Master of Science in Electrical Engineering (MSEE), Wright State University, 2024, Electrical Engineering

This study examines the aging effects of GaN HEMTs, focusing on the CG2H40010 device under conditions that mimic the high-power, high-frequency environments of wireless communication systems. With the increasing adoption of GaN technology in RF applications, understanding its degradation mechanisms under CW stress and modulated signal characterization is essential for predicting device lifetime and ensuring performance standards for modern communication systems. RFALT was employed to stress the device using CW signals, while key performance metrics, such as gain compression, gate leakage, ACP, and EVM, were assessed using W-CDMA signals to replicate real-world dynamic stresses. The findings reveal that CW stress accelerates thermal and electrical degradation in GaN HEMTs, while W-CDMA characterization highlights the impact of complex modulation on linearity and spectral containment. Degradation mechanisms such as ohmic contact wear and dielectric failure significantly affect performance, especially under high peak-to-average ratio conditions. This research underscores the importance of combining CW-based RFALT with modulation-specific testing to evaluate device reliability comprehensively. By addressing thermal management, enhancing dielectric materials, and employing linearization techniques, these insights pave the way for optimizing GaN HEMTs to meet the stringent requirements of 5G and future wireless communication systems.

Master of Science (MS), Wright State University, 2024, Computer Science

In today's technological landscape, hardware devices are integral to critical applications such as industrial automation, autonomous vehicles, and medical equipment, relying on advanced platforms like FPGAs for core functionalities. However, the multi-stage manufacturing process, often distributed across various foundries, introduces substantial security risks, notably the potential for hardware Trojan insertion. These malicious modifications compromise the reliability and safety of hardware systems. This research addresses the detection of hardware Trojans through side-channel analysis, utilizing power and electromagnetic signal data, combined with meta-learning techniques, specifically model stacking. By employing diverse base models and a meta-model to consolidate predictions, this non-invasive approach effectively identifies Trojans without requiring direct access to internal circuitry. The methodology demonstrates robust classification capabilities, achieving an accuracy of 88.0\%, precision of 81.0\%, and recall of 95.0\%, even on previously unseen data. The results highlight the superior performance of meta-learning over traditional detection methods, offering an efficient and reliable solution to enhance hardware security.

Doctor of Philosophy (PhD), Wright State University, 2024, Electrical Engineering

In modern wideband receiver standards, efficient frequency spectrum utilization is essential to meet demands for high data rates, reduced latency, and enhanced connectivity. The Fast Fourier Transform (FFT) stands as a pivotal technology, particularly in radar signal processing, where it supports tasks such as target detection, range estimation, and velocity estimation by analyzing the frequency content of the received radar signals. This dissertation introduces the design of an advanced digital wideband receiver featuring a high dynamic range for multiple signals, with a focus on improved performance, compact size, and reduced power consumption, implemented on an FPGA using custom hardware. Key optimizations include converting floating-point data to 10-bit integers and replacing complex multipliers in the FFT module with simplified operations. The design begins with an FFT implementation using a 12-bit analog-to-digital converter (ADC) operating at a 2 GHz sampling rate, capturing 512 data points. Improvements such as a multiple-input selection block enhance weak signal amplification while preserving dynamic range, and an upgraded square-root approximation using Chebyshev coefficients reduces FFT output errors. These advancements improve weak signal detection accuracy even in the presence of strong signals, minimizing hardware requirements. The implementation utilized the Xilinx UltraScale+ RFSoC 1275 board, which integrates both RF and digital processing components onto a single chip, offering a compact and efficient solution for wideband receiver designs. The FFT module processes sampled data every 256 ns, evaluating frequencies from 64 MHz to 940 MHz. Experimental results demonstrate the lowest detectable signal strength of 500 uVpp with an approximate dynamic range of 60 dB for a single signal. For two-tone signals, the achievable instantaneous dynamic range is about 40 dB, with the lowest detectable signal strength in the presence of the s (open full item for complete abstract)

Master of Science in Electrical Engineering (MSEE), Wright State University, 2024, Electrical Engineering

