act). However, Shannon capacity measures bits while intelligence operates on concepts, necessitating a dual-channel model separating physical bandwidth B_io from representational capacity B_rep. Intelligence emerges as multi-dimensional rather than scalar, with components exhibiting different bandwidth dependencies. Surpassing T_h requires either new sensing methods expanding B, enhanced representational frameworks, or reconceptualization within …

A Sysml V2 Metamodel For -Ilities: Application To A Digital Twin Complex System, Pacifique Munezero, Holly Ann Heine Handley

Engineering Management & Systems Engineering Faculty Publications

Multidisciplinary knowledge and exchange of information are two of the most important aspects of complex system design within digital engineering and model-based systems engineering (MBSE) environments. The next generation of systems modeling language, SysML v2, developed by the Object Management Group, improves the implementation and maturation of model-based approaches. This article focuses on using SysML v2 to model reusable parameterized templates for nonfunctional requirements, referred to as -ilities, to address how they can be treated with the same rigor as functional requirements and mapped directly and explicitly from the requirements engineering phase within digital engineering and MBSE environments. The selected …

Frequency And Phase Synchronization Of Antenna Arrays With Phase Incoherent Microcontrollers, Harley W. Byrd

University of Kentucky Master's Theses

The construction of phased array antennas has traditionally been an expensive and complex task. It has recently been claimed that it is possible to construct high-performance Wi Fi antenna arrays using inexpensive consumer components. Motivated by some limited data available from online presentations of such devices, this thesis covers an attempt to construct a phased-array Wi-Fi antenna using several ESP32 chips, which have native Wi-Fi modulation and demodulation capabilities. While there are many positive aspects associated with utilizing ESP32 chips for this purpose, a key challenge is the inherent phase incoherence of their internal Phase Locked Loops (PLLs). The approach …

Cost-Effective Design Of Home Energy Management System With Pv-Wind And Battery Storage In A Grid-Tied Microgrid, Md Mamun Ur Rashid, Jiefeng Hu, Md Alamgir Hossain, Md Shafiul Alam, Ismot Tasmary Salsabil, Bo Long

Research outputs 2022 to 2026

The rising energy demand and the limited availability of traditional energy sources have driven the search for renewable energy alternatives. Over the past few years, more renewable energy sources and significant efforts have been carried out around the World to decrease carbon emissions in the utility sector. To support this development, microgrids have emerged as a smart component of the future power grid. Microgrids, powered by local energy resources, are the most effective for building the new power grid. In microgrids, an Energy Management System (EMS) is an essential element in scheduling the local energy flows. This paper focuses on …

Stem-Fit And Soil-Fit: Integrated Plant And Soil Nitrogen-Hormone Sensing With Machine Learning-Based Forecasting For Next-Generation Precision Agriculture, Nafize I. Hossain, Mohammad Solaiman, A.K.M. Ahsanul Habib, Md Al Mahmud Hossain Al Hadi, Shawana Tabassum

Electrical Engineering Faculty Publications and Presentations

Inefficient fertilizer application in agriculture leads to reduced crop productivity, nutrient losses, and reduced crop resilience, highlighting the urgent need for real-time monitoring of plant–soil nutrient and stress dynamics. This research aims to develop and validate a multiplexed sensing platform for real-time, in-situ measurement of key soil nutrients (Soil-FIT) and plant phytohormones (Stem-FIT) involved in nitrogen signaling and stress regulation. The proposed sensor suite integrates 3D-printed modules for continuous monitoring of nitrate, ammonium, and pH in both soil and plant sap, along with salicylic acid (SA), indole-3-acetic acid (IAA), methyl jasmonate (MeJA), and ethylene (ET) in plant sap. The sensors, …

Development Of Feature Tokenizer Deep Learning Model For Fault Diagnosis In Marine Propulsion System, Pratik Anand Deshpande, J. Preetha Roselyn, Prabha Sundaravadivel

Electrical Engineering Faculty Publications and Presentations

The fault diagnostics in Brushless Direct Current (BLDC) motor drive system is critical for operational safety and system lifespan in propulsion system applications. However, signature parameters such as currents, voltages, speed, and torque have provided nonlinear behavior, which limits the usefulness of traditional model-based approaches. This research provides a deep learning based intelligent system to monitor the failures in marine propulsion system. Each signal feature is represented as a structured token, with a specific class token used to collect global contextual information. The proposed model captures both local temporal dynamics and global inter-feature interdependence multi-layer self-attention processes, allowing for the …

Decentralized Linear Solvers: Communication Cost And Privacy, Nelson G. Brasil, Vinay A. Vaishampayan

