Engineering Commons^™

Neural Networks For Improved Signal Source Enumeration And Localization With Unsteered Antenna Arrays, John T. Rogers Ii

Theses and Dissertations

Direction of Arrival estimation using unsteered antenna arrays, unlike mechanically scanned or phased arrays, requires complex algorithms which perform poorly with small aperture arrays or without a large number of observations, or snapshots. In general, these algorithms compute a sample covriance matrix to obtain the direction of arrival and some require a prior estimate of the number of signal sources. Herein, artificial neural network architectures are proposed which demonstrate improved estimation of the number of signal sources, the true signal covariance matrix, and the direction of arrival. The proposed number of source estimation network demonstrates robust performance in the case …

Asset Cueing Nuclear Radiation Anomaly Detection Using An Embedded Neural Network Resource, April Inamura

Department of Electrical and Computer Engineering: Dissertations, Theses, and Student Research

Nuclear radiation detection is inherently a challenging task, coupled with a high background variation or increase in anomalies, the accuracy for detection can plummet. A key factor in the success of nuclear detection hinges on the sensor’s ability to generalize its model and directly leads to the model’s robustness. The goal of this project is to develop algorithms suitable for use on the University of Nebraska-Lincoln’s Pingora chip, a low-power, system-on-chip device with an active neural processing unit (NPU) made for nuclear radiation detection. The thesis aims to improve Pingora’s overall generalization ability in nuclear radiation source detection. A multiphase …

Long-Term Human Video Activity Quantification In Collaborative Learning Environments, Venkatesh Jatla

Electrical and Computer Engineering ETDs

Research on video activity detection has mainly focused on identifying well-defined human activities in short video segments, often requiring large-parameter systems and extensive training datasets. This dissertation introduces a low-parameter, modular system with rapid inference capabilities, capable of being trained on limited datasets without transfer learning from large-parameter systems. The system accurately detects specific activities and associates them with students in real-life classroom videos. Additionally, an interactive web-based application is developed to visualize human activity maps over long classroom videos.

Long-term video activity detection in classrooms presents challenges, such as multiple simultaneous activities, rapid transitions, long-term occlusions, duration exceeding 15 …

Hybrid Model For Making Decision Methods In Wireless Sensor Networks Through Neuro-Fuzzy Inference System, Martha Lucia Torres

Open Access Theses & Dissertations

Considering the complexity and multiple alternatives for technology decisions in Wireless Sensor Networks (WSNs), a multicriteria selection method (MCDM) is an appropriate approach for choosing the best option in technical projects. Purely quantitative decision-making procedures have currently been created based on client requirements and recommendations from industry professionals in many domains. In this context, implementation and operational costs could be increasing due to technical problems and additional processes. In order to prevent future difficulties and obtain a more accurate technology selection, a new method was being developed to involve qualitative and quantitative parameters taken from real scenarios and technical literature …

Computational Models To Detect Radiation In Urban Environments: An Application Of Signal Processing Techniques And Neural Networks To Radiation Data Analysis, Jose Nicolas Gachancipa

Beyond: Undergraduate Research Journal

Radioactive sources, such as uranium-235, are nuclides that emit ionizing radiation, and which can be used to build nuclear weapons. In public areas, the presence of a radioactive nuclide can present a risk to the population, and therefore, it is imperative that threats are identified by radiological search and response teams in a timely and effective manner. In urban environments, such as densely populated cities, radioactive sources may be more difficult to detect, since background radiation produced by surrounding objects and structures (e.g., buildings, cars) can hinder the effective detection of unnatural radioactive material. This article presents a computational model …

Building Marginal Pattern Library With Unbiased Training Dataset For Enhancing Model-Free Load-Ed Mapping, Qiwei Zhang, Fangxing Li, Wei Feng, Xiaofei Wang, Linquan Bai, Rui Bo

