Digital Commons Network^™

Robust Visual Observer And Controller Design For System Modeled On Se(3) With Camera Measurements, Tong Zhang

Electronic Theses and Dissertations

This dissertation presents a controlling method that utilizes visual observation to address the challenge of driving an autonomous vehicle with a kinematic model on SE(3), by using the data obtained from the onboard cameras. This controlling framework is applicable when direct retrieval of the vehicle's attitude information is not possible or when precision control of the vehicle is necessary. However, addressing several challenges to attain closed-loop stability for the observer-based control for autonomous vehicles on SE(3) is essential. 1. The analysis is further complicated by the presence of multiple equilibria in the error dynamics associated with the suggested method. 2. …

Memristor-Based Digital Circuit Design, Khalid Mohammed Alammari

Electronic Theses and Dissertations

Further miniaturization in CMOS technology has faced severe challenges, such as increased leakage power and reduced circuit reliability, which has caused fundamental restrictions on advancing efficient computing architecture. These problems can be addressed by the new emerging devices known as memristors owing to their nanoscale size and the ability to integrate with the exciting CMOS technology. Memristors are passive devices with variable resistance. The resistance value will remain constant when no electrical field is applied, giving this device a unique behavior by saving its last state. This unique behavior makes this device very appealing to a wide variety of applications …

Skin Cancer Detection By Deep Learning Algorithms, Faezeh Mohammadi Aydoghmishi

Electronic Theses and Dissertations

Skin cancer, characterized by the abnormal growth of skin cells, is a globally prevalent and serious condition. Despite advancements in digital diagnosis techniques, many existing skin cancer detection methods often fail to achieve satisfactory accuracy levels. In our first study, we propose a novel superpixel-based segmentation method that significantly surpasses the traditional k-means clustering in performance. After segmentation, we utilize Convolutional Neural Networks such as VGG16, ResNet50, DenseNet, Xception, Inception, and Mobilenet for feature extraction and classification. We also conducted a comparative analysis with other existing techniques. Our results suggest that our superpixel-based segmentation method notably enhances the accuracy of …

Enhancing Breast Cancer Detection Through Combination Of Contrastive Learning And Adversarial Domain Adaptation, Mahnoosh Torabi

Electronic Theses and Dissertations

The most common cancer diagnosed worldwide is breast cancer and early detection is essential for reducing mortality. The best standard for early detection of breast cancer is digital mammography, which can aid physicians in treating the illness when it is still curable. However, inaccurate mammography diagnoses are frequent and can cause patients to undergo unnecessary examinations and therapies. This study aims to explore deep-learning techniques that can be utilized to implement and train a model to identify breast cancer cases in mammograms. Current deep learning-based diagnostic techniques are hindered by two fundamental issues: the expensive and time-consuming task of data …

Digital Realization Of Spiking Neural Network, Mahsasadat Seyedbarhagh

Electronic Theses and Dissertations

Spiking Neural Network which is known as the third generation of artificial neural networks can imitate the same biological patterns of human brain. Neuromorphic Computing is a multidisciplinary research topic which employs digital and analog platforms to implement bio-inspired systems and is able to operate parallelly with low-power consumption. This research study presents the digital design for several bio-inspired systems such as biophysical spiking neuron, calcium signaling astrocyte, and a neuron-astrocyte interaction models using various approximation methodologies. Astrocytes have an essential impact within the neural network and their role is to maintain support for neuron, control the ion hemostasis and …

A Reinforcement Learning Algorithm For Training A Spiking Neural Network Agent, Seyede Narjes Zamani

Electronic Theses and Dissertations

Over the past two decades, Spiking Neural Networks (SNNs), as the third generation of Artificial Neural Networks, have gained widespread use due to their ability to closely mimic human brain neuron behavior. Being spike-based and resembling biological neurons, SNNs have demonstrated superior energy efficiency compared to conventional counterparts. However, to fully harness their potential, effective training methods are essential. Despite significant progress in the field of biological implementation of SNNs, there remains a need for research to identify the most optimal learning approach. To address this, we employed Policy-based stochastic reinforcement learning to train a spiking neural network using Izhikevich …

