Electrical and Computer Engineering Commons^™

Voltage Scaled Low Power Dnn Accelerator Design On Reconfigurable Platform, Rourab Paul, Sreetama Sarkar, Suman Sau, Sanghamitra Roy, Koushik Chakraborty, Amlan Chakrabarti

Electrical and Computer Engineering Faculty Publications

The exponential emergence of Field-Programmable Gate Arrays (FPGAs) has accelerated research on hardware implementation of Deep Neural Networks (DNNs). Among all DNN processors, domain-specific architectures such as Google’s Tensor Processor Unit (TPU) have outperformed conventional GPUs (Graphics Processing Units) and CPUs (Central Processing Units). However, implementing low-power TPUs in reconfigurable hardware remains a challenge in this field. Voltage scaling, a popular approach for energy savings, can be challenging in FPGAs, as it may lead to timing failures if not implemented appropriately. This work presents an ultra-low-power FPGA implementation of a TPU for edge applications. We divide the systolic array of …

Understanding Timing Error Characteristics From Overclocked Systolic Multiply–Accumulate Arrays In Fpgas, Andrew Chamberlin, Andrew Gerber, Mason Palmer, Tim Goodale, Noel Daniel Gundi, Koushik Chakraborty, Sanghamitra Roy

Electrical and Computer Engineering Faculty Publications

Artificial Intelligence (AI) hardware accelerators have seen tremendous developments in recent years due to the rapid growth of AI in multiple fields. Many such accelerators comprise a Systolic Multiply–Accumulate Array (SMA) as its computational brain. In this paper, we investigate the faulty output characterization of an SMA in a real silicon FPGA board. Experiments were run on a single Zybo Z7-20 board to control for process variation at nominal voltage and in small batches to control for temperature. The FPGA is rated up to 800 MHz in the data sheet due to the max frequency of the PLL, but the …

Cost Minimization For Charging Electric Bus Fleets, Daniel Mortensen, Jacob Gunther, Greg Droge, Justin Whitaker

Electrical and Computer Engineering Student Research

Recent attention for reduced carbon emissions has pushed transit authorities to adopt battery electric buses (BEBs). One challenge experienced by BEB users is extended charge times, which create logistical challenges and may force BEBs to charge when energy is more expensive. Furthermore, BEB charging leads to high power demands, which can significantly increase monthly power costs and may push the electrical infrastructure beyond its present capacity, requiring expensive upgrades. This work presents a novel method for minimizing the monthly cost of BEB charging while meeting bus route constraints. This method extends previous work by incorporating a more novel cost model, …

Cost Minimizing Energy Management Control Scheme For Microgrids Considering Dynamic Electricity Prices, Levi T. Miller

All Graduate Theses and Dissertations, Fall 2023 to Present

As countries develop and technology improves, the world is using more energy than ever before. This fact along with several other political, social, and economic factors has resulted in simultaneous energy and climate crises. A partial solution to both problems is bringing clean energy sources of electricity closer to the customers who use that energy. A microgrid is a smaller version of the national electric grid where smaller electricity generators are networked with local consumers and controlled independently of the main grid. Because control of electricity sources and loads are transferred to local controllers, the flexibility with which they can …

Gps-Denied Navigation Using Synthetic Aperture Radar, Tucker Hathaway

All Graduate Theses and Dissertations, Fall 2023 to Present

In most modern navigation systems, GPS is used to determine the precise location of the vehicle; however, GPS signals can easily be blocked, jammed, or spoofed. These signals can be blocked by canyons or tall buildings. Additionally, adversaries can transmit signals that either make GPS signals difficult to interpret or that imitate real GPS signals and cause a navigation system to think it is somewhere other than its true location. GPS-denied (GPS-D) navigation is the process of navigating in the absence of GPS.

Many methods of performing GPS-D navigation have been proposed and explored. One such method is to use …

Virtual Structures Based Autonomous Formation Flying Control For Small Satellites, Tyson K. Smith

All Graduate Theses and Dissertations, Fall 2023 to Present

Many space organizations have a growing need to fly several small satellites close together in order to collect and correlate data from different satellite sensors. To do this requires teams of engineers monitoring the satellites orbits and planning maneuvers for the satellites every time the satellite leaves its desired trajectory or formation. This task of maintaining the satellites orbits quickly becomes an arduous and expensive feat for satellite operations centers. This research develops and analyzes algorithms that allow satellites to autonomously control their orbit and formation without human intervention. This goal is accomplished by developing and evaluating a decentralized, optimization-based …

A Cohesive Simulation And Testing Platform For Civil Autonomous Aerial Sensing And Operations, Stockton G. Slack

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Drones (also known as sUAS or small Uncrewed Aerial Systems) are often flown with cameras to take images of an area of land. These images can then be used to create a map by stitching these images together. This map can then be analyzed using scientific principles to learn things about the land and make decisions or take action based on the information.