Publications and Research

We consider the problem of multiple parties iteratively solving a system of linear equations in a decentralized manner. Specifically, we solve for $\negr{x} \in \R^n$, the $n \times n$ system of linear equations $M\negr{x} = \negr{b}$ when each party only knows their row of the matrix $M$ and a single component of $\negr{b} \in \R^n$. Our objective is to determine the tradeoff between the accuracy of the solution and the total communication cost measured in bits. A fully connected, reliable mesh network is assumed to connect the different parties. We develop a general formulation that applies to a large class …

Adaptive Multi-Grade Deep Learning For Highly Oscillatory Fredholm Integral Equations Of The Second Kind, Jie Jiang, Yuesheng Xu

Mathematics & Statistics Faculty Publications

This paper studies the use of Multi-Grade Deep Learning (MGDL) for solving highly oscillatory Fredholm integral equations of the second kind. We provide rigorous error analyses of continuous and discrete MGDL models, showing that the discrete model retains the convergence and stability of its continuous counterpart under sufficiently small quadrature error. We identify the DNN training error as the primary source of approximation error, motivating a novel adaptive MGDL algorithm that selects the network grade based on training performance. Numerical experiments with highly oscillatory (including wavenumber 500) and singular solutions confirm the accuracy, effectiveness and robustness of the proposed approach.

High Temperature Mold Gap Measurement Using Optical Fiber Interferometry, Jie Huang, Ronald O'Malley, Dinesh Reddy Alla, Deva Prasaad Neelakandan, Muhammad Roman

Patents

A temperature insensitive fiber optic Fabry-Perot interferometer based mold gap measurement system that can be employed in foundry and continuous casting facilities for both non-ferrous and ferrous production applications. For steel continuous casting, the sensor is also capable of detecting crystallite formation in the mold flux layer within the mold gap by detecting reflections from both the flux layer and the steel shell to facilitate direct mold lubrication monitoring for continuous casting. These interferometers can also be easily multiplexed to perform interface shape measurements or monitor gap variations at various positions within a complex mold design. The ability to measure …

Transverse Magnetic Mode Split Post Dielectric Resonator, Donghyun Kim, Xiaoning Ye, Chaofeng Li

Patents

A device for measuring a property of a substrate comprises an enclosure,two resonators,and two probes.The substrate has a top surface, and the enclosure defines a cavity for placing the substrate. The resonators extend into the cavity of the enclosure. The probes extend into the cavity, and each of the probes comprises a loop that extends about an axis that is substantially parallel to the top surface of the substrate so that the probes are operable to excite a transverse magnetic mode signal in the cavity.

The Operation And Control Strategy Of Microgrids: A Brief Review, Mohammad Alam, Salsabila Zaman

Electrical Engineering Student Publications

A microgrid is an effective approach for integrating various distributed energy resources to meet local energy demand. It possesses the capability to operate independently or in conjunction with the main utility grid. Typically, a microgrid is a small-scale power system (usually several megawatts or less) comprising three primary components: the ability to function in both grid-connected and islanded modes, distributed power generation sources, and autonomous control centers. Microgrids enhance system resilience by ensuring uninterrupted operation during grid disturbances. Additionally, they are environmentally sustainable and contribute to improved power quality. As a crucial element of modern power systems, microgrids can effectively …

Biosecure-Llm Framework: Protecting Llms From Cyberbiosecurity Threats And The Case For Independent Ai Safety Governance, Xavier-Lewis Palmer, Lucas Potter, Srdjan Lesaja, Sotirios Karathanasis, Mohammad Ghasemigol

Computer Science Faculty Publications

Large Language Models (LLMs) are becoming critical infrastructure in scientific, healthcare, and governmental contexts. As frontier AI laboratories increasingly partner with government agencies, a fundamental question arises: Who should control the safety and policy-enforcement layers that constrain model behavior? Current safety mechanisms (LLM guardrails) are typically designed for generic "harmlessness" and operate by detecting semantic patterns and refusing requests. However, they are inadequate governance instruments because they cannot implement auditable, domain-specific controls tied to external regulatory policy objects (e.g., control lists or rules governing personally identifying information). Even a perfectly aligned model is not able to express institution-specific policy without …

Formalizing Asymmetric Control-Telemetry Separation In Distributed Industrial Control Systems, Andrew Manison

College of Graduate Studies: Theses & Dissertations

Distributed industrial control systems often place control and telemetry traffic on the same communication substrate even though the two workloads impose different requirements. Control paths need bounded request-response latency and predictable acknowledgement semantics, whereas telemetry paths benefit from scalable publish-subscribe fanout and tolerance for consumer-side delay. This thesis argues that, for the tested class of mixed workloads on shared commodity infrastructure, these communication roles should be separated architecturally rather than forced through a single protocol. To evaluate that claim, the thesis formalizes an asymmetric control- telemetry pattern and instantiates it in the Asymtra framework using gRPC for synchronous control and …

