Engineering Commons^™

Methods For Object Tracking With Machine Vision, Zachary Simon Stamler

Dissertations and Theses

As machine learning and deep learning systems continue to find applications in science and engineering, the problem of providing these systems with high-quality data continues to increase in importance. Many of these systems utilize machine vision as their primary source of information, and in order to maximally leverage their abilities it is important to be able to provide them with high quality, accurate data. Unfortunately, many sets of tracking data extracted from video suffer from the problem of missing frames, which can arise from a multitude of causes depending on the system. These missing frames can result in confusion between …

Clustered Hyperspectral Target Detection, Sean Onufer Stalley

Dissertations and Theses

Aerial target detection is often used to search for relatively small things over large areas of land. Depending on the size and signature of the target, detection can be a very easy or very difficult task. By capturing images with several hundred color channels, hyperspectral sensors provide a new way of looking at this task, both literally and figuratively. Hyperspectral sensors can be used in many aerial target detection tasks such as identifying unhealthy trees in a forest, searching for minerals at a mining site, or finding the sources of chemical leaks at a factory. The high spectral resolution of …

Combining Algorithms For More General Ai, Mark Robert Musil

Undergraduate Research & Mentoring Program

Two decades since the first convolutional neural network was introduced the AI sub-domains of classification, regression and prediction still rely heavily on a few ML architectures despite their flaws of being hungry for data, time, and high-end hardware while still lacking generality. In order to achieve more general intelligence that can perform one-shot learning, create internal representations, and recognize subtle patterns it is necessary to look for new ML system frameworks. Research on the interface between neuroscience and computational statistics/machine learning has suggested that combined algorithms may increase AI robustness in the same way that separate brain regions specialize. In …

Real-Time Object Detection And Tracking On Drones, Tu Le

Undergraduate Research & Mentoring Program

Unmanned aerial vehicles, also known as drones, have been more and more widely used in recent decades because of their mobility. They appear in many applications such as farming, search and rescue, entertainment, military, and so on. Such high demands for drones lead to the need of developments in drone technologies. Next generations of commercial and military drones are expected to be aware of surrounding objects while flying autonomously in different terrains and conditions. One of the biggest challenges to drone automation is the ability to detect and track objects of interest in real-time. While there are many robust machine …

Design Optimization For A Cnc Machine, Alin Resiga

Dissertations and Theses

Minimizing cost and optimization of nonlinear problems are important for industries in order to be competitive. The need of optimization strategies provides significant benefits for companies when providing quotes for products. Accurate and easily attained estimates allow for less waste, tighter tolerances, and better productivity. The Nelder-Mead Simplex method with exterior penalty functions was employed to solve optimum machining parameters. Two case studies were presented for optimizing cost and time for a multiple tools scenario. In this study, the optimum machining parameters for milling operations were investigated. Cutting speed and feed rate are considered as the most impactful design variables …

Generalized Differential Calculus And Applications To Optimization, R. Blake Rector

Dissertations and Theses

This thesis contains contributions in three areas: the theory of generalized calculus, numerical algorithms for operations research, and applications of optimization to problems in modern electric power systems. A geometric approach is used to advance the theory and tools used for studying generalized notions of derivatives for nonsmooth functions. These advances specifically pertain to methods for calculating subdifferentials and to expanding our understanding of a certain notion of derivative of set-valued maps, called the coderivative, in infinite dimensions. A strong understanding of the subdifferential is essential for numerical optimization algorithms, which are developed and applied to nonsmooth problems in operations …

Bayesian Optimization For Refining Object Proposals, Anthony D. Rhodes, Jordan Witte, Melanie Mitchell, Bruno Jedynak

Computer Science Faculty Publications and Presentations

We develop a general-purpose algorithm using a Bayesian optimization framework for the efficient refinement of object proposals. While recent research has achieved substantial progress for object localization and related objectives in computer vision, current state-of-the-art object localization procedures are nevertheless encumbered by inefficiency and inaccuracy. We present a novel, computationally efficient method for refining inaccurate bounding-box proposals for a target object using Bayesian optimization. Offline, image features from a convolutional neural network are used to train a model to predict an object proposal’s offset distance from a target object. Online, this model is used in a Bayesian active search to …

Algorithm For Premature Ventricular Contraction Detection From A Subcutaneous Electrocardiogram Signal, Iris Lynn Shelly

Dissertations and Theses

Cardiac arrhythmias occur when the normal pattern of electrical signals in the heart breaks down. A premature ventricular contraction (PVC) is a common type of arrhythmia that occurs when a heartbeat originates from an ectopic focus within the ventricles rather than from the sinus node in the right atrium. This and other arrhythmias are often diagnosed with the help of an electrocardiogram, or ECG, which records the electrical activity of the heart using electrodes placed on the skin. In an ECG signal, a PVC is characterized by both timing and morphological differences from a normal sinus beat.

