Computer Engineering Commons^™

Optimized Learning Using Fuzzy-Inference-Assisted Algorithms For Deep Learning, Miroslava Barua

Open Access Theses & Dissertations

For years, researchers in Artificial Intelligence (AI) and Deep Learning (DL) observed that performance of a Deep Learning Network (DLN) could be improved by using larger and larger datasets coupled with complex network architectures. Although these strategies yield remarkable results, they have limits, dictated by data quantity and quality, rising costs by the increased computational power, or, more frequently, by long training times on networks that are very large. Training DLN requires laborious work involving multiple layers of densely connected neurons, updates to millions of network parameters, while potentially iterating thousands of times through millions of entries in a big …

Development Of An Automated Electronic Prototyping System, Cesar Yahir Sanchez Zambrano

Open Access Theses & Dissertations

Prototyping systems with interconnected components can be a time and resource expensive process. The process consists of three main phases (design, build and analysis) with each having their own associated cost. For the case of electronic circuits, the building phase is the costliest phase among the three, being prone to human errors which causes the circuit to fail. All three phases of the prototyping process are important. However, often a disproportionate amount of time is spent on the build phase due to the difficulty of making and troubleshooting circuits by hand. In this thesis we will discuss a system that …

Hardware For Quantized Mixed-Precision Deep Neural Networks, Andres Rios

Open Access Theses & Dissertations

Recently, there has been a push to perform deep learning (DL) computations on the edge rather than the cloud due to latency, network connectivity, energy consumption, and privacy issues. However, state-of-the-art deep neural networks (DNNs) require vast amounts of computational power, data, and energyâ??resources that are limited on edge devices. This limitation has brought the need to design domain-specific architectures (DSAs) that implement DL-specific hardware optimizations. Traditionally DNNs have run on 32-bit floating-point numbers; however, a body of research has shown that DNNs are surprisingly robust and do not require all 32 bits. Instead, using quantization, networks can run on …

Development Of The Payload System And Obc Microcontroller Coding For A Cubic Satellite Performing An Additive Self-Repair Experiment In Space, Eduardo Macias-Zugasti

Open Access Theses & Dissertations

Additive manufacturing, which is also known as three-dimensional printing, in space is one of the most promising technologies advancing current capabilities for in-orbit space manufacturing and assembly. Additive manufacturing contributes to the reduction of cost per kilogram and number of launches, thus facilitating extraterrestrial colonization and deep-space exploration. The state of the art includes advancing efforts inside the International Space Station (ISS). However, the ISS is a controlled environment and, to the best of our knowledge, no spacecraft or satellite has performed additive manufacturing tasks in the extreme environment of outer space. In this work a 1U CubeSat named Orbital …

Tram System Automation For Environmental Spectroscopy And Vegetation Monitoring, Enrique Anguiano Chavez

Open Access Theses & Dissertations

Spectroscopy is the science of studying the interactions of matter and electromagnetic radiation (EMR). In particular, field spectroscopy takes place in a natural environment with a natural source of EMR. The paper presents progress towards the development and automation of a tram cart system. The new system in development collects high resolution, hyperspectral images and data from a spectrometer. Alternatives for a sensor cover mechanism to provide cover for the sensors mounted while the system is not operating are discussed, analyzing and comparing the benefits and disadvantages. An implementation for a charging station in an environment isolated from the electric …

Artificial Intelligence In The Assessment Of Transmission And Distribution Systems Under Natural Disasters Using Machine Learning And Deep Learning Techniques In A Knowledge Discovery Framework, Rossana Villegas

Open Access Theses & Dissertations

Warming trends and increasing temperatures have been observed and reported by federal agencies, such as the National Oceanic and Atmospheric Administration (NOAA). Extreme-weather events, especially hurricanes, tornadoes and winter storms, are among the highly devastating natural disasters responsible for massive and prolonged power outages in Electrical Transmission and Distribution Systems (ETDS). Moreover, the failure rate probability of any system component under extreme-weather tends to increase in the impacted geographic area. This Dissertation proposes an Artificial Intelligence (AI) Decision Support System that can predict damage in the ETDS and allow operators to mitigate disastrous extreme weather events. The document reports the …

Improving Time-Of-Flight And Other Depth Images: Super-Resolution And Denoising Using Variational Methods, Salvador Canales Andrade

Open Access Theses & Dissertations

Depth information is a new important source of perception for machines, which allow them to have a better representation of the surroundings. The depth information provides a more precise map of the location of every object and surfaces in a space of interest in comparison with conventional cameras. Time of flight (ToF) cameras provide one of the techniques to acquire depth maps, however they produce low spatial resolution and noisy maps. This research proposes a framework to enhance and up-scale depth maps by using two different regularization terms: Total Generalized Variation (TGV) and Total Generalized Variation with a Structure Tensor …