The scientific application of drones is very advantageous, but flying a drone is inherently dangerous, impacting the safety of the airspace (particularly in the event of a crash), and drones are more dangerous the bigger they are. …

A Scalable, Cost-Oriented Approach For Computing Charge Plans For Electric Bus Fleets, Daniel T. Mortensen

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Recent attention for reduced carbon emissions has pushed transit authorities to adopt battery electric buses (BEBs). One challenge experienced by BEB users is extended charge times, which create logistical challenges and may force BEBs to charge when energy is more expensive. Furthermore, BEB charging leads to high power demands, which can significantly increase monthly power costs and may push electrical infrastructure beyond its present capacity, requiring expensive upgrades. This work presents a novel method for minimizing the monthly cost of BEB charging while meeting bus route constraints. This method extends previous work by incorporating a more novel cost model, effects …

Development Of The Digital Signal Processing For The Space Weather Probes Version 2 Sensor Using The Matlab/Simulink Environment, Benjamin J. Lewis

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Space Weather Probes (SWP) is an instrument that provides measurements of the plasma environment of the ionosphere. SWP was flown on the Scintillation Prediction Observation Task (SPORT) mission, a joint mission between the United States of America and Brazil. This thesis will develop the digital signal processing (DSP) hardware design for the Space Weather Probes version 2 (SWP2). The data from these instruments will be used to determine the density and temperature of the local plasma, as well as the electric field in the local plasma. This thesis presents the design and testing of the DSP designs for all of …

Developing Firmware For Space Weather Probes 2 Using Hdl Coder, Nicholas L. Wallace

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

GPS and wireless communications are affected by interference from the ionosphere. Space weather affects plasma in the ionosphere, causing communication disruptions and reliability issues. To better understand how space weather affects the ionosphere, instruments are flown in space to collect data about the electrical characteristics of plasma in the ionosphere. Space systems require a lot of time and effort to develop and test. This thesis explores how a high level tool can be used to simplify the process and some obstacles that still exist with developing some space systems. To do this, the firmware architecture of a new version of …

Super-Resolution Textured Digital Surface Map (Dsm) Formation By Selecting The Texture From Multiple Perspective Texel Images Taken By A Low-Cost Small Unmanned Aerial Vehicle (Uav), Bikalpa Khatiwada

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Textured Digital Surface Model (TDSM) is a three-dimensional terrain map with texture overlaid on it. Utah State University has developed a texel camera which can capture a 3D image called a texel image. A TDSM can be constructed by combining these multiple texel images, which is much cheaper than the traditional method. The overall goal is to create a TDSM for a larger area that is cheaper and equally accurate as the TDSM created using a high-cost system.

The images obtained from such an inexpensive camera have a lot of errors. To create scientifically accurate TDSM, the error presented in …

Mission Planning Techniques For Cooperative Leo Spacecraft Constellations, Skylar A. Cox

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

This research develops a mission planning approach that allows different systems to cooperate in accomplishing a single mission goal. Using the techniques described allows satellites to cooperate in efficiently maneuvering, or collecting images of Earth and transmitting the collected data to users on the ground. The individual resources onboard each satellite, like fuel, memory capacity and pointing agility, are used in a manner that ensures the goals and objectives of the mission are realized in a feasible way. A mission plan can be generated for each satellite within the cooperating group that collectively optimize the mission objectives from a global …

Reclaiming Fault Resilience And Energy Efficiency With Enhanced Performance In Low Power Architectures, Noel Daniel Gundi

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Rapid developments of the AI domain has revolutionized the computing industry by the introduction of state-of-art AI architectures. This growth is also accompanied by a massive increase in the power consumption. Near-Theshold Computing (NTC) has emerged as a viable solution by offering significant savings in power consumption paving the way for an energy efficient design paradigm. However, these benefits are accompanied by a deterioration in performance due to the severe process variation and slower transistor switching at Near-Threshold operation. These problems severely restrict the usage of Near-Threshold operation in commercial applications. In this work, a novel AI architecture, Tensor Processing …

Producing Analytical Models Of The Sport Spacecraft For Science Data Processing, Jason L. Powell

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

The Scintillation Prediction Observation Research Task (SPORT) is a joint United States of America and Brazil 6U CubeSat mission. The SPORT mission is to understand how the Earth’s ionosphere evolves during day and night transitions. Space Weather Probes, a suite of instruments developed by Utah State University, monitors the ionosphere by observing currents and voltages around the SPORT spacecraft body. This thesis presents the data collected by Space Weather Probes and how it is used to measure the ionosphere. Modeling techniques are also presented to demonstrate how the Space Weather Probes instrument is expected to respond to the ionosphere.