The uncertainty of the standard calibration procedure for radar cross-section (RCS) measurement is studied for different targets measured in the near-field from 550 to 700 GHz. Using common calibration spheres and squat cylinders mounted on a styrofoam pedestal at waterline (zero-degrees elevation), the calibration difference measure is determined for each target. Similarly, the difference metric is determined for square trihedral and tophat targets placed on a ground plane and measured at different elevation angles. The mean calibration measure is calculated using the dual calibration target method and repeated measurements in an anechoic chamber. The specific THz system is described and the results show how the near field scattering behaviors degrade the accuracy of the scattering measurement. Additional analysis shows the measurement uncertainty to be within a few decibels for frequencies within 580 to 650 GHz.

Master of Science, Miami University, 2025, Electrical and Computer Engineering

This work introduces a proof-of-concept implementation of a new type of Hardware Trojan (HT), which we call Glitch Hardware Trojan (GHT). We design a GHT on an Altera DE2-115 FPGA to demonstrate the attack vector of such HTs, verify its behavior and information leaking potential, and integrate it in a functional benchmark - a Serial Peripheral Interface (SPI) module. We understand that a significant portion of semiconductor fabrication is conducted overseas due to the high costs associated with the process. This global supply chain introduces several potential vulnerabilities that leave companies open to malicious actors. To stay ahead of malicious actors, we hope to create these GHTs and devise defense methods to safeguard against their threat. This will enhance the security and reliability of integrated circuits and FPGA-based applications.

Master of Science in Electrical Engineering (MSEE), Wright State University, 2024, Electrical Engineering

To address the issues of limited target data in the Synthetic Aperture Radar Automatic Target Recognition (SAR ATR) problem set, synthetic data is often used to aid in filling the gap. This paper covers an in depth look at the use of colorization, dynamic range adjustment, and target extraction as data augmentation techniques to improve the accuracy of deep learning networks trained on synthetic SAR data. The use of multiple different data augmentations combine to dramatically improve the accuracy of a common Convolutional Neural Network (CNN) over the use of standard synthetic data. A comparison of increasing fraction of measured data were used to show that the less measured data there is available the more critical these data augmentation techniques are to improve target recognition.

Doctor of Philosophy (Ph.D.), University of Dayton, 2024, Electrical and Computer Engineering

Additive manufacturing (AM) technology is rapidly advancing across diverse fields. For instance, the use of robotic arms in various AM processes has led to significant gains in printing flexibility and manufacturing scalability. However, despite these advancements, there remains a notable research gap concerning the mechanical properties of parts 3D-printed with robotic arms. This study focuses on developing a robotic fused filament fabrication (FFF) 3D-printing process with a layer resolution of 50 μm to 300 μm. The impact of the robotic printing process on the mechanical properties of printed parts is investigated and benchmarked against a commercial FFF 3D printer. In addition, we propose a novel tool path that can vary contour layer thickness within an infill layer to improve mechanical strength by minimizing air gaps between contours. SEM images suggest that this new tool path strategy leads to a significant reduction in the fraction of the void area within the contours, confirmed by a nearly 6% increase in the ultimate tensile strength. Furthermore, a novel strategy for non-planar contours is proposed, specifically designed for thin-shell 3D models. This approach aligns tool paths parallel to the Z-axis, organized into triangular segments, and utilizes planar slicing techniques. The method involves segmenting the point cloud and systematically printing non-planar contours on top of the planar contours. Axial compression testing reveals that samples produced using this strategy exhibit mechanical properties comparable to those of conventional 3D printing. However, distinct fracture patterns are observed: in conventional 3D-printed samples, fractures occur on both inner and outer surfaces, while in non-planar printed samples, fractures are confined to the inner surfaces (planar contours) and do not propagate to the outer non-planar contours. This demonstrates the potential of non-planar printing for improved structural integrity.