Structural And Electrical Characterization Of Tin Oxide Resistive Switching, Arka Talukdar

Open Access Theses & Dissertations

Resistive switching in metal oxide is a phenomenon in which the metal oxide changes its resistance upon application of electric field and thus giving two states; high resistance state (HRS) and low resistance state (LRS). Many metal oxides have been investigated however very little is known about unipolar resistive switching in SnO2 though it has shown excellent resistive switching characteristics. Defects in the material play a vital role in resistive switching of the metal oxides. In this work, the role of defects in resistive switching of SnO2 are investigated in Ti/SnO2/Au structures. Two methods were used to control the concentration …

Safety Airway For Small Unmanned Aerial Vehicles Using A Gas Particles Behavior Analogy, Pablo Rangel

Open Access Theses & Dissertations

The United States Federal Aviation Administration (FAA) implemented the Part 107 legislation to allow the flight of Unmanned Aerial Vehicles (UAV) for commercial use (i.e. package deliveries, power transmission line inspections, etc.) in the National Airspace System (NAS). As a consequence of the newly introduced rules, there is an increased risk for accidents involving injured bystanders or damaged to property. The work within this document defines a UAV to UAV safety distance model that acts as a range sensor enabled "elastic bubble". The length of the UAV safety bubble contracts and expands upon changing airway wind speed conditions. It also …

G-Code Generation For Multi-Process 3d Printing, Callum Peter Bailey

Open Access Theses & Dissertations

Since the inception of stereolithography in the 1980s, interest in 3D printing has exploded, with desktop 3D printers now commercially accessible to the general public. In recent years, next-generation multifunctional technologies have been developed, which combine 3D printing with other technologies such as wire embedding, foil embedding, CNC machining, and robotic component placement, enabling complex parts to be made on a single multifunctional machine.

However, the complexity of these integrated processes exceeds the capabilities of established design tools. To this end, this Thesis aims to develop a multi-functional design solution that can automatically generate final control code for next-generation multifunctional …

Adaptive Switched Capacitor Voltage Boost For Thermoelectric Generation, Rene A. Brito

Open Access Theses & Dissertations

Thermoelectric generators (TEG) and other forms of energy harvesting often provide voltages that are not directly usable by traditional electronics as levels are too low from the TEG. While increasing the number of thermoelectric elements can ultimately increase the power output, there is a tradeoff between size and power. By implementing charge pumps, a proposed circuit technique is described that can boost the TEG output to levels that can be used for energy harvesting applications. Current voltage boost circuits for TEGs simply boost a voltage by a set amount. The proposed circuit consists of an analog chip, to provide several …

Low Power Design Techniques For Data Acquisition, Praveen Kumar Palakurthi

Open Access Theses & Dissertations

Semiconductor technology advancement continues to lead to smaller device geometries. Digital circuits have benefited from the technology scaling whereas analog circuits often suffer in terms of performance and noise immunity. As functionality continues to be integrated onto systems-on-chips, analog circuits consume increasingly more power than digital circuits and the objective of this Dissertation is to explore low power techniques in the context of both digital and analog circuits residing on the same silicon real estate.

As natural signals are in analog form, a necessity exists to convert these into digital form and then back again in order to exploit the …

Computation Offloading Decisions For Reducing Completion Time, Salvador Melendez

Open Access Theses & Dissertations

Mobile devices are being widely used in many applications such as image processing, computer vision (e.g. face detection and recognition), wearable computing, language translation, and battlefield operations. However, mobile devices are constrained in terms of their battery life, processor performance, storage capacity, and network bandwidth. To overcome these issues, there is an approach called Computation Offloading, also known as cyber-foraging and surrogate computing. Computation offloading consists of migrating computational jobs from a mobile device to more powerful remote computing resources. Upon completion of the job, the results are sent back to the mobile device. However, a decision must be made; …

Frequency Selective Surfaces For Extreme Applications, Jay Houston Barton

Open Access Theses & Dissertations

It is known that for high-power microwaves and other extreme environments, the use of resonant metallic elements in frequency selective surfaces can be problematic. The solution developed within this Dissertation to solve these problems was to use guided-mode resonance phenomenon to create all-dielectric frequency selective surfaces that could survive these extreme environments.

