Electrical and Computer Engineering Commons^™

Reinforcement Learning Based Security Schemes For Distributed Ai Systems, Ashan Chamath Gunawardena

Dissertations and Doctoral Documents, University of Nebraska-Lincoln, 2023–

Distributed machine learning (DML) is a component of modern intelligent systems, enabling collaborative training across devices such as mobile clients, vehicles, and edge networks. However, the decentralized nature of these systems introduces vulnerabilities, particularly data poisoning attacks that compromise model integrity and degrade performance. Traditional defenses, such as statistical filtering, robust aggregation, and privacy-preserving techniques, often struggle to adapt to overwhelming adversaries or operate under strict privacy and real-time constraints. This dissertation proposes the use of reinforcement learning (RL) and deep reinforcement learning (DRL) based misbehavior detection schemes that dynamically identify poisoning attempts in distributed AI systems, including federated learning, …

Rapid State-Of-Health Estimation Of Batteries Using Machine Learning With Limited Eearly-Discharge Voltage Data, Mohammad Bakhtiari

Durham School of Architectural Engineering and Construction: Dissertations, Theses, and Student Research

The utilization of lithium-ion batteries has been rapidly expanding across diverse sectors, including electric transportation, stationary energy storage systems, and the built environment. Ensuring a high level of reliability in these applications is essential, as the performance and safety of such systems depend strongly on the accurate assessment of the battery’s State of Health (SOH). Conventional SOH estimation techniques—often based on complex electrochemical models or extensive laboratory testing—tend to require a large number of measurements, advanced instrumentation, and high computational cost. These factors make them impractical for large-scale deployment or real-time monitoring. This study introduces a simplified machine-learning-based approach for …

A Study On Utilizing Coherently Coupled Orbital Angular Momentum Beams For Maritime Sensing And Communication, Evan Robertson

All Dissertations

A large portion of the world is covered in water which introduces a couple of key challenges in communication and sensing systems. The impact of particulates in the water will limit the ability to successfully transmit information through the water. The study of how the channel impacts specific frequencies and the ability to transmit more information at a single time can limit the impact of this environment in how it degrades an optical communication system. In sensing applications, it is important to detect information related to an object’s motion which will either be towards or away from a system, or …

Application Of Natural Language Processing And Machine Learning For Analyzing Mining Accident Reports And Automating The Process Of Root Cause Analysis, Siddhartha Agarwal, Y. P. Chugh, Atul Singh, Vikram Sakinala, Ayan Mukherjee, Balbir Prasad, Cihan Dagli, Yuhao Zou

Engineering Management and Systems Engineering Faculty Research & Creative Works

Coal mining accidents are a major concern worldwide, necessitating effective safety measures and comprehensive analysis to prevent future accidents. Our proposed solution is the first attempt for Indian mines, inspired by the potential of Natural Language Processing (NLP) that can read and analyze vast repositories of accident records in seconds. In combination with machine learning (ML), NLP algorithms can extract unstructured text by eliminating manual data entry errors, reading poorly scanned reports, and understanding multiple versions of the event and cluster documents based on types that would otherwise take months to collate. In the case of accident records, it can …

Data Driven Design Of Ultra High Performance Concrete Prospects And Application, Bryan K. Aylas-Paredes, Taihao Han, Advaith Neithalath, Jie Huang, Ashutosh Goel, Aditya Kumar, Narayanan Neithalath

Electrical and Computer Engineering Faculty Research & Creative Works

Ultra-high-performance concrete (UHPC) is a specialized class of cementitious composites that is increasingly used in various applications, including bridge decks, connections between precast components, piers, columns, overlays, and the repair and strengthening of bridge elements. The mechanical and durability properties of UHPC are significantly influenced by factors such as low water-to-binder ratios, the inclusion of supplementary cementitious materials (SCMs), and fiber reinforcement. Machine learning (ML) has been employed to predict the performance of UHPC and optimize its mixture designs by using various raw materials. This study first provides a comprehensive review of ML applications in UHPC, focusing on predicting workability, …

Ai-Driven Electromagnetic Design And Performance Prediction Of Microstrip Antennas, Eduardo Javier Vazquez

Theses and Dissertations

This thesis investigates the application of Deep Learning to automate and accelerate microstrip antenna inverse design. The initial investigation was to predict microstrip antenna performance from its geometry parametric input, and found out that forward prediction with adopted geometric representation results in ill-posed scenario, high ambiguity and unstable mapping. The study later mostly focuses on the inverse prediction by machine learning from S11 parameter input to predict antenna patch geometry parameters instead.