New Memristive Architecture For Digital Circuits Design, Farzad Mozafari

Electronic Theses and Dissertations

Before 1971, all the electronics were based on three basic circuit elements. Until a professor from UC Berkeley reasoned that another basic circuit element exists, which he called memristor; characterized by the relationship between the charge and the flux-linkage. A memristor is essentially a resistor with memory. The resistance of a memristor (Memristance) depends on the amount of current that is passing through the device. In 2008, a research group at HP Labs succeeded to build an actual physical memristor. HP's memristor was a nanometer scale titanium dioxide thin film, composed of two doped and undoped regions, sandwiched between two …

Mitigation Of Asymmetrical Currents In Permanent Magnet Synchronous Machine With Unbalanced Stator Winding Impedance For Electric Vehicle Application, Khagendra Thapa

Electronic Theses and Dissertations

The electric drive system of a permanent magnet synchronous machine (PMSM) with stator winding impedance asymmetry draws unbalanced three-phase stator currents if the conventional current control scheme is employed. Due to these asymmetrical stator currents, some issues such as uneven heating of the inverter phases and switches, additional losses in the motor due to skin and proximity effects, vibration and noise, uneven heating of the motor phases, and torque ripple appear in the PMSM drive. These effects reduce the reliability, efficiency, and lifespan of the machine. The conventional proportional-integral (PI) current controller in a PMSM drive is unable to mitigate …

Cognitive Load Detection Based On Speech In Automation Driving Systems, Obiajuru Onwunamoghor Ninduwezuor

Electronic Theses and Dissertations

Driver inattention is one of the leading causes of road accidents and fatal crashes. To mitigate these risk, Driver Monitoring System (DMS) has been developed and extensively experimented by the automobile industry. An overview of DMS is presented in this thesis with specific focus on driver inattention and cognitive state of the drivers. Research on the sources of driver inattention are reviewed, and a comprehensive classification is provided. Various safety systems that measure driver inattention based on driving behavior, hybrid measures and physiological measures are investigated. In particular, a non-invasive speech-based measure of physiological signal for detecting the cognitive load …

Machine Learning Approaches For Healthcare Analysis, Bashier Omar Elkarami

Electronic Theses and Dissertations

Machine learning (ML)is a division of artificial intelligence that teaches computers how to discover difficult-to-distinguish patterns from huge or complex data sets and learn from previous cases by utilizing a range of statistical, probabilistic, data processing, and optimization methods. Nowadays, ML plays a vital role in many fields, such as finance, self-driving cars, image processing, medicine, and Speech recognition. In healthcare, ML has been used in applications such as the detection, prognosis, diagnosis, and treatment of diseases due to Its capability to handle large data. Moreover, ML has exceptional abilities to predict disease by uncovering patterns from medical datasets. Machine …

High Radix And Efficient Hardware Implementation Of Modular Integer Multiplication For Iot Cryptosystems, Fahimeh Pakzadalinodehi

Electronic Theses and Dissertations

This thesis presents a new design for a radix-4 Montgomery Modular Multiplier that is based on field-programmable gate array (FPGA) implementation. This work is an improvement of the radix-4 Montgomery Modular Multiplier structure that requires no multiplication or subtraction operations in the computation process, resulting in a reduced critical path delay and increased maximum frequency. The proposed Montgomery modular multiplication design was implemented on Virtex-7 FPGA platform. The final result shows that this work runs one complete modular multiplication for 256-bit operands, in 0.566 micro seconds with maximum clock frequency of 256.5 MHz by consumption of 4534 number of lookup …

Multicriteria Consensus Models To Support Intelligent Group Decision-Making, Hossein Hassani

Electronic Theses and Dissertations

The development of intelligent systems is progressing rapidly, thanks to advances in information technology that enable collective, automated, and effective decision-making based on information collected from diverse sources. Group decision-making (GDM) is a key part of intelligent decision-making (IDM), which has received considerable attention in recent years. IDM through GDM refers to a decision-making problem where a group of intelligent decision-makers (DMs) evaluate a set of alternatives with respect to specific attributes. Intelligent communication among DMs aims to give orders to the available alternatives. However, GDM models developed for IDM must incorporate consensus support models to effectively integrate input from …

Computational Analysis Investigation Of A Piezoelectrically Actuated Micropump For Air Sniffing Applications, Yameema Babu Lopez

