School of Engineering & Computer Science

Permanent URI for this communityhttps://hdl.handle.net/2104/4749

Browse

Now showing 1 - 20 of 313

A deep convolutional neural network approach for biomedical applications.
(December 2022) Nguyen, Hanh Hong, 1991-; Schubert, Keith Evan.
Deep learning is a subset of machine learning that uses multi layer neural networks to perform desired tasks by using trained models. Neural networks are nonlinear mapping systems whose structure and function are loosely modeled on the physical structure of the nervous systems in humans and animals. In deep learning, convolutional neural networks (CNNs) have been used to analyze visual tasks for more than 40 years. Since the mid-2000s, they have revolutionized image processing and analysis. The goal of this dissertation is designing a deep CNN approach for biomedical applications, including automation of the process of colon polyps classification as well as single particle identification in radiation therapy.
A method for replicating ice accretion roughness using multi-scale analog distributions.
(2018-03-23) Clemenson, John-Mark, 1994-; McClain, Stephen Taylor.
A new approach of creating an analog surface for a surface with natural roughness is presented based on the hypothesis that the dominant skin friction and heat transfer augmentation mechanisms are 1) the vortex shedding from the roughness elements and 2) the interaction of the shed vortices with the next downstream roughness elements. An autocorrelation function was employed on a real ice surface to capture the root-mean-square-roughness height, the primary streamwise wavelength, and the aspect ratio of surface microscales. Two distributions of deterministic roughness, one of ellipsoids and one of elliptical cones, were created to match the primary features of the real ice surface. This work describes the analog surface creation method and characterizes the convective enhancement and velocity boundary layer development of the analog surfaces. The convective enhancement and velocity boundary layer results associated with the analog surfaces are then compared to those of the real ice surface.
A model of clocked electric field inputs for molecular quantum-dot cellular automata.
(2019-11-08) Henry, Jackson Alan, 1994-; Blair, Enrique Pacis.
Quantum-dot cellular automata (QCA) is a low-power, high-speed, beyond- CMOS approach to general-purpose computing [1]. Elementary devices called “cells” are implemented using mixed-valence molecules with redox centers having a few quan- tum dots. These support three distinct localized electronic states labeled “0”, “1”, and “Null”. Cells can be clocked to either the “Null” state or an active (“0” or “1”) state using the vertical component of an applied electric field. Clocking provides power gain for restoring weakened signals and allows synchronous control of QCA circuits. In this paper, clocked molecular QCA circuits are simulated in the presence of an applied input field, using the intercellular Hartree-Fock approximation [2]. In- put circuits and down-stream circuits function in the presence of the input field and unwanted field fringing from electrodes. This emphasizes that widely-available fabri- cation techniques may be used to form electrodes for writing bits to molecular QCA circuits.
A neural inspired grid cell aid for robot intertial navigation.
(2018-05-08) Martinez, Moises, 1993-; Koziol, Scott M.
Accurate position information is needed for a robot's guidance and control systems. Therefore, the navigation system is foundational for a robot's interactions with the world. This thesis explores a biologically inspired method of navigation with the goal of improving navigational accuracy. The desire to integrate brain-inspired methods with conventional signal processing methods is based on animals' innate ability to successfully navigate through habitats. Therefore, it is reasonable to explore the way animals process brain signals to navigate and leverage it for robot navigation. This thesis uses a combinatorial model of map formation and localization with grid cells to aid dead reckoning with an accelerometer and gyroscope to show grid cells are a viable aid in idiothetic navigation. The results show that the grid cell aided navigation systems shows better performance with longer paths and higher noise values with an improvement of 57.24 cm for a 32 m path with 3 σ noise.
A photogrammetric method for measuring transient frost surface roughness.
(2018-11-26) Miyauchi, Taber Scott, 1994-; McClain, Stephen Taylor.
Cold-soaked fuel frost (CSFF) is a form of aircraft wing contamination that results in aerodynamic degradation. Unless a certification exemption for a specific aircraft model is provided by the FAA, an operator of an aircraft is not allowed to takeoff with CSFF on aircraft wing surfaces. To assist manufacturers in the design of air vehicles and to assist the FAA in evaluating exemption requests, frost roughness evolution must be characterized and modeled. However measuring transient frost roughness is difficult because of the optical and material properties of frost. This experimental investigation presents a nonintrusive, in-situ method of measuring frost roughness using structure-from-motion photogrammetry. This method was validated with fabricated rough surfaces and was used on a closed-loop psychrometric wind tunnel that is capable of producing frost formations in various environmental conditions. The resulting roughness parameters and detailed height distributions were analyzed to determine the effects of each environmental variable.