Doctor of Philosophy (Ph.D.), University of Dayton, 2024, Electro-Optics

The aim of this research is to explore a new concept in holography, termed digital volume reflection holography, for implementing 3D displays, numerically and optically, with excellent color discrimination. Conventionally, digital holograms are transmission holograms, recorded with a CCD camera as the interference pattern between nominally co-propagating light from a reference and the object. Realizing that a digital volume reflection hologram can be simulated through periodic longitudinal extension of the digital transmission hologram, a wavelength-multiplexed composite digital volume reflection hologram is derived from respective simulated transmission holograms. Readout from the composite digital volume reflection hologram is numerically achieved using coupled wave theory, and the dependence of diffraction efficiency during reconstruction on incident angle of the reading beam, volume grating thickness and wavelength is studied. Alternatively, optical readout of digital volume reflection hologram is possible, by using a liquid crystal spatial light modulator as a holographic optical element. This is achieved by computing the optical field at the exit of a composite digital volume reflection hologram generated numerically from multiple digital transmission holograms recorded at different wavelengths, which is then appropriately introduced as an electrical input into the liquid crystal spatial light modulator, operated in both phase only and amplitude only modes. Also in this work, the composite digital volume reflection hologram is generated from point-based methods. The main advantage of the point-based method is its ability to capture 3D visual cues effectively; however, its key disadvantage is the computational volume. A layer-based method is used with point cloud method to reduce the computational intensity. Finally, realizing the importance of phase retrieval in the imaging of phase objects or objects with 3D topograp (open full item for complete abstract)

Master of Science in Electrical Engineering, University of Dayton, 2024, Electrical and Computer Engineering

This paper presents the robotic control vision system used in Cold Spray (CS) repairs. The robot used, is a standard six axis degree of freedom (6DOF) arm to hold the CS system. Show the calibration techniques for the eye to hand vision system. Show how the pictures are analyzed with python code using OpenCV to identify the repairs, and plan the path of the CS. How an industry programmable logic controller (PLC) is used a mediator device between the computer and the robotic controller. The robot then runs a loop to go from repair to repair using the data interpreted by the vision system. This paper will also measure the accuracy of the points found and how accurate the point found are to the point transition to see if this falls in the tolerance of the cameras specifications.

PhD, University of Cincinnati, 2024, Engineering and Applied Science: Electrical Engineering

Modern heterogeneous architectures contain multiple cores that perform a variety of processing, storage, and communication tasks. The complexity of interactions among the cores and of the cores themselves introduces potential security vulnerabilities that can be exploited by malicious actors to mount a variety of attacks. To address these vulnerabilities, it is critical to conduct systematic security analysis, enforce relevant security policies, and verify designs through formal methods before fabrication. However, current SoC designs require a time-consuming and resource-intensive process to identify and verify security assets against applicable security policies. This gap highlights the need for efficient abstraction techniques that streamline the specification and verification of security policies, reducing both the verification cost and design overhead. As these complex architectures rely on information transfer between the cores, the significance of a well-established interconnect such as Network-on-Chip (NoC) is paramount. NoC architectures have gained prominence in modern computing systems for their scalability and efficiency. However, the globalization of NoC design and fabrication exposes them to security threats. The shared hardware resources between secure and malicious IPs in NoC create vulnerabilities that are exploited by the attacker to implement explicit and implicit data leakages. Quantitative analysis plays an important role in exposing vulnerabilities by quantifying packet information and traffic flow across components and pathways. It uses numerical data and mathematical models to understand complex systems, revealing patterns, and anomalies through qualitative methods. This dissertation introduces a comprehensive methodology to address the challenges associated with SoC and NoC security. First, we propose a systematic approach for security analysis using Assertion-Based Verification (ABV), focusing on cataloging SoC vulnerabilities and d (open full item for complete abstract)

MS, University of Cincinnati, 2024, Engineering and Applied Science: Electrical Engineering