Electronic Theses and Dissertations

Micro-Electro-Mechanical Systems (MEMS) refers to a technology which uses microfabrication technology to fabricate miniaturized devices and systems. MEMS based technologies have paved the way for miniaturization and manufacturing of devices with applications ranging from biological to fluid engineering. This interdisciplinary nature of MEMS has given rise to an emerging field called microfluidics which studies devices that pump, control and sense small volumes of fluids. This work aims to investigate a piezoelectric micropump that can mimic mammalian olfaction for e-nose applications. A comprehensive literature review was conducted on all the available micropumps. A piezoelectric micropump was chosen as the candidate for …

Control And Protection Solutions For Resilient Protective Relaying Of Modern Power Systems, Abdallah Alaa Mohieldien Aboelnaga

Electronic Theses and Dissertations

Renewable energy sources (RESs) are permeating the power grid due to their importance in reducing air pollution and fuel consumption. These sources require synchronization with the power grid using inverters that must meet grid code (GC) requirements. During fault conditions, GCs enforce inverter interfaced RESs (IIRESs) to follow reactive current generation (RCG) requirements to enhance grid stability. However, it could adversely affect protection functions, e.g., phase selection methods (PSMs), operations. The main objective of this dissertation is to enhance the power system resiliency by determining the faulty phase(s) accurately. This is achieved by investigating the root causes behind the failure …

Unsupervised Bidirectional Mr To Ct Synthesis Based On Generative Adversarial Networks, Jiayuan Wang

Electronic Theses and Dissertations

Magnetic resonance (MR) and computer tomography (CT) images are two typical types of medical images that provide mutually-complementary information for accurate clinical diagnosis and treatment. However, obtaining both images may be limited due to some considerations such as cost, radiation dose, and modality missing. Recently, medical image synthesis has aroused gaining research interest to cope with this limitation. In this thesis, we proposed unsupervised bidirectional learning models based on generative adversarial networks (GANs) to synthesize medical images from unpaired data. The first model is dual contrast CycleGAN (DC-cycleGAN), where a dual contrast (DC) loss is introduced into the CycleGAN's discriminators …

Single Hydrophone Underwater Localization Approach In Sallow Waters, Faraz Talebpour

Electronic Theses and Dissertations

Applications of underwater signal processing are essential for environmental monitoring. Remote monitoring and passive sound source localization in an underwater environment can provide great insight into geological studies, environmental changes and marine lives monitoring. While various methods are available for Localization, they mostly employ arrays of hydrophones, requiring synchronization or prior knowledge of the source signals, which can prove costly, complicated, and hard to maintain. Remote monitoring applications require very high-range passive localization methods; and, given the frequency-selective nature of ambient noise and other channel parameters, current localization methods have short-distance range estimation or high localization error for long distances. …

2d Hybrid Analytical Model Based Performance Optimization For Linear Induction Motors, Michael Thamm

Electronic Theses and Dissertations

In this thesis the domain of double-layer, single-sided, 3-phase, integral slot winding, linear induction motor (LIM)s is analyzed. Motor meta parameters such as slots and poles are difficult to optimize since they drastically effect the configuration of the motor and require heuristic optimization implementations.

A non-dominated sorting genetic algorithm II (NSGAII) was implemented with the Platypus-Opt Python library. It serves as a robust, yet flexible integration while maximizing thrust and minimizing the mass of each motor iteration. Each iteration was accurately modelled using the hybrid analytical model (HAM), producing the necessary performance parameters for the NSGAII’s objective function. Field plotting …

Rapid Prototyping And Functional Verification Of Power Efficient Ai Processor On Fpga, Vivek Liladhar Ladhe

Electronic Theses and Dissertations

Prototyping a design on a Field Programmable Gate Array (FPGA) involves different stages such as developing a design, performing synthesis, handling placement and routing and finally generating the programming bit file for the FPGA. After successful completion of the above stages, it is important to functionally verify the design. This thesis addresses the challenges involved in rapid prototyping and functional verification of a low power AI processor provided by the industry partner. This research also addresses the methodology used in generating programming bit file and testing the design. Traditional method of testing a design using RTL level testbench utilises more …