A dataset of 5,000 ANSYS HFSS simulated antennas (later filtered to 4,136 valid samples) was generated using cubic spine -described geometry profiles. Multiple neural architectures were adopted for …

Extreme Bandgap Recessed-Gate Metal Oxide Semiconductor Heterostructure Field Effect Transistors With Drain Current 0.28 A Mm−1 And Threshold Voltage −1.5 V, Abdullah Al Mamun Mazumder, Abdullah Mamun, Kenneth Stephenson, Kamal Hussain, Tariq Jamil, Grigory Simin, Asif Khan

Faculty Publications

Herein, the first demonstration of hybrid high-k oxide (ZrO2-Al2O3) incorporation into extreme bandgap (EBG) Al_0.87Ga_0.13N/Al_0.64Ga_0.36N metal-oxide-semiconductor heterostructure field-effect transistors (MOSHFETs) is presented, with both planar and recessed-gate designs on the same AlN/sapphire template with a state-of-the-art low contact resistance of 1.4 Ω mm (contact resistivity, ρc ≈ 5.7 × 10−6 Ω cm2). The recessed-gate MOSHFETs achieve a threshold voltage shift of ΔVTH = 5.8 V, highlighting improved channel control. Static output measurements reveal a peak drain current (IDS) of 340 mA mm⁻¹ for the planar …

An Audio Data-Driven Roadway Digital Twin And Its Underlying Framework For A Digitized Transportation Construction Environment, Anisha Deria, Pedro J. Chacon Dominguez, Yong-Cheol Lee, Jin W. Choi

Michigan Tech Publications

The increasing need to build and maintain transportation systems has led project managers to manage multiple projects simultaneously. Roadway projects often entail several miles of job site, making it difficult to keep track of progress and maintenance activities. To improve the situation, this study proposes an audio data-driven roadway digital twin framework for real-time and remote monitoring of construction projects. The latent characteristics of a digital twin required for establishing a digitized work environment were investigated. As a primary method of seamlessly linking virtual and physical environments, audio data classified and analyzed by deep neural network (DNN) has been employed …

Complementary Color Laser Illumination For Perceptual Contrast Enhancement In Structurally Colored Samples, Tomoshree Dash

All Theses

Structural color, a phenomenon arising from nanoscale interaction between light and materials, has the potential to unveil biological, chemical, physical, and mechanical characteristics of the sample. It enables direct visualization of the sample by human users with high spatiotemporal resolution. However, its impact is often constrained by challenges in the perceptual differentiation of subtle color variations. This thesis introduces complementary color laser illumination (C²LI) as an imaging technique that enhances color perception of the Human Visual System (HVS) by accessing the psychophysical non-linearities in chromatic color perception. C²LI offers a platform optimized for the HVS by …

Optimization Of Size And Siting Of Distributed Generation In Unbalanced Distribution Systems: A Literature Review, Pema Dorji, Stefan Lachowicz, Octavian Bass

Research outputs 2022 to 2026

Renewable energy sources (RES) are essential for meeting the rising global electricity demand while reducing greenhouse gas emissions from conventional generation. As traditional systems approach capacity saturation, the integration of RES into power grids becomes increasingly vital. However, the intermittent and variable nature of RES introduces significant technical, economic, and operational challenges. This review focuses on the optimal planning and integration of distributed generation in unbalanced distribution systems, which more accurately reflect real-world power network conditions. Emphasis is placed on siting and sizing strategies aimed at enhancing voltage stability, minimizing power losses, and reducing system costs and emissions. The review …