An implantable cardiac …

Task And Participant Scheduling Of Trading Platforms In Vehicular Participatory Sensing Networks, Heyuan Shi, Xiaoyu Song, Ming Gu, Jiaguang Sun

Electrical and Computer Engineering Faculty Publications and Presentations

The vehicular participatory sensing network (VPSN) is now becoming more and more prevalent, and additionally has shown its great potential in various applications. A general VPSN consists of many tasks from task, publishers, trading platforms and a crowd of participants. Some literature treats publishers and the trading platform as a whole, which is impractical since they are two independent economic entities with respective purposes. For a trading platform in markets, its purpose is to maximize the profit by selecting tasks and recruiting participants who satisfy the requirements of accepted tasks, rather than to improve the quality of each task. This …

Hydrologic Modeling In Dynamic Catchments: A Data Assimilation Approach, Sahani Darshika Pathiraja, Ashish Sharma, Lucy Marshall, Hamid Moradkhani

Civil and Environmental Engineering Faculty Publications and Presentations

The transferability of conceptual hydrologic models in time is often limited by both their structural deficiencies and adopted parameterizations. Adopting a stationary set of model parameters ignores biases introduced by the data used to derive them, as well as any future changes to catchment conditions. Although time invariance of model parameters is one of the hallmarks of a high quality hydrologic model, very few (if any) models can achieve this due to their inherent limitations. It is therefore proposed to consider parameters as potentially time varying quantities, which can evolve according to signals in hydrologic observations. In this paper, we …

Evaluation Of Data-Path Topologies For Self-Timed Conditional Statements, Navaneeth Prasannakumar Jamadagni

Dissertations and Theses

This research presents a methodology to evaluate data path topologies that implement a conditional statement for an average-case performance that is better than the worst-case performance. A conditional statement executes one of many alternatives depending on how Boolean conditions evaluate to true or false. Alternatives with simple computations take less time to execute. The self-timed designs can exploit the faster executing alternatives and provide an average-case behavior, where the average depends on the frequency of simple and complex computations, and the difference in the completion times of simple and complex computations. The frequency of simple and complex computations depends on …

Fault Localization In Reversible Circuits Is Easier Than For Classical Circuits, Kavitha Ramasamy, Radhika Tagare, Edward Perkins, Marek Perkowski

Electrical and Computer Engineering Faculty Publications and Presentations

There is recently an interest in test generation for reversible circuits, but nothing has been published about fault localization in such circuits. This paper deals with fault localization for binary reversible (permutative) circuits. We concentrate on functional test based fault localization, to detect and locate “stuck-at” faults in a reversible circuit by creating an adaptive tree. A striking property of reversible circuits is that they exhibit “symmetric” adaptive trees. This helps considerably by being able to generate only half of the tree, and the other half is created as the mirror image of the first half. Because each test covers …

Evolutionary Algorithms For Beginners, Marek Perkowski

Electrical and Computer Engineering Faculty Publications and Presentations

This presentation offers a detailed introduction into the concepts and applications of evolutionary ideas in computing, adapted from genetics and evolution.

Design And Evaluation Of A Specialized Computer Architecture For Manipulating Binary Decision Diagrams, Robert K. Hatt

Dissertations and Theses

Binary Decision Diagrams (BDDs) are an extremely important data structure used in many logic design, synthesis and verification applications. Symbolic problem representations make BDDs a feasible data structure for use on many problems that have discrete representations. Efficient implementations of BOD algorithms on general purpose computers has made manipulating large binary decision diagrams possible. Much research has gone into making BOD algorithms more efficient on general purpose computers. Despite amazing increases in performance and capacity of such computers over the last decade, they may not be the best way to solve large, specialized problems. A computer architecture designed specifically to …

Training Strategies For Critic And Action Neural Networks In Dual Heuristic Programming Method, Christian Peter Paintz

Dissertations and Theses

This thesis discusses strategies for and details of training procedures for the Dual Heuristic Programming (DHP) methodology. This and other approximate dynamic programming approaches (HDP, DHP, GDHP) have been discussed in some detail in the literature, all being members of the Adaptive Critic Design (ACD) family. The example applications used here are the inverted pendulum problem and a fully nonlinear constant velocity bicycle steering model. The inverted pendulum has been successfully controlled using DHP, as reported in the literature. This thesis suggests and investigates several alternative D HP training procedures and compares their performance with respect to convergence speed and …

New Approaches To Column Compatibility Checking And Column-Based Input/Output Encoding For Curtis Decompositions Of Completely Or Incompletely Specified Switching Functions, Michael A. Burns

Dissertations and Theses

Cube calculus is an algebraic model used to process boolean functions. Cube calculus operations are widely used in logic optimization, logic synthesis, image processing and recognition, machine learning, and other applications which require massive logic operations.