The advent of the single purpose microcontrollers, coupled with the evolution in battery technology and wireless communication has accelerated the development and widespread of IoT devices. These omnipresent devices, due to their inherent benefit, have been employed in every aspect of our lives including the very critical ones from medical devices to security and defense. Despite their myriads of benefits, IoT devices have for the few years been one of the favorite targets of bad cyber actors due to diverse reasons. Securing IoT devices is challenging because they are not only deployed in remote places where control and supervision is unfeasible, but they also have limited computation resources which renders the current security infrastructure obsolete. To fix this issue methodologies that use hardware malware detectors (HMD) have been employed. An HMD is a security device deployed to detect and combat malicious software by analyzing activities at the hardware level. This methodology uses either built in performance monitor units and machine learning algorithms to create models capable of detecting malware operation or malware detection units created using different heuristics capable of differentiating benign or malware programs. One such heuristic is the negative selection algorithm from the field of artificial immune system which allows to build a detector set capable of differentiating between self and non-self-samples. Since malware is spread in diverse families and its development tends to have a rapid evolution, it is necessary to find a way to store critical detection information without requiring too much memory to detect a vast array of attacks. A methodology for reducing the size of the detector set and a partial implementation in hardware to assess the power and area implication of the reduction is proposed in this work. The methodology entitled Reduction of Detector Set for a Hardware (open full item for complete abstract)

PhD, University of Cincinnati, 2024, Engineering and Applied Science: Electrical Engineering

The comprehensive evaluation of hardware security applications poses significant challenges due to limitations in currently available benchmarks, characterized by small size, lack of scalability, and inflexibility. In this research work, we developed a new framework named Attributed Circuit Transformation (ACT), designed to address these challenges by generating synthetic benchmark circuits that are both flexible and scalable. The ACT framework comprises a new language named the ACT language and the ACT system which is based on attributed graph grammars. Attributed graph grammars have been used for generating interesting and constraint-satisfying structures in various domains of design. In this work, we demonstrate the compilation process of programs written in the ACT language into attributed graph grammars, followed by their execution using the ACT system. The proposed research aims to showcase the generation of synthetic gate-level circuit structures that meet arbitrary design constraints, both with and without utilizing frequently existing patterns from current circuits. Additionally, this work explores the generation of RTL-level circuits. Furthermore, this work demonstrates the practical applications of the generated gate-level circuits in hardware security, particularly focusing on SAT (Satisfiability), SeqSAT (Sequential Satisfiability), SCOAP Analysis, and Trojan insertion and detection. Additionally, the research involves developing RTL-level circuits customized for various types of modules and processor.

Master of Science (M.S.), University of Dayton, 2024, Electrical Engineering

As Machine Learning (ML) technology advances, the devices and servers required to collect, process, and store data are becoming increasingly complex. Tiny Machine Learning (TinyML) addresses this challenge by enabling simple ML models to run on small, low-power devices, such as microcontrollers. By processing data locally, TinyML reduces the amount of data that must be transferred to a server, resulting in greater spectral efficiency and faster response times. One practical application of TinyML is drone detection in modern warfare. A low-cost microcontroller with an integrated microphone can be equipped with a TinyML model to detect drone presence, provide immediate warnings, and operate with an extended battery life—enhancing both functionality and portability in field conditions.

Master of Science (M.S.), University of Dayton, 2024, Electrical Engineering

RGB imagery provides detail which is usually sufficient to perform computer vision tasks. However, images taken in low-light appear vastly different from well-lit imagery due to the diversity in light intensity. Polarimetric data provides additional detail which focuses on the orientation of the light rather than intensity. Scaling our classic RGB images using polarimetric data can maintain the RGB image type, while also enhancing image contrast. This allows transfer learning using pre-trained RGB models to appear more feasible. Our work focuses on developing a large dataset of paired polarimetric RGB images in a highly controlled laboratory environment. Then, we perform transfer learning on a pre-trained image segmentation model with each of our image product types. Finally, we compare these results in both well-lit and low-light scenarios to see how our polarimetrically enhanced RGB images stack up against regular RGB images.

MS, University of Cincinnati, 2024, Engineering and Applied Science: Environmental Engineering

We created the GLObal river Width from Sentinel-2 (GLOW-S), derived from Sentinel-2 imagery, to examine river width seasonality globally. GLOW-S, containing 2.1 billion observations across 797,394 river reaches for 2017-2022, represents an 8.9-fold increase in data frequency and a 2-fold increase in spatial coverage compared to previous studies, enabling the seasonality analysis of river width. Results indicate that 9.4% of rivers maintain steady widths, 30.2% exhibit sinusoidal seasonality, and 44.4% display non-sinusoidal seasonal patterns; additionally, we identified 16% of global rivers that are non-seasonal, suggesting complex environmental interactions that, along with the non-sinusoidal but seasonal rivers, require further targeted research. Larger, more regulated rivers tend to have steadier widths than smaller, free-flowing ones. The timing of peak widths varies regionally, with 31.4% occurring in April and June. This study has important implications for freshwater hydrology and ecosystems (e.g., nutrient and carbon exchange) and provides data support for future studies in this avenue and beyond.