Electrical and Computer Engineering Faculty Research & Creative Works

Input-output mapping for a given power system problem, such as loads versus economic dispatch (ED) results, has been demonstrated to be learnable through artificial intelligence (AI) techniques, including neural networks. However, the process of identifying and constructing a comprehensive dataset for the training of such input-output mapping remains a challenge to be solved. Conventionally, load samples are generated by a pre-defined distribution, and then ED is solved based on those load samples to form training datasets, but this paper demonstrates that such dataset generation is biased regarding load-ED mapping. The marginal unit and line congestion (i.e., marginal pattern) exhibit a …

Spanish And English Phoneme Recognition By Training On Simulated Classroom Audio Recordings Of Collaborative Learning Environments, Mario J. Esparza Perez

Electrical and Computer Engineering ETDs

Audio recordings of collaborative learning environments contain a constant presence of cross-talk and background noise. Dynamic speech recognition between Spanish and English is required in these environments. To eliminate the standard requirement of large-scale ground truth, the thesis develops a simulated dataset by transforming audio transcriptions into phonemes and using 3D speaker geometry and data augmentation to generate an acoustic simulation of Spanish and English speech. The thesis develops a low-complexity neural network for recognizing Spanish and English phonemes (available at github.com/muelitas/keywordRec). When trained on 41 English phonemes, 0.099 PER is achieved on Speech Commands. When trained on 36 Spanish …

Analog Spiking Neural Network Implementing Spike Timing-Dependent Plasticity On 65 Nm Cmos, Luke Vincent

Graduate Theses and Dissertations

Machine learning is a rapidly accelerating tool and technology used for countless applications in the modern world. There are many digital algorithms to deploy a machine learning program, but the most advanced and well-known algorithm is the artificial neural network (ANN). While ANNs demonstrate impressive reinforcement learning behaviors, they require large power consumption to operate. Therefore, an analog spiking neural network (SNN) implementing spike timing-dependent plasticity is proposed, developed, and tested to demonstrate equivalent learning abilities with fractional power consumption compared to its digital adversary.

Utilizing Trajectory Optimization In The Training Of Neural Network Controllers, Nicholas Kimball

Master's Theses

Applying reinforcement learning to control systems enables the use of machine learning to develop elegant and efficient control laws. Coupled with the representational power of neural networks, reinforcement learning algorithms can learn complex policies that can be difficult to emulate using traditional control system design approaches. In this thesis, three different model-free reinforcement learning algorithms, including Monte Carlo Control, REINFORCE with baseline, and Guided Policy Search are compared in simulated, continuous action-space environments. The results show that the Guided Policy Search algorithm is able to learn a desired control policy much faster than the other algorithms. In the inverted pendulum …

Synchrophasor-Based Fault Location Detection And Classification, In Power Systems, Using Artificial Intelligence, Hemal Falak

Graduate Theses and Dissertations

With the introduction of sophisticated electronic gadgets which cannot sustain interruption in the provision of electricity, the need to supply uninterrupted and reliable power supply, to the consumers, has become a crucial factor in the present-day world. Therefore, it is customary to correctly identify fault locations in an electrical power network, in order to rectify faults and restore power supply in the minimum possible time. Many automated fault location detection algorithms have been proposed, however, prior art requires topological and physical information of the electrical power network. This thesis presents a new method of detecting fault locations, in transmission as …

Artificial Intelligence In The Context Of Human Consciousness, Hannah Defries

Senior Honors Theses

Artificial intelligence (AI) can be defined as the ability of a machine to learn and make decisions based on acquired information. AI’s development has incited rampant public speculation regarding the singularity theory: a futuristic phase in which intelligent machines are capable of creating increasingly intelligent systems. Its implications, combined with the close relationship between humanity and their machines, make achieving understanding both natural and artificial intelligence imperative. Researchers are continuing to discover natural processes responsible for essential human skills like decision-making, understanding language, and performing multiple processes simultaneously. Artificial intelligence attempts to simulate these functions through techniques like artificial neural …