A Study of chemical treatments and processing for banana fiber-reinforced LDPE composites.
(2016-07-25) Chester, Patrick, 1991-; Jordan, William Mark, 1950-
Natural fiber-reinforced polymeric composites have been gaining popularity due to their low weight and high sustainability potential. Banana pseudo-stem fibers in particular present a unique opportunity for reinforcement since they are widely grown and are currently considered an agricultural waste product. However, poor bonding between the hydrophilic fiber and a comparatively hydrophobic thermoplastic matrix, such as LDPE, have limited composite properties. This study looks at two known natural fiber treatments designed to promote interfacial bonding, peroxide and permanganate treatment, and documents their effect on banana pseudo-stem fibers and a banana fiber/LDPE injection molded composite. It was found that peroxide treatment enhances the strength of individual fibers but is thermally unstable. Permanganate treatment decreases the moisture absorption of the fibers and improves interfacial bonding but provides no advantage over untreated fibers in composite properties. This study also looks into special processing considerations for injection molding natural fiber composite parts.
A study of frost growth and roughness on a cold flat plate under forced convection.
(2020-03-19) Zhang, Tongxin, 1987-; O'Neal, Dennis.
The growth and roughness of frost formed on a cold flat plate under forced convection was characterized through experimental observations and analysis. The results indicated that the evolution of frost roughness was mainly determined by the complex structures of frost crystals appearing on the top of the frost layer. The frost crystal type, frost crystal size, and frost crystal distribution were strongly affected by environmental conditions and collectively determined the formation of frost surface roughness. Frost roughness parameters, including frost root-mean-square height and frost surface skewness, were measured by utilizing a 3-D photogrammetric method. The growth of frost roughness under various environmental conditions were then assessed. Both frost root-mean-square height and surface skewness increased with time until reaching a maximum value (up to 0.69 mm and 2.34, respectively), before starting to decrease. The frost surface skewness peak, the maximum frost root-mean-square height, and the time for the appearance of frost root-mean-square height typically varied linearly with the surrounding environmental conditions. The experimental investigation also showed that frost roughness varied over both time and location on the test surfaces. The frost equivalent sand-grain roughness height, calculated by using the frost root-mean-square height and frost surface skewness, was applied to represent the unevenness of frost roughness on cold substrates. Through an ANOVA test on the data of the frost equivalent sand-grain roughness height, air humidity and air temperature with p-values less than 0.05 were indicated to be the key factors determining the changes in the variation of the frost roughness in space. Empirical correlations were developed to describe the relationships between the appearance of the frost roughness peak and environmental conditions. Finally, based on the frost roughness data collected in this study, a predictive model of frost roughness parameters, including frost root-mean-square height, frost surface skewness, and frost equivalent sand-grain roughness height, was established with the R-square values of 0.57, 0.65, and 0.53, respectively.
Ab-initio models of quantum dot cellular automata molecules.
(December 2022) Liza, Nishat Tasnim, 1993-; Blair, Enrique Pacis.