The cube calculus operations can be carried out on general-purpose computers. Since these operations can involve several levels of nested loops, this approach has poor performance.

A cube calculus machine which has a special data path designed to speed up cube calculus operations is presented in this thesis. This c-qbe calculus machine can execute cube calculus operations 10 to 25 times faster than the …

Erp Analysis Using Matched Filtering And Wavelet Transform, Xueming Lin

Dissertations and Theses

Event related potentials (ERP's) carry very important information that relates to the performance of the brain functions of the human being. Further studies have identified that one component, in particular, P₃₀₀, is affected by the memory process. Matched filter is used to improved the SNR of signal ERP's. We use the output of the matched filter to distinguish the difference of the waveforms between normal subjects and memory-impaired subjects. In our study, we found that the peak values of the matched filtering output were different between normal subjects and memoryimpaired subjects. Also, as an application, wavelet transform is …

Adaptive Notch Filter, Yuchen Huang

Dissertations and Theses

The thesis presents a new adaptive notch filter (ANF) algorithm that is more accurate and efficient and has a faster convergent rate than previous ANF algorithms. In 1985, Nehorai designed an infinite impulse response (UR) ANF algorithm that has many advantages over previous ANF algorithms. It requires a minimal number of parameters with constrained poles and zeros. It has higher stability and sharper notches than any ANF algorithm until now. Because of the special filter structure and the recursive prediction error (RPE) method, however, the algorithm is very sensitive to the initial estimate of the filter coefficient and its covariance. …

Unified Bias Analysis Of Subspace-Based Doa Estimation Algorithms, Yang Lu

Dissertations and Theses

This thesis presents the unified bias analysis of subspace-based DOA estimation algorithms in terms of physical parameters such as source separation, signal coherence, number of senors and snapshots. The analysis reveals the direct relationship between the performance of the DOA algorithms and signal measurement conditions. Insights into different algorithms are provided. Based upon previous first-order subspace perturbations, second-order subspace perturbations are developed which provide basis for bias analysis and unification. Simulations verifying the theoretical bias analysis are presented.

A New Approach To The Optimal Filtering Of Differential Phase Measurements Of Gps Signal In The Precision Survey, Shengan Wang

Dissertations and Theses

The Global Positioning System (GPS) has become popular research and application interests in surveying and many other areas. Nowadays, the accuracy of the Differential GPS can easily reach the order of a few meters. Yet, there are still many ways to exploit the GPS system signal carrier to improve the accuracy to less than meter level. In this thesis, a new approach to improve the accuracy to less than meter level is presented while the observer is in the dynamic situation. In order to reach the sub-meter accuracy, we measure on the carrier phase difference (The L1 carrier frequency is …

Designing And Simulating A Multistage Sampling Rate Conversion System Using A Set Of Pc Programs, David Joseph Hagerty

Dissertations and Theses

The thesis covers a series of PC programs that we have written that will enable users to easily design FIR linear phase lowpass digital filters and multistage sampling rate conversion systems. The first program is a rewrite of the McClellan-Parks computer program with some slight modifications. The second program uses an algorithm proposed by Rabiner that determines the length of a lowpass digital filter. Rabiner used a formula proposed by Herrmann et al. to initially estimate the filter length in his algorithm. The formula, however, assumes unity gain. We present a modification to the formula so that the gain of …

Performance Analysis Of Doa Estimation Algorithms Using Physical Parameters, Hui Liu

Dissertations and Theses

Analytical performance analysis on Direction-Of-Arrival (DOA) estimation algorithms has attracted much excellent research in recent years, various statistical properties have been revealed. However, in most of these analyses, insights of the performance were masked because of the involvement of singular values and singular vectors which depend on the character of the algorithms and data structures in a complex and nonlinear manner.

Decomposition Of Measured Contours Into Geometric Features For Dimensional Inspection, Devaraj Rajkumar

Dissertations and Theses

Image processing systems used in Vision Assisted Dimensional Inspection usually output a set of boundary pixels representing the part edges. This boundary information must be divided into several subsets representing the various edges of the actual object, so that comparisons with the nominal part can be made.