Viability And Performance Of Rf Source Localization Using Autocorrelation-Based Fingerprinting, Joseph L. Ipson

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Finding the source location of a radio-frequency (RF) transmission is a useful capability for many civilian, industrial, and military applications. This problem is particularly challenging when done “Blind,” or when the transmitter was not designed with finding its location in mind, and relatively little information is available about the signal before-hand. Typical methods for this operation utilize the time, phase, power, and frequency viewable from received signals. These features are all less predictable in indoor and urban environments, where signals undergo transformation from multiple interactions with the environment. These interactions imprint structure onto the received signal which is dependent on …

Dynamic Inductive Power Transfer Systems With Reflexive Tuning Networks Designed By Machine Learning, Shuntaro Inoue

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

This dissertation proposes a new way to make dynamic wireless charging systems more affordable. Instead of using one inverter for each transmitter coil, as is typically done, the proposed system uses a single inverter that is connected to multiple transmitter coils. This approach is made possible by reflexive tuning, which allows for high currents to be achieved only on the transmitter coil in use. The system was tested with a 50kW prototype, designed using a combination of neural networks and genetic algorithms. The prototype was tested on both automated rail and vehicle systems. The measured dc-dc efficiency with single and …

Communication‐Less Synchronous Rectification For In Motion Wireless Charging, Joshua B. Larsen

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

This thesis puts forward a control scheme to allow for synchronous rectification for dynamic wireless power transfer. The automotive industry is transitioning away from internal combustion engines (ICEs) and towards electric vehicles (EVs). This transition is spurred by the environmental and economic benefits EVs offer over ICEs. However, further improvements can still be made to how electric vehicles operate. One of these improvements is the technology of in motion wireless charging or dynamic wireless power transfer. In motion wireless charging offers the ability to remove existing range anxiety concerns for EVs. It also offers the potential for a reduction in …

The Active Cryocubesat Technology: Active Thermal Control For Small Satellites, Lucas S. Anderson

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Modern CubeSats and Small Satellites have advanced in capability to tackle science and technology missions that would usually be reserved for more traditional, large satellites. However, this rapid growth in capability is only possible through the fast-to-production, low-cost, and advanced technology approach used by modern small satellite engineers. Advanced technologies in power generation, energy storage, and high-power density electronics have naturally led to a thermal bottleneck, where CubeSats and Small Satellites can generate more power than they can easily reject. The Active CryoCubeSat (ACCS) is an advanced active thermal control technology (ATC) for Small Satellites and CubeSats, which hopes to …

Markov Chain Analysis Comparison, Thomas Robert Prouty

Undergraduate Honors Capstone Projects

This Honors Capstone was proposed as a method to develop a greater understanding of modern model-checking tools. To do this the student chose to analyze and compare the results of the tools Prism, Storm, and Stamina. To evaluate the effectiveness of each tool, comparisons were made of the results for running each tool on a simplified communication network.

The simplified communication network model used was a CTMC (Continuous-Time Markov Chain) model that employed correct signal transitions and erroneous bit flipping transitions. This was done in an effort to simulate possible errors and faults that can occur between a provider and …

Compressive Sensing Via Variational Bayesian Inference Under Two Widely Used Priors: Modeling, Comparison And Discussion, Mohammad Shekaramiz, Todd K. Moon

Electrical and Computer Engineering Faculty Publications

Compressive sensing is a sub-Nyquist sampling technique for efficient signal acquisition and reconstruction of sparse or compressible signals. In order to account for the sparsity of the underlying signal of interest, it is common to use sparsifying priors such as Bernoulli-Gaussian-inverse Gamma (BGiG) and Gaussian-inverse Gamma (GiG) priors on the compounds of the signal. With the introduction of variational Bayesian inference, the sparse Bayesian learning (SBL) methods for solving the inverse problem of compressive sensing have received significant interest as the SBL methods become more efficient in terms of execution time. In this paper, we consider the sparse signal recovery …

Path Planning For Aircraft Under Threat Of Detection From Ground-Based Radar With Uncertainty In Aircraft And Radar States, Austin D. Costley

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Mission planners for manned and unmanned aircraft operating within the detection range of ground-based radar systems are often concerned with the probability of detection. Several factors influence the probability of detection, including aircraft position and orientation, radar position, and radar performance parameters. Current path planning algorithms assume that these factors are known with certainty, but in practice, these factors are estimated and have some uncertainty.

This dissertation explores methods to consider the uncertainty in the detection factors for an aircraft path planner. First, the detection model is extended to include uncertainty in the aircraft position and orientation, radar position, and …

Numerical Modeling And Simulating Thermal Performance Of Printed Circuit Boards, Paul R. Ziegenfelder

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

This research investigates how heat transfer theory can be applied to evaluate the thermal performance of printed circuit board (PCB) designs. A model was developed for correlating theory to the physical parameters relating to the construction of the PCB and a computer application was developed to perform the thermal analysis. The application allows the importation of common PCB design files for a straightforward and quick analysis of a PCB design’s thermal performance. The application is open-source to allows for further development and free access from the public.