Doctor of Philosophy, Case Western Reserve University, 0, EECS - Electrical Engineering

Timely characterization of the hemostatic system at the point-of-care/point-of-injury (POC/POI) is clinically important in traumatically injured and critically ill patients to guide therapeutic interventions and improve survival outcomes. This work advances the development of a microfluidic dielectric sensor, termed ClotChip, as a platform technology for POC/POI assessment of whole blood coagulation by enhancing its readout characteristics to precisely identify a range of blood coagulation disorders, including dysfunctions in fibrin formation and fibrinolysis. Specifically, two new distinct readout parameters in the ClotChip readout curve, namely, lysis time (LT) and maximum lysis rate (MLR) are identified and shown to be sensitive to the fibrinolytic status in whole blood. LT identifies the time that it takes from the onset of coagulation for the fibrin clot to mostly dissolve in the blood sample during fibrinolysis, whereas MLR captures the rate of fibrin clot lysis. A third new parameter, Smax, is also identified that represents the maximum permittivity slope and is shown to be sensitive to fibrin-polymerization defects during clot formation. These findings are validated through correlative measurements with rotational thromboelastometry (ROTEM) – a clinical, viscoelastic-based, global assay of blood coagulation. This work also includes the development of next-generation miniaturized dielectric coagulometry featuring multiple channels with bioreagent-functionalized electrodes that uniquely and specifically elicit differential responses from the multifactorial process of blood coagulation. Specifically, a microfluidic sensor is developed with physisorption of tissue factor and aprotinin on the electrode surfaces to probe the fibrinolytic function. The dual-coated microsensor can detect the hemostatic rescue in the hyperfibrinolytic profile of whole blood coagulation induced by tissue plasminogen activator as well as the hemostatic dysfunction due to concurrent pl (open full item for complete abstract)

Doctor of Philosophy (Ph.D.), University of Dayton, 2024, Electrical and Computer Engineering

This dissertation investigates p-polarized electromagnetic (EM) wave propagation through chiral metamaterials using Fresnel coefficients (FCs), examining thereby such characteristics as possible emergence or mitigation of Brewster phenomena, total internal reflection (TIR) or the more unusual inverse TIR (ITIR) effect at a chiral-achiral (CAC) boundary, and the behavior of chiral Fabry-Perot type slabs (especially in the thin film limit) with particular attention to their resonance characteristics. Chiral materials belong among the broader group of metamaterials known for their ability to interact with polarized light, leading to bimodal propagation of right- and left-circularly polarized (RCP and LCP) states with interesting and potentially useful optical properties. In this dissertation, the propagation of a parallel (p-) polarized plane wave across an achiral/chiral (ACC) interface is investigated under different material index assumptions (rarer to denser (R→D) and denser to rarer (D→R) configurations). Subsequently, a chiral/achiral (CAC) interface was investigated using the transmitted circularly polarized waves out of the first interface as incident waves for the second interface. Multiple optical phenomena have been found to be tunable and/or controllable by adjusting the material parameters such as dimensionless chirality factor, medium index, and/or material thickness. These optical phenomena include Brewster effect, TIR and ITIR, evanescence and tunable critical angles. Since chiral materials are inherently lossy material, following an initially lossless analysis with some related results, the problem was extended to include dielectric losses (defined via the imaginary part of the complex chiral dielectric permittivity). With dielectric losses present, it turns out that the EM boundary conditions with complex material parameters and the related phase matching led to entirely modified algebraic results (several of which, as with the lossless cases) are c (open full item for complete abstract)

Master of Science, The Ohio State University, 2024, Electrical and Computer Engineering