The purpose of this project is to devise a method to divide the set of pixels obtained from the image processing system into subsets of pixels. Each of these subsets represent an edge in the contour of the actual object. This method must also detect transition points between the adjacent features. This project …

An Effective Cube Comparison Method For Discrete Spectral Transformations Of Logic Functions, Ingo SchäFer

Dissertations and Theses

Spectral methods have been used for many applications in digital logic design, digital signal processing and telecommunications. In digital logic design they are implemented for testing of logical networks, multiplexer-based logic synthesis, signal processing, image processing and pattern analysis. New developments of more efficient algorithms for spectral transformations (Rademacher-Walsh, Generalized Reed-Muller, Adding, Arithmetic, multiple-valued Walsh and multiple-valued Generalized Reed- Muller) their implementation and applications will be described.

Investigation Of Techniques For High Speed Cmos Arbitrary Waveform Generation, Albert Henry Nehl

Dissertations and Theses

Today a growing number of applications in design engineering, production and environmental testing, and system service require specific analog waveforms and digital patterns. Such requirements are neither satisfactorily nor easily met by the use of standard function or single purpose, custom generators.

Traditional methods of waveform generation suffer from undesirable complexity or mediocre performance and are otherwise limited. For the majority of arbitrary waveform generation applications, including medical engineering, modal analysis and electronic engineering, direct digital synthesis techniques are satisfactory. Direct digital synthesis, based generally on periodic retrieval of predetermined amplitude values, may be used to 2 generate such waveforms. …

Endless State-Of-Polarization Control For Coherent Optical Communication Systems Using Nematic Liquid Crystal, Scott H. Rumbaugh

Dissertations and Theses

One of the obstacles to coherent fiber-optic communications is the unpredictable polarization drift which necessitates the use of an active polarization controlling system to match the polarizations of the signal and local oscillator. The polarization match must be maintained during the reset of any of the finite range components to prevent loss of data. We present a novel and practical system which uses three liquid crystal devices for the polarization matching process. Also, the required reset control algorithm and its derivation are described in detail.

A Comparative Study Of The Performance Of Various Image Analysis Methods For Dimensional Inspection With Vision Systems, Ralf Heinrich Koeppe

Dissertations and Theses

Dimensional inspection with Vision Systems requires a careful selection of image analysis methods in order to obtain accurate information about the geometry of the parts to be measured.

The purpose of this project is to study, implement and compare different image evaluation methods and to show their strengths and weaknesses with respect to dimensional inspection. Emphasis is made on the inspection of circular features. The criteria of comparison for these methods are discussed. Using synthetically generated images, various analysis methods are compared and conclusions for their use are drawn. Results of the comparison show that the selection of a method …

Invariant Pattern Recognition Algorithm Using The Hough Transform, Duwang Li

Dissertations and Theses

A new algorithm is proposed which uses the Hough Transform to recognize two dimensional objects independent of their orientations, sizes and locations. The binary image of an object is represented by a set of straight lines. Features of the straight lines, namely the lengths and the angles of their normals, their lengths and the end point positions are extracted using the Hough Transform. A data structure for the extracted lines is constructed so that it is efficient to match the features of the lines of one object to those of another object, and determine if one object is a rotated …

A New General Purpose Systolic Array For Matrix Computations, Hai Van Dinh Le

Dissertations and Theses

In this thesis, we propose a new systolic architecture which is based on the Faddeev's algorithm. Because Faddeev's algorithm is inherently general purpose, our architecture is able to perform a wide class of matrix computations. And since the architecture is systolic based, it brings massive parallelism to all of its computations. As a result, many matrix operations including addition, multiplication, inversion, LU-decomposition, transpose, and solutions to linear systems of equations can now be performed extremely fast. In addition, our design introduces several concepts which are new to systolic architectures:

- It can be re-configured during run time to perform different …

Two New Parallel Processors For Real Time Classification Of 3-D Moving Objects And Quad Tree Generation, Farjam Majd

Dissertations and Theses

Two related image processing problems are addressed in this thesis. First, the problem of identification of 3-D objects in real time is explored. An algorithm to solve this problem and a hardware system for parallel implementation of this algorithm are proposed. The classification scheme is based on the "Invariant Numerical Shape Modeling" (INSM) algorithm originally developed for 2-D pattern recognition such as alphanumeric characters. This algorithm is then extended to 3-D and is used for general 3-D object identification. The hardware system is an SIMD parallel processor, designed in bit slice fashion for expandability. It consists of a library of …