Molecular quantum-dot cellular automata (QCA) is a charge based, low-power, energy-efficient alternative to transistor-based, general-purpose computation. In molecular QCA, redox centers of a mixed-valence (MV) molecule function as coupled quantum dots, and localized charge states of the molecule encode binary information useful for classical computing. Molecular QCA promises ultra-high device densities, THz-scale switching speeds and room temperature readout. While the fundamental principle of molecular QCA have been tested and established, major challenges must be overcome to successfully implement molecular QCA. This work applies ab-initio techniques in the design and modeling of candidate MV molecules for QCA. Here, we study and characterize ∼ 1-nm-scale MV QCA molecules using first principle calculations. The structural and electronic properties of QCA molecules are calculated utilizing Hartree-Fock, Post-Hartree-Fock and Density functional theory (DFT) methods. Asymmetric, cationic, MV molecules are designed for spectroscopic state readout of QCA devices at room temperature. Tip-enhanced Raman spectroscopy is proposed to detect the state of QCA devices in a circuit if the QCA molecules have slightly dissimilar quantum dots. Clocked zwitterionic three-dot QCA molecules with built-in counterions at the center of the molecules are modeled. The choice and design of the central linkers of these molecules determines number of mobile charges in the molecules for encoding the device states on the three quantum dots. These molecules show different device responses to applied clocking electric field based on different central linkers designed and used, similar to the complementary responses of PMOS and NMOS transistors to gated voltage control. Counterion effects on QCA candidate molecules are also explored in terms of electron transfer parameters. The complete active space self consistent field (CASSCF) method is used to calculate electron transfer (ET) matrix element and inner-sphere reorganization energy of the molecules in the presence of nearby counterions. Results demonstrate that randomly placed externel counterions may degrade device states by causing mobile charge to localize in undesirable ways on the QCA molecule. New zwitterionic molecules with a built-in counterion are proposed to eliminate unpredictable effects of external counterions in QCA circuits. Novel organometallic zwitterionic QCA molecules with ferrocene dots are designed and proposed for synthesis. The chemical stability of these ferrocene based molecules are evaluated by theoretical calculations. The synthesis of these stable zwitterionic molecules by collaborating experimental chemists is in progress and may open a new path to realize molecular QCA computing. A new machine-learning-based DFT functional, DM21 is investigated and benchmarked against traditional methods by comparing the calculated ET matrix elements of several QCA molecules. Preliminary results calculated from DM21 functional did not show significant improvements in accuracy and computational cost. Modification and improvement of the neural network used in the development of the functional, as well as the underlying code is proposed which might open new path to computationally inexpensive QCA calculations.
Accelerating path planning algorithms with high level synthesis tools and FPGAs.
(2013-05-15) Trower, John W.; Duren, Russell Walker.; Electrical and Computer Engineering.; Baylor University. Dept. of Electrical and Computer Engineering.
Accelerating path planning algorithms with field programmable gate arrays (FPGA) allows the designer to achieve significant performance increases over using a traditional central processing unit (CPU). Converting an algorithm to run on an FPGA is a complicated and time consuming process. This thesis develops and verifies a design framework that demonstrates how to design a path planning algorithm in a high level language, then convert the algorithm into hardware description languages using high level synthesis tools. This design framework will be used to demonstrate the acceleration of a genetic algorithm.
Adaptive load impedance optimization for power amplifiers in reconfigurable radar transmitters.
(2013-05-15) Martin, Joshua Lee.; Baylis, Charles Passant, 1979-; Electrical and Computer Engineering.; U.S. Naval Research Laboratory.; Baylor University. Dept. of Electrical and Computer Engineering.
A fundamental tradeoff exists in radar transmitter design between linearity and efficiency due to the signal amplification. These transmitters are driven into saturation in order to increase efficiency, but may potentially violate regulatory spectral mask limitations. An adaptive method for optimizing linearity and efficiency for power amplifiers in radar transmitters is presented. This approach uses intelligent search techniques with load-pull measurements for power-added efficiency (PAE) and adjacent channel power ratio (ACPR) to dynamically maximize the PAE while meeting spectral requirements. Using load-tuning, an algorithm performs a steepest ascent search for the PAE optimum load reflection coefficient, followed by a steepest descent search for ACPR. The steepest descent search, when begun at the PAE optimum, approximates the Pareto optimal frontier between the two objectives. This trace enables PAE to be maximized for an imposed limit on ACPR, optimizing the performance of adaptive radar transmitters under spectral mask constraints.
Aerodynamic and aeroacoustic design considerations for small-scale, fixed-pitch, horizontal-axis wind turbines operating at low reynolds numbers.
(2017-03-22) Hays, Andrew W., 1993-; Van Treuren, Kenneth W.
Renewable energy is needed now more than ever to meet demands as conventional energy sources are being depleted. Wind energy has the potential to provide a substantial amount of renewable energy. Of increasing interest is the development of small wind turbines for residential and urban applications. In these settings, a new concern is the noise that these turbines produce due to their operation at higher rotational speeds. Experimental noise data for the NREL S823, Eppler 387, NACA 4412 and NACA 0012 2D airfoils were taken for Reynolds number from 50,000 to 200,000. From these studies, trends based upon airfoil shape were gathered and used in choosing a new airfoil for testing a new rotor design. From these trends, two new airfoils were chosen to compare to the baseline rotor that utilized the S823 airfoil. The two new rotors, each utilizing the Eppler 216 and SD 7062 airfoils, showed an increase of power production from 5-60%. These rotors were also able to decrease the generated noise by up to 9.5 dB(A).