As machine learning (ML) and artificial intelligence (AI) approaches gain prominence and make significant impacts within biomedical science, the need for a specialized knowledge base to archive published research findings in this field has become increasingly apparent. While PubMed remains a widely used database within the biomedical science community, it is not specifically designed to annotate ML/AI applications in biomedical research. To address this gap, Biomed-ML, a comprehensive knowledge portal for ML and AI applications, was developed. Biomed-ML was generated using BioBERT models trained on a meticulously curated, manually labeled dataset, achieving an F1-score of 0.953 in classifying ML/AI applications. The portal categorizes articles into Clinical Science (n=49,627) and Basic Science (n=25,319) studies, enabling researchers to query ML/AI-related publications and applications through integrated Medical Subject Headings (MeSH) terms. For the first time, researchers in the biomedical science community have access to a high-accuracy tool to explore ML/AI applications in areas such as disease diagnosis, therapeutic outcome prediction, pre-clinical drug screening, target discovery, article analysis techniques, and other advanced research domains. This resource bridges a critical gap, enhancing accessibility and fostering innovation in biomedical research.

Master of Science, The Ohio State University, 2024, Electrical and Computer Engineering

Automated driving needs lane-level accurate localization. However, automated driving systems face significant challenges in environments where GPS signals are unavailable or compromised. Several techniques have been introduced over time to address this issue. However, each technique presents its own set of challenges.To address the lane-level localization challenges, this study proposes a kinematic dead reckoning system utilizing vehicle onboard sensor data, which is crucial for vehicle operation itself. Onboard sensors provide data such as steering angle, steering rate, yaw rate, and wheel speed sensors through the vehicle's Controller Area Network (CAN). However, dead reckoning is susceptible to drift over time, compromising localization accuracy. To mitigate this drift, an innovative arc-length-based map matching method is introduced, which leverages a digital 2D map of road and lane geometry to correct the dead reckoning estimates.The proposed methodology enhances vehicle localization by combining the temporal prediction of a kinematic model with spatial information from static map data, effectively correcting drift without GPS support. This approach was tested in multiple safety-critical scenarios suggested by NHTSA in distinct road geometry, speed, and maneuvers, demonstrating consistent localization accuracy. The overall results showed reliable drift correction for all tested scenarios. Furthermore, we evaluated the outage performance for each scenario at different times during the scenario test, revealing a bound error in the localization method. Furthermore, the proposed method calculates a confidence interval to identify overestimation and underestimation.This novel arc-length-based map matching ensures continuous and dependable navigation for automated vehicles in GPS-denied situations, significantly enhancing safety and operational reliability. The findings of this study highlight a scalable and effective solution to maintain automated vehicle localizati (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Electrical and Computer Engineering

Medical imaging has revolutionized patient care, enabling noninvasive and detailed visualization critical for diagnosing and treating numerous conditions. However, interpreting complex medical images remains challenging, requiring specialized expertise often unavailable in rural or underserved areas. Advances in artificial intelligence (AI) have introduced significant potential to automate and enhance medical imaging, yet the size and computational demands of current deep learning models hinder scalability in real world applications, particularly in resource constrained settings. This dissertation addresses these challenges by developing advanced, lightweight neural network architectures that bring efficient and high-performing AI-driven medical imaging solutions to a broader audience. In particular we explore various deep learning approaches to systematically address key bottlenecks, drawing on lessons learned from each method to drive improvements across others. We investigate Graph Neural Networks (GNNs) as a compact, efficient alternative to traditional deep learning models, enabling targeted analysis that minimizes unnecessary background processing. We further enhance contextual understanding through lightweight Vision Transformers, optimizing self attention complexity to improve model efficiency. Additionally, we introduce an efficient Convolutional Neural Network (CNN) design that captures larger receptive fields without increasing computational burden, supporting both 2D and 3D imaging. Finally, we propose hybrid CNN Transformer architectures that combine CNNs' local feature extraction strengths with Transformers' global context capabilities, offering scalable, high performing solutions for diverse medical imaging applications. These contributions advance the development of accessible, AI based imaging tools that can be deployed on handheld devices and support critical applications in underserved areas, military operations, and emergency response, ult (open full item for complete abstract)