To fully understand how these devices work, three different computational electromagnetic methods are formulated and implemented. The formulation of these methods start with the differential form of Maxwellâ??s equation and are derived all the way down to the final simulation state. This is done sequentially and all the …

Efficient, Scalable, Parallel, Matrix-Matrix Multiplication, Enrique Portillo

Open Access Theses & Dissertations

For the past decade, power/energy consumption has become a limiting factor for large-scale and embedded High Performance Computing (HPC) systems. This is especially true for systems that include accelerators, e.g., high-end computing devices, such as Graphics Processing Units (GPUs), with terascale computing capabilities and high power draws that greatly surpass that of multi-core CPUs. Accordingly, improving the node-level power/energy efficiency of an application can have a direct and positive impact on both classes of HPC systems.

The research reported in this thesis explores the use of software techniques to enhance the execution-time and power-consumption performance of applications executed on a …

Asynchronous Logic Design With Increased Fault Tolerance And Optimized For Subthreshold Operation, Ivan Santos

Open Access Theses & Dissertations

Synchronization has always been an essential feature in electronic circuits, in which functionality is achieved by an appropriate flow of data with respect to time. An ideal concept of logic design would require every operation to happen instantly; however, the application of this principle to electronics is governed by the laws of physics, which influence the time to execute a process. In fact, the speed at which a microprocessor can perform logic operations is determined by the time that charging and discharging capacitive loads requires, and the time that electric signals spend when traveling through the circuitry. Consequently, the variation …

Anasic Design And Test Methodology For An Undergraduate Design And Fabrication Project, Arun Joseph Kurian

Open Access Theses & Dissertations

During the 1990's the main focus of chip design methodologies was on the timings and area constraints. Power consumption was considered significant only after a drastic increase of device densities from 130nm on as well as dramatic increases in sub threshold leakage. As technology advanced from 130nm to 90nm and below there was a significant increase in leakage current due to lower threshold voltage and the influence of the deep submicron effects. High power consumption causes different problems such as increasing the cost of the product, reducing the reliability, reducing the battery life among others. Therefore EDA tools were designed …

Multi-Physics Modeling Of Induction-Based Additive Manufacturing Of Metals, Danny W. Muse

Open Access Theses & Dissertations

The shift towards 3D printing of functional products has provided a glimpse of the future in terms of manufacturing technologies and the creation of volumetrically complex structures made possible only through these technologies. 3D printing has emerged from the culmination of several technologies covering multiple industries and a myriad of materials. Unlike subtractive technologies such as mills, lathes and CNC machines, where material is removed from a bulk slab or billet to produce a desired part, 3D printing uses an additive layered approach to deposit material based on cross sections of the desired part. Both approaches produce parts based on …

Investigation Of The Divcon Neuron To Increase The Performance Of A Traditional Feed Forward Multi-Layer Perceptron And Its Hardware Implementation, Jovan Saenz

Open Access Theses & Dissertations

ABSTRACT

Artificial Neural Networks (ANNs) have been developed in an attempt to emulate the information processing capabilities of the biological brain. They offer an alternate computing approach to problems in which mathematical modeling is complicated, such as pattern recognition and pattern classification.

Since ANNs were proposed in the early 1940s, there has been a great amount of research effort dedicated to the development of new models that improve performance. Consequently, different architectures, a variety of activation functions, and distinct learning algorithms have been developed and implemented in different disciplines such as medicine, engineering, and science. In addition, ANNs have been …

The Application Of Fuzzy Granular Computing For The Analysis Of Human Dynamic Behavior In 3d Space, Murad Mohammad Alaqtash

Open Access Theses & Dissertations

Human dynamic behavior in space is very complex in that it involves many physical, perceptual and motor aspects. It is tied together at a sensory level by linkages between vestibular, visual and somatosensory information that develop through experience of inertial and gravitational reaction forces. Coordinated movement emerges from the interplay among descending output from the central nervous system, sensory input from the body and environment, muscle dynamics, and the emergent dynamics of the whole neuromusculoskeletal system.

There have been many attempts to directly capture the activities of the neuronal system in human locomotion without the ability to clarify how the …

Video Mosaicing Using Inertial Sensors, Arturo Maya Pereyra

Open Access Theses & Dissertations

A method to reduce the process complexity of Video Mosaicing could rely on information concerning the orientation of the camera instead of image feature tracking. Inertial sensors (e.g. accelerometers and gyros) could form part of a low cost motion based technology which can be used in this case and this thesis details an approach for Video Mosaicing based on such technology that requires non-computationally intensive algorithms.