Aerodynamic and aeroacoustic design of small unmanned aircraft system propellers at low Reynolds numbers.
(2020-11-19) Sanchez, Ricardo D., 1996-; Van Treuren, Kenneth W.
The Small Unmanned Aircraft System (sUAS) has become an overwhelmingly important asset for military intelligence, surveillance, and reconnaissance in addition to a multitude of needs in the commercial industry. More research should investigate sUAS propulsion systems and specifically the propellers, largely responsible for noise generation and inefficiencies in power consumption at low Reynolds numbers. Experimental noise data compared stock, modified, and five bladed propellers reducing tip vortex strength and noise generation. Results showed measurable far field sound decay and five bladed noise reductions of 5 dBA. Three motors compared propeller power consumption and resulted in increased electrical efficiencies of 14.5% and 31.3%. An airfoil study showed the GOE358 as the most aerodynamically efficient airfoil tested. A Prandtl bell-shaped lift distribution, minimum induced loss design, was applied to a propeller resulting in decreased power consumption and improved electrical efficiency by 18.51% with a SPL reduction of 11.15 dBA compared against the Baseline propeller. The Baseline propeller used an industry standard minimum loss propeller design.
Aerodynamic design considerations for small-scale, fixed-pitch, horizontal-axis wind turbines operating in class 2 winds.
(2012-08-08) Burdett, Timothy A.; Van Treuren, Kenneth W.; Engineering.; Baylor University. Dept. of Mechanical Engineering.
Renewable sources of energy, such as wind, are necessary to meet the growing demand as conventional energy sources are depleted. Very little research has been accomplished to improve wind turbine performance in Class 2 winds. This work experimentally analyzed techniques to improve the aerodynamic performance of small-scale, fixed-pitch, horizontal-axis wind turbines in Class 2 winds. Experimental data for the S823 airfoil was taken for Reynolds numbers from 50,000 to 200,000. A trip strip was shown to improve airfoil performance for Reynolds numbers below 100,000. Additional wind tunnel studies validated the wind turbine testing procedure. Using blade element theory (BET) and blade element momentum theory (BEMT), wind turbine blades with optimum angle of twist were designed and tested. Results suggest the BEMT-optimized blades will perform better at the design point. The sensitivity of the design angle on power production was also examined, resulting in negligible difference for the conditions tested.
Age classification from facial images for detecting retinoblastoma.
(2012-11-29) Chiam, Tak Chien.; Hamerly, Gregory James, 1977-; Computer Science.; Baylor University. Dept. of Computer Science.
Facial age estimation from images is a difficult problem, both because it is naturally difficult to tell the exact age of a person visually, and because of the variations in images, such as illumination, pose, and expression. We want to classify people into two groups, children (age ≤ 5) and adults (age > 5), to facilitate the detection of retinoblastoma, a type of pediatric cancer. Current regression based methods are ineffective, as they usually have mean absolute error of 5 years, which is too high for our purposes. We study the facial anthropometric measurements of humans at different ages, and build a system based on these growth patterns. We detect 76 facial landmarks using Active Shape Models, analyze all possible ratios computable from these landmarks, and use the best ratios as input into a Support Vector Machine. Our final system does very well on our problem, correctly classifying 85% of images.
Algorithmic specified complexity.
(2013-09-24) Ewert, Winston.; Marks, Robert J., II (Robert Jackson), 1950-; Electrical and Computer Engineering.; Baylor University. Dept. of Electrical and Computer Engineering.
Information theory is a well developed field, but does not capture the essence of what information is. Shannon Information captures something in its definition of improbability as information. But not all improbable events convey information. Kolmogorov complexity captures the idea of information as something easily described. But not all easily described objects are information. The proposed Algorithmic Specified Complexity takes into account both Shannon Information and Kolmogorov complexity to gain a fuller evaluation of information. We demonstrate this concept and develop several examples. We show the low probability of high Algorithmic Specified Complexity. We apply the concept to both images and functional machines from the Game of Life.
Algorithms for fast power amplifier load impedance and input power optimization using the Power Smith Tube.