Event-Based Obstacle Detection With Commercial Lidar, Chaz Cornwall

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Computerized obstacle detection for moving vehicles is becoming more important as vehicle manufacturers make their systems more autonomous and safe. However, obstacle detection must operate quickly in dynamic environments such as driving at highway speeds. A unique obstacle detection system using 3D changes in the environment is proposed. Furthermore, these 3D changes are shown to contain sufficient information for avoiding obstacles. To make the system easy to integrate onto a vehicle, additional processing is implemented to remove unnecessary dependencies. This system provides a method for obstacle detection that breaks away from typical systems to be more efficient.

Modeling And Control Of Battery Management Systems With High-Frequency Ac Link Coupled Multiport Series Resonant Converters For 2nd Life Battery Applications, Brooks Jace Maughan

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

While the use and production of Electric Vehicles becomes more prevalent, it is also important to make this economical and ensure the reduction of a carbon footprint. Second-life batteries can satisfy both problems as batteries can be used in a second-life application for lower power purposes such as supplementing the grid so the infrastructure needed to charge the expanding fleet of Electric Vehicles can be easily supplied. This thesis goes through the process of Active Cell Balancing which will produce equal capacities, or similar batteries, that can be more efficiently used in these and other types of second-life applications. The …

High-Isolation Antenna Technique For Cubesat-Borne, Continuous-Waveform Radar, Logan E. Voigt

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Radar is important in target tracking, imaging, and weather prediction applications. As technology is increasingly miniaturized, there is a push for smaller radar. Research and exploration in outer space also benefit from small, low-power technologies. The NASA Jet Propulsion Laboratory’s RainCube was the first successful CubeSat-borne radar. A CubeSat is a type of small satellite that conforms to specific size and weight standards. Radar technology benefits from additional research on how to further miniaturize radar payloads.

Integrating the transmit and receive antennas on the solar panels removes the need for antenna-deployment mechanisms, preserving space on the CubeSat. This thesis also …

Analysis Of Optically Transparent Antennas Designed From Different Transparent Conductors, Rakib Hasan

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

This master’s thesis work presents an investigative study about the performance of transparent antennas. The antennas have been fabricated using two advanced transparent conductive films (TCF) namely Nano-C Hybrid and Silver-Nanowire (AgNW). Nano-C Hybrid is a new transparent conductive film which combines thin conductors like Carbon Nanotube and Silver Nanowire. The antennas made from Nano-C Hybrid and AgNW are of monopole types and are mounted on a ground plane made out of a printed circuit board and the antennas are excited using an SMA connector through that printed circuit board. Simulation works have been performed to find out the optimum …

Collaborative Research: Expedite Csi Processing With Lightweight Ai In Massive Mimo Communication Systems, Rose Hu

Funded Research Records

No abstract provided.

Efficient And Secure Resource Allocation In Mobile Edge Computing Enabled Wireless Networks, Qun Wang

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

To support emerging applications such as autonomous vehicles and smart homes and to build an intelligent society, the next-generation internet of things (IoT) is calling for up to 50 billion devices connected world wide. Massive devices connection, explosive data circulation, and colossal data processing demand are driving both the industry and academia to explore new solutions.

Uploading this vast amount of data to the cloud center for processing will significantly increase the load on backbone networks and cause relatively long latency to time-sensitive applications. A practical solution is to deploy the computing resource closer to end-users to process the distributed …

Regression Tree Predictive Filter, Jarren Worthen

Undergraduate Honors Capstone Projects

Many algorithms have been developed to predict future samples of a signal. These algorithms, such as the recursive least squares predictive filter, rely on the assumption that the system generating the signal can be modeled as a linear system of equations. These systems perform poorly when used to predict signals generated by non-linear systems. To predict a non-linear signal, non-linear methods must be used. Regression trees are a simple form of machine learning that is non-linear in nature and can predict output based on a set of given input. The goal of this capstone project was to develop an algorithm …

The Surface Conditions Of Spacecraft Panels May Significantly Affect Spacecraft Survivability, Trace Taylor

Research on Capitol Hill

USU junior Trace grew up in Brigham City and studies physics and electrical engineering. The majority of spacecraft failure is caused by electron charging on the outer surfaces of the craft. Additionally, contaminants on the craft can cause a film over surface panels, increasing the problem. Trace is studying how roughness on panels can mitigate this contamination as it affects the charging that can lead to craft failure. This research will help determine what optimal panel materials should be used in future spacecraft construction. Trace started research almost as soon as he came to campus in his freshman year, and …