Solar Concentrators Manufacture And Automation, Ernst Kussul, Tetyana Baydyk, Alberto Escalante Estrada, Maria Tersa Rodriguez Gonzalez, Donald C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

Solar energy is one of the most promising types of renewable energy. Flat facet solar concentrators were proposed to decrease the cost of materials needed for production. They used small flat mirrors for approximation of parabolic dish surface. The first prototype of flat facet solar concentrators was made in Australia in 1982. Later various prototypes of flat facet solar concentrators were proposed. It was shown that the cost of materials for these prototypes is much lower than the material cost of conventional parabolic dish solar concentrators. To obtain the overall low cost of flat facet concentrators it is necessary to …

The Hilbert-Huang Transform: A Theoretical Framework And Applications To Leak Identification In Pressurized Space Modules, Kenneth R. Bundy

Electronic Theses and Dissertations

Any manned space mission must provide breathable air to its crew. For this reason, air leaks in spacecraft pose a danger to the mission and any astronauts on board. The purpose of this work is twofold: the first is to address the issue of air pressure loss from leaks in spacecraft. Air leaks present a danger to spacecraft crew, and so a method of finding air leaks when they occur is needed. Most leak detection systems localize the leak in some way. Instead, we address the identification of air leaks in a pressurized space module, we aim to determine the …

Deep Neural Network Architectures For Modulation Classification Using Principal Component Analysis, Sharan Ramjee, Shengtai Ju, Diyu Yang, Aly El Gamal

The Summer Undergraduate Research Fellowship (SURF) Symposium

In this work, we investigate the application of Principal Component Analysis to the task of wireless signal modulation recognition using deep neural network architectures. Sampling signals at the Nyquist rate, which is often very high, requires a large amount of energy and space to collect and store the samples. Moreover, the time taken to train neural networks for the task of modulation classification is large due to the large number of samples. These problems can be drastically reduced using Principal Component Analysis, which is a technique that allows us to reduce the dimensionality or number of features of the samples …

Iris Biometric Identification Using Artificial Neural Networks, Kevin Joseph Haskett

Master's Theses

A biometric method is a more secure way of personal identification than passwords. This thesis examines the iris as a personal identifier with the use of neural networks as the classifier. A comparison of different feature extraction methods that include the Fourier transform, discrete cosine transform, the eigen analysis method, and the wavelet transform, is performed. The robustness of each method, with respect to distortion and noise, is also studied.

Lionfish Detection System, Carmelo Furlan, Andrew Boniface

Computer Engineering

Deep neural networks have proven to be an effective method in classification of images. The ability to recognize objects has opened the door for many new systems which use image classification to solve challenging problems where conventional image classification would be inadequate. We trained a large, deep convolutional neural network to identify lionfish from other species that might be found in the same habitats. Google’s Inception framework served as a powerful platform for our fish recognition system. By using transfer learning, we were able to obtain exceptional results for the classification of different species of fish. The convolutional neural network …

Association Learning Via Deep Neural Networks, Trevor J. Landeen

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Deep learning has been making headlines in recent years and is often portrayed as an emerging technology on a meteoric rise towards fully sentient artificial intelligence. In reality, deep learning is the most recent renaissance of a 70 year old technology and is far from possessing true intelligence. The renewed interest is motivated by recent successes in challenging problems, the accessibility made possible by hardware developments, and dataset availability.