2d Gaussian Object Motion Detection, Miguel Angel Chaidez

Open Access Theses & Dissertations

Dr. John Moya, and associated research assistants, have previously created an image-change recognition algorithm (JESSE) to mark changes within an image. The focus of this thesis is to present a physical application and modification of this algorithm in order to detect a surgeon's hand and verify chip placement on a printed circuit board.

There are different techniques in implementing visual recognition and motion detection with smart systems but the high cost and complicated calibration of these systems make them impractical. The goal was to create a system that is simple, inexpensive and applicable to multiple applications that will allow the …

Development Of Load Balancing Algorithm Based On Analysis Of Multi-Core Architecture On Beowulf Cluster, Damian Valles

Open Access Theses & Dissertations

In this work, analysis, and modeling were employed to improve the Linux Scheduler for HPC use. The performance throughput of a single compute-node of the 23 node Beowulf cluster, Virgo 2.0, was analyzed to find bottlenecks and limitations that affected performance in the processing hardware where each compute-node consisted of two quad-core processors with eight gigabytes of memory. The analysis was performed using the High Performance Linpack (HPL) benchmark.

In addition, the processing hardware of the compute-node was modeled using an Instruction per Cycle (IPC) metric that was estimated using linear regression. Modeling data was obtained by using the Tuning …

Utepcam: A Scalable Wireless Vision Sensor Architecture For Computational, Power And Bandwidth Constrained Scenarios, Ricardo Zepeda

Open Access Theses & Dissertations

UTEPcam is a low cost and power vision sensor node system. UTEPcam is composed of an Atmel atmega32 8-bit microcontroller, a CY7C09099V static RAM chip, an OV6620 CMOS image sensor, a XBEE transceiver and a SD Flash memory card, and four logic gates. UTEPcam's simple yet efficiently architecture enables it to capture video at one frame per second. At its absolute highest, it is estimated that UTEPcam consumes only 1.321 Amps. When in standby, UTEPcam consumes 21 microamps. Furthermore, UTEPcam's program takes up only 1Kbyte of memory space.

UTEPcam's CPU, a simple 8-bit MCU, is unlike most vision sensor node …

Collaborative And Distributed Algorithms For Localization In Wireless Sensor Networks Based On The Solution Of Spatially Constrained Local And Sub-Local Problems, Juan De Dios Cota

Open Access Theses & Dissertations

In this research we present algorithms for the distributed and collaborative localization of nodes for applications in wireless sensor networks. The algorithms are distributed in the sense that each node can estimate its own position using only range information and position estimates from neighboring nodes. The algorithms aim at achieving good accuracy with low computational complexity and low energy consumption. We consider the full localization process consisting of an initialization stage followed by a refinement stage.

For initialization, we propose a \emph{bilateration} algorithm where each node uses a set of anchors and their respective ranges to solve a set of …

Algorithms For Training Large-Scale Linear Programming Support Vector Regression And Classification, Pablo Rivas Perea

Open Access Theses & Dissertations

The main contribution of this dissertation is the development of a method to train a Support Vector Regression (SVR) model for the large-scale case where the number of training samples supersedes the computational resources. The proposed scheme consists of posing the SVR problem entirely as a Linear Programming (LP) problem and on the development of a sequential optimization method based on variables decomposition, constraints decomposition, and the use of primal-dual interior point methods. Experimental results demonstrate that the proposed approach has comparable performance with other SV-based classifiers. Particularly, experiments demonstrate that as the problem size increases, the sparser the solution …

Design Of A Low Voltage Analog To Digital Converter, Praveen Kumar Palakurthi

Open Access Theses & Dissertations

For the past 40 years, the number of transistors per chip has been increasing at an exponential rate confirming Moore's Law regarding the growth of chip complexity. Increases in transistor count have led to enhancements in functionality and made possible products with features that were unimaginable ten years ago. Consequently, the resulting increase in digitization in all electronics, across a wide range of applications, requires the Analog-to-Digital Converters (ADCs) with a higher resolution and lower power consumption. The evolution of the integrated circuit technologies and scaling methodologies partially helps in providing faster circuits and allowing complex functionalities in a given …

Health Prognisis Of Electronics Via Power Profiling, Jonathan Amilcar Cervantes

Open Access Theses & Dissertations

The objective of this research is to investigate a new approach for the early detection of latent defects in electronic devices in the field. Reliability is assessed through the non-traditional approach of recording and evaluating the power profile of electronic devices within a deterministic state of operation. Traditionally, measuring the quiescent current (Iddq) of a device has been employed in manufacturing tests to detect defective parts prior to deployment to the field. However, the monitoring of the deterministic power signature (i.e. boot up or during a self-test routine) has never been exploited to monitor the health of a device in …