Stabilized Multi-Objective Optimization For Camouflage Attack In Remote Sensing, Omid Rajabi Rostami, Kazi Aminul Islam, Liuwan Zhu, Rui Ning, Chungsheng Xin, Hongyi Wu, Jiang Li

Electrical & Computer Engineering Faculty Publications

We investigate the vulnerability of multi-task maritime surveillance systems in which object detection and semantic segmentation operate concurrently. We propose Stabilized Multi-Objective Optimization (SMO), a framework that jointly targets Mask R-CNN and U-Net with a single universal adversarial texture in a digital white-box setting, with additional cross-model transfer evaluation. SMO combines a logarithmic-inversion loss transformation, feature inconsistency, and a systematic comparison of six optimization strategies for seven heterogeneous objectives. The loss transformation controls loss-dependent gradient scaling, while a U-Net generator supplies a spatially coherent texture and a hull-aligned application function accommodates target geometry without explicit Expectation over Transformation (EoT) sampling. …

Real-Time Neck Posture Classification Using A Lightweight Wearable Imu Pendant, Muntaser Syed, Alfred Sjöqvist, Noah Sedlik, Tsing Liu

Electrical Engineering and Computer Science Student Publications

Poor neck posture during prolonged device use contributes to musculoskeletal disorders affecting millions worldwide. Existing posture monitoring solutions rely on camera-based systems or complex multi-sensor arrays, limiting their practicality for continuous daily use. We present a lightweight, chest-worn pendant using a single 6-axis IMU (accelerometer and gyroscope) for real-time classification of seven neck posture states: neutral, mild flexion, moderate flexion, severe flexion, extension, lateral tilt, and lying. Our approach employs an ensemble architecture combining bidirectional LSTM, Transformer encoder, and 1D-CNN models with learnable fusion weights. To address limited training data, we apply aggressive data augmentation (30x multiplication) including noise injection, …

Dystop: Dynamic Staleness Control And Topology Construction For Asynchronous Decentralized Federated Learning, Yizhou Shi, Qianpiao Ma, Yan Xu, Junlong Zhou, Ming Hu, Yunming Liao

Research Collection School Of Computing and Information Systems

Federated Learning (FL) has emerged as a potential distributed learning paradigm that enables model training on edge devices (i.e., workers) while preserving data privacy. However, its reliance on a centralized server leads to limited scalability. Decentralized federated learning (DFL) eliminates the dependency on a centralized server by enabling peer-to-peer model exchange. Existing DFL mechanisms mainly employ synchronous communication, which may result in training inefficiencies under heterogeneous and dynamic edge environments. Although a few recent asynchronous DFL (ADFL) mechanisms have been proposed to address these issues, they typically yield stale model aggregation and frequent model transmission, leading to degraded training performance …

Parking Information And Supervision System, Joshua A. Thum, Alex J. Kinch, Jacob A. Dye

Williams Honors College, Honors Research Projects

In densely populated areas, finding parking can be an arduous and time-consuming struggle, especially in large and tall parking decks. Drivers would benefit from a convenient way to find an open parking spot without having to scour the entire lot first. The goal of this project is to sense available parking spots in a parking garage or parking lot using physical object detection and visual detection with computer vision verification, and display the open spots to drivers entering the lot. This information will be displayed locally at the lot and in an app, with the latter allowing someone to see …

Dashboard And Racing Telemetry, Cole Barach, Jacob Koshel, Ethan Zifzal, Matthew Sullivan

Williams Honors College, Honors Research Projects

The main goal of the project is to design and manufacture a combined dashboard and data logger for the vehicles produced by the Zips Racing design team. The dashboard will intuitively display real-time information to the driver and record all received information while driving. This information may be pulled off the device later for performing data analysis. This project will incorporate custom PCB design, surface mount soldering, embedded software development, and the CAN communication protocol.

Robotic Air Hockey Table, William Forcey, Andrew Piunno, Kaden Carpenter, Xander Zavatchen

Williams Honors College, Honors Research Projects

Air hockey, a popular arcade game, is traditionally designed for two players. This limits the game’s accessibility for individuals who wish to practice or enjoy it as a single player. To solve this problem, a robotic system was implemented to play air hockey against a human player. The speed and acceleration of the puck and mallet were measured from a game played between humans to inform the required movement capabilities of the robot. The robotic opponent implemented observes the location of the puck on the table using a camera and predicts where it will be in the future. A Cartesian …

Automated Pill Dispenser, Ryan Oderkirk, Connor Beaven, Rachelle Labrie, Josue Panchana

Williams Honors College, Honors Research Projects

The project we propose is an automated system for dispensing dosages of medication throughout the day. It will be able to alert a user when their pills need to be taken and give them the correct dosages of up to four different medications. These dosages are configurable as well as the scheduled time they are to be taken. In addition, the pill dispenser will alert users when they are low on medications and need to refill the machine.