The predecessor to deep learning, commonly known as the artificial neural network, is a computational network setup to mimic the biological neural structure found in brains. However, unlike human brains, artificial …

Training Neural Networks To Pilot Autonomous Vehicles: Scaled Self-Driving Car, Jason Zisheng Chang

Senior Projects Spring 2018

This project explores the use of deep convolutional neural networks in autonomous cars. Successful implementation of autonomous vehicles has many societal benefits. One of the main benefits is its potential to significantly reduce traffic accidents. In the United States, the National Highway Traffic Safety Administration states that human error is at fault for 93% of automotive crashes. Robust driverless vehicles can prevent many of these collisions. The main challenge in developing autonomous vehicles today is how to create a system that is able to accurately perceive and process the world around it. In 2016, NVIDIA successfully trained a deep convolutional …

Switching Neural Network Systems For Nonlinear Tracking, Manoj Ghimire

Browse all Theses and Dissertations

In this thesis, we consider the problem of tracking in complex nonlinear dynamical systems. While the Kalman filter is known to be the mean-squared error optimal tracker under linear dynamics and linear measurements, more sophisticated models and algorithms are required for complex dynamics. Here, we consider switching systems where the dynamical properties vary (''switch modes") over time. For example, the dynamics of a vehicle may switch as it transitions from interstate to urban conditions, human speech dynamics switch as speakers change, and stock market dynamics switch with discrete news events. In this work, we use mode-dependent neural networks to capture …

Modeling And Simulation Of Microgrid, Ahmad Alzahrani, Mehdi Ferdowsi, Pourya Shamsi, Cihan H. Dagli

Electrical and Computer Engineering Faculty Research & Creative Works

Complex computer systems and electric power grids share many properties of how they behave and how they are structured. A microgrid is a smaller electric grid that contains several homes, energy storage units, and distributed generators. The main idea behind microgrids is the ability to work even if the main grid is not supplying power. That is, the energy storage unit and distributed generation will supply power in that case, and if there is excess in power production from renewable energy sources, it will go to the energy storage unit. Therefore, the electric grid becomes decentralized in terms of control …

Solar Irradiance Forecasting Using Deep Neural Networks, Ahmad Alzahrani, Pourya Shamsi, Cihan H. Dagli, Mehdi Ferdowsi

Electrical and Computer Engineering Faculty Research & Creative Works

Predicting solar irradiance has been an important topic in renewable energy generation. Prediction improves the planning and operation of photovoltaic systems and yields many economic advantages for electric utilities. The irradiance can be predicted using statistical methods such as artificial neural networks (ANN), support vector machines (SVM), or autoregressive moving average (ARMA). However, they either lack accuracy because they cannot capture long-term dependency or cannot be used with big data because of the scalability. This paper presents a method to predict the solar irradiance using deep neural networks. Deep recurrent neural networks (DRNNs) add complexity to the model without specifying …

Neural Networks Skin Tumor Diagnostic System, Zhao Zhang, William V. Stoecker, Randy Hays Moss

Zhao Zhang

In this study, a malignant melanoma diagnostic system is designed using a straightforward neural network with the back-propagation learning algorithm. Eleven features are automatically extracted from skin tumor images. The correct diagnostic rate of this system is better than the average rate of 16 dermatologists who based their diagnosis with only the slide images.

Dc-Dc Converter Control System For The Energy Harvesting From Exercise Machines System, Alexander Sireci

Master's Theses

Current exercise machines create resistance to motion and dissipate energy as heat. Some companies create ways to harness this energy, but not cost-effectively. The Energy Harvesting from Exercise Machines (EHFEM) project reduces the cost of harnessing the renewable energy. The system architecture includes the elliptical exercise machines outputting power to DC-DC converters, which then connects to the microinverters. All microinverter outputs tie together and then connect to the grid. The control system, placed around the DC-DC converters, quickly detects changes in current, and limits the current to prevent the DC-DC converters and microinverters from entering failure states.

An artificial neural …

Neuron Clustering For Mitigating Catastrophic Forgetting In Supervised And Reinforcement Learning, Benjamin Frederick Goodrich

Doctoral Dissertations

Neural networks have had many great successes in recent years, particularly with the advent of deep learning and many novel training techniques. One issue that has affected neural networks and prevented them from performing well in more realistic online environments is that of catastrophic forgetting. Catastrophic forgetting affects supervised learning systems when input samples are temporally correlated or are non-stationary. However, most real-world problems are non-stationary in nature, resulting in prolonged periods of time separating inputs drawn from different regions of the input space.