Design Of A Novel Robust Adaptive Fractional-Order Model Predictive Controller For Boost Converter Using Grey Wolf Optimization Algorithm, Chao Peng, Seyyed Morteza [email protected] Ghamari, Hasan Mollaee, Omid Rezaei

Research outputs 2022 to 2026

Boost converters play a crucial role in power electronics but present control challenges due to their non-minimum phase behavior and nonlinear dynamics at high switching frequencies. To address these issues, this work proposes a Fractional-order adaptive Model Predictive Control (FO-MPC) framework incorporating Exponential Regressive Least Squares (ERLS) for system identification. Traditional MPC frameworks often rely on accurate mathematical models, which are difficult to obtain in real-world scenarios. This adaptive modelling approach based on ERLS identification method eliminates the need for precise system models, improving robustness and adaptability under parameter variations. Additionally, a FO derivative term enhances damping, stability, and noise …

Machine Learning Assisted Quality Control In Metal Additive Manufacturing: A Review, Zeqi Hu, Changlin Huang, Lechun Xie, Lin Hua, Yujie Yuan, Lai Chang Zhang

Research outputs 2022 to 2026

Additive manufacturing (AM) promotes the production of metallic parts with significant design flexibility, yet its use in critical applications is hindered by challenges in ensuring consistent quality and performance. Process variability often leads to defects, insufficient geometric accuracy and inadequate material properties, which are difficult to effectively manage due to limitations of traditional quality control methods in modeling high-dimensional nonlinear relationships and enabling adaptive control. Machine learning (ML) offers a transformative approach to model intricate process-structure-property relationships by leveraging the rich data environment of AM. The study presents a comprehensive examination of ML-driven quality assurance implementations in metallic AM. First, …

Next-Generation Underwater Localization: Artificial Intelligence-Based And Energy-Aware Approaches, Mainul Islam Chowdhury, Quoc Viet Phung, Iftekhar Ahmed, Walid K. Hasan, Daryoush Habibi

Research outputs 2022 to 2026

Designing accurate, reliable, and energy-efficient localization techniques for underwater acoustic networks is highly challenging due to factors such as large propagation delays, the absence of Global Positioning System (GPS), node mobility, and limited acoustic link capacity. In any underwater sensor network (UWSN) monitoring application, data collected by underwater nodes becomes more meaningful when accompanied by location information. However, traditional localization methods often rely on geometric models and statistical filters that are highly sensitive to sensor noise and communication constraints. Energy consumption is another primary concern in UWSNs, not only because replacing and recharging underwater batteries are challenging, but also due …

Data-Driven Evaluation Of Sustainable Waste-To-Energy Pathways For Intelligent Urban Systems, Izech Brian O. Edwin, King Harold A. Recto

Electronics, Computer, and Communications Engineering Faculty Publications

The handling of municipal solid waste (MSW) in swiftly urbanizing Philippine cities poses intricate energy and governance challenges. In Baguio City, reliance on landfills has reached critical levels due to diminishing capacity, rising transport costs, and opposition to trash transfers by nearby LGUs. Although shaped by unique topographical and governance constraints, Baguio’s situation reflects issues faced by other rapidly growing Philippine cities; therefore, analyzing it offers insights for national MSW decision-making. This study applies a triple bottom line (TBL) framework to assess three management scenarios: (1) Status Quo, where all MSW is landfilled with no energy recovery; (2) Landfill with …

Investigating The Efficiency Of Ingan P-N-P-N Homojunction Solar Cells, Moath Alhejji, Mohammad Alavijeh, Jacob Kupernik, Mirsaeid Sarollahi, Abbas Jammali, Seyed Taghavi, Reem Alhelais, Md Hel Uddin Maruf, Morgan Ware

Electrical Engineering and Computer Science Faculty Publications and Presentations