Reinforcement learning represents a worst-case scenario when it comes to precipitating catastrophic forgetting in neural networks. …

Implementation Of A Neuromorphic Development Platform With Danna, Jason Yen-Shen Chan

Masters Theses

Neuromorphic computing is the use of artificial neural networks to solve complex problems. The specialized computing field has been growing in interest during the past few years. Specialized hardware that function as neural networks can be utilized to solve specific problems unsuited for traditional computing architectures such as pattern classification and image recognition. However, these hardware platforms have neural network structures that are static, being limited to only perform a specific application, and cannot be used for other tasks. In this paper, the feasibility of a development platform utilizing a dynamic artificial neural network for researchers is discussed.

Optimisation Of Stand-Alone Hydrogen-Based Renewable Energy Systems Using Intelligent Techniques, Adel Brka

Theses: Doctorates and Masters

Wind and solar irradiance are promising renewable alternatives to fossil fuels due to their availability and topological advantages for local power generation. However, their intermittent and unpredictable nature limits their integration into energy markets. Fortunately, these disadvantages can be partially overcome by using them in combination with energy storage and back-up units. However, the increased complexity of such systems relative to single energy systems makes an optimal sizing method and appropriate Power Management Strategy (PMS) research priorities.

This thesis contributes to the design and integration of stand-alone hybrid renewable energy systems by proposing methodologies to optimise the sizing and operation …

Digital-To-Analog Converter Interface For Computer Assisted Biologically Inspired Systems, Nicholas Conley Poore

Masters Theses

In today's integrated circuit technology, system interfaces play an important role of enabling fast, reliable data communications. A key feature of this work is the exploration and development of ultra-low power data converters. Data converters are present in some form in almost all mixed-signal systems; in particular, digital-to-analog converters present the opportunity for digitally controlled analog signal sources. Such signal sources are used in a variety of applications such as neuromorphic systems and analog signal processing. Multi-dimensional systems, such as biologically inspired neuromorphic systems, require vectors of analog signals. To use a microprocessor to control these analog systems, we must …

Smart Weights, Luke W. Rafla-Yuan, Austin C. Fox

Electrical Engineering

The goal of this project is to design and implement weights which can record and analyze work out patterns. Motivation for this project stems from the high cost of personal training. The hope is that this device will provide many of the benefits a user receives from personal training at only a fraction of the cost. The Smart Weight is designed with an on-board Inertial Measurement Unit providing acceleration, gyroscope, and magnetometer data. A microcontroller records and analyzes changes in motion, feeding this data into Multiplicative Recurrent Neural Network (MRNN) for exercise classification. A Raspberry Pi was chosen as the …

Affine Image Registration Using Artificial Neural Networks, Pramod Gadde

Master's Theses

This thesis deals with image registration of MRI images using neural networks. Image registration combines multiple images of the same subject that were taken at different points in time, from different sensors, or from different points of views into a single image and coordinate system. Image registration is widely used in medical imaging and remote sensing. In this thesis feed forward neural networks and wavelet neural networks are used to estimate the parameters of registration. Simulations show that the wavelet networks provide significantly more accurate results than feed forward networks and other proposed methods including genetic algorithms. Both methods are …

Neural Networks And The Natural Gradient, Michael R. Bastian

All Graduate Theses and Dissertations, Spring 1920 to Summer 2023

Neural network training algorithms have always suffered from the problem of local minima. The advent of natural gradient algorithms promised to overcome this shortcoming by finding better local minima. However, they require additional training parameters and computational overhead. By using a new formulation for the natural gradient, an algorithm is described that uses less memory and processing time than previous algorithms with comparable performance.