This research investigates the development of a novel p-n-p-n homostructure solar cell, through semiconductor simulations using the Nextnano software. InGaN was used as a model system in order to achieve a bandgap with optimized efficiency for a p-n homojunction solar cell. By increasing the uniform doping concentration from 1.5*10(16) cm(-3) to 1.5*10(17) cm(-3), the open circuit voltage (V-oc) increased while the short-circuit current density (J(sc)) decreased, as expected in simple p-n junctions. The p-n-p-n structure achieved a peak efficiency of 32.91% at a doping level of 6.5*10(16) cm(-3), a similar to 7% improvement over a conventional p-n junction's 25.31% efficiency …

Bridging Cybersecurity Practice And Law: A Hands-On, Scenario-Based Curriculum Using The Nice Framework To Foster Skill Development, Colman Mcguan, Aadithyan Vijaya Raghavan, Komala M. Mandapati, Chansu Yu, Brian Ray, Debbie Jackson, Sathish Kumar

Electrical and Computer Engineering Faculty Publications

In an increasingly interconnected world, cybersecurity professionals play a pivotal role in safeguarding organizations from cyber threats. To secure their cyberspace, organizations are forced to adopt a cybersecurity framework such as the NIST National Initiative for Cybersecurity Education Workforce Framework for Cybersecurity (NICE Framework). Although these frameworks are a good starting point for businesses and offer critical information to identify, prevent, and respond to cyber incidents, they can be difficult to navigate and implement, particularly for small-medium businesses (SMBs). To help overcome this issue, this paper identifies the most frequent attack vectors to SMBs (Objective 1) and proposes a practical …

Robust Data-Driven Predictive Control Of Nonlinear Systems Under Modeling Uncertainty, Pegah Ghafghanbari

All Dissertations

Data-driven predictive control enables designing controllers directly from data, making it attractive for complex systems with hard-to-model dynamics. However, practical deployment is challenged by modeling inaccuracies and changing operating conditions. This dissertation develops predictive control frameworks that incorporate robustness and adaptability to address these issues in uncertain nonlinear systems.

The first part employs the Linear Parameter-Varying (LPV) framework, which represents nonlinear dynamics through simple linear form representation. To characterize the plant-model-mismatch often caused by limited data and numerical calculations, Bayesian Neural Networks (BNNs) are used, and their uncertainty estimates are integrated into two robust control approaches. The first is a …

Towards Trustworthy Federated Learning, Alina Basharat

Theses and Dissertations

Federated learning is a collaborative training model in which multiple clients optimize aglobal model by transmitting updates to a coordinating server while keeping raw data on-device, thereby reducing direct data exposure and enabling iterative global improvement. However, the iterative communication process is vulnerable to malicious attackers that either deliberately destroy the model or curious to infer raw data. Moreover, learning from multiple agents may result in unfair results. To enhance trustworthiness within this setting, we employ two-sided norm-based screening (TNBS) that removes both abnormally large and abnormally small updates, pair it with a q-fair objective to emphasize high-loss (disadvantaged) clients, …

Realization Of Deterministic Quantum Circuits For Non-Deterministic Or Incompletely Specified Quantum State Machines, Manjith Kumar, Marek Perkowski

Electrical and Computer Engineering Faculty Publications and Presentations

In classical logic design, there are machine learning methods based on converting a set of input-output traces to non-deterministic automata that are then converted to deterministic automata and synthesized using logic gates. This approach has not yet been extended to quantum automata. In this paper, we present a method to convert a set of input-output traces to a non-deterministic automaton, which is then converted to an incompletely specified multi-output Boolean function. The existing logic synthesis approaches for designing quantum circuits are insufficient to handle incompletely specified functions. So, we present a novel algorithm to synthesize logic functions with don’t cares …

Hallucination Techniques For Self-Supervised Synthetic Datasets For Mobile Robots, Wyatt D. Colburn

Master's Theses

Classical techniques in autonomous navigation struggle in tightly constrained spaces. Machine learning has been shown to perform better in these difficult environments but most techniques require large amounts of navigation experience for training. Using a new machine learning paradigm learning from hallucination (LfH), training data can be collected in a safe environment and not require supervision. Data is collected in real time while an agent performs a random walk in free space, supervision is not required as there are no obstacles for the robot to run into. After a random walk a post processing pipeline will hallucinate a safety corridor …