(2016-04-08) Barkate, Joseph G., 1991-; Baylis, Charles Passant, 1979-
The increasing number of devices occupying the finite wireless broadband spectrum has led to serious concerns regarding spectral congestion. To facilitate radar and communication spectral coexistence, reconfigurable, adaptive amplifiers are expected to be a critical component in future cognitive radar transmitters. This work details a method to visualize and simultaneously optimize the load reflection coefficient and input power of a power amplifier device in order to achieve the highest possible efficiency while meeting a predefined spectral spreading constraint. The proposed vector-based search utilizes gradients as well as momentum in the Power Smith Tube in order to optimize accurately with as few experimental queries as possible. This method is proven to be feasible in higher dimensions, allowing for future implementation in real-time reconfigurable power amplifiers.
Alternatives using the Leap Motion to extend mid-air word-gesture keyboards.
(2015-12-14) Benoit, Garrett.; Poor, G. Michael.
Lately, the use of touchless, mid-air, gesture-based interactions has increased significantly due to the popularity of augmented and virtual reality and advances in other industries (e.g., medicine, gaming), and with this wide-spread application comes the need for effecient, mid-air text-entry. Word-gesture keyboards have garnered attention in recent years, now coming standard on most Android devices, offering efficient means of gesture-based text-entry. For the first time, Markussen et al. combined the two with the inception of Vulture \cite{ref_vulture}, the first mid-air, word-gesture keyboard, providing the fastest means of mid-air text-entry yet. This thesis builds on the findings of Markussen et al. and presents alternatives means for word separation in mid-air text-entry for word-gesture keyboards, exploring and identifying the problems of new techniques and presenting possible solutions. Of the new techniques, a bimodal approach shows great promise, reaching a mean text-entry rate of 15.8 Words Per Minutes for a single session with no training.
Ambiguity function magnitude inversion and applications of morphological dilation in POCS.
(2018-11-15) Yu, Albert Reynold, 1984-; Marks, Robert J., II (Robert Jackson), 1950-
This dissertation examines morphological dilation for applications in Projection onto Convex Sets (POCS) as well as the inversion of ambiguity function magnitude. In general, POCS solvers implement Least-Squares (LS) algorithms which intuitively minimize the Euclidean distance or L^2-norm of a proposed solution. However, there are situations where other error metrics can be advantageous. One such metric is the weighted minimized-maximum error, or minimax which minimizes the L^∞-norm. Multiple methods for evaluating the weighted, minimax error are investigated, and this dissertation will introduced a modified alternating projections algorithm utilizing morphological dilation on context sets to solve for the minimax. This is shown to have notable improvements over standard POCS solvers for selective signal synthesis applications, including Fresnel diffraction synthesis and Computed Tomography (CT) and associative memory image reconstruction. When multiple, conflicting objective functions are present, minimax solvers can be demonstrated to be an unbiased solver among conflicting constraints, avoiding the Least-Squares tendency to shift a solution towards the centroid. In addition, the ambiguity function magnitude inversion is shown to be possible and a regularized method for quickly inverting a given function to a valid family of source signals is detailed. The ambiguity function is a fundamental aspect of radar signal processing that is frequently described as non-invertible from its magnitude as the transform is not one-to-one. In the past, an inversion to constant phase shift is possible with the full magnitude and phase of the function, but the phase information is frequently stripped as extraneous for analysis. Unfortunately this practice prevents a clear inversion. However, this paper demonstrates that an inversion to a valid spawning signal is possible, and outlines a regularized method for achieving the desired magnitude response. This will give radar designers direct control over crafting ambiguity functions with mission-critical characteristics.
An algorithmic solution to mission planning via auto-routing algorithms.
(2016-04-08) Furgerson, Jase Lloyd.; Thompson, Michael Wayne.
Mission planning for radar jamming escort missions is a tedious and complex problem to solve. For years this type of mission planning has taken many hours to solve and used multiple pilots to develop a solution. This thesis discusses the development of a MATLAB solution for auto-routing aircraft for a mission planning scenario. The unique contribution of this work involves the development and implementation of an auto-router algorithm called the Augmented Mission Planning (or AMP) algorithm. The AMP algorithm is developed by combining techniques for Jamming Acceptability Region (JAR) construction, weighted map creation from DTED data, and an augmented version of the A* path finding algorithm. This auto-router preforms within a typical mission planning system framework and we demonstrate the effectiveness of this approach for determining mission planning in a timely manner.
An anisotropic topology optimization method for carbon fiber-reinforced Fused Filament Fabrication.
(2016-07-26) Hoglund, Robert M., 1992-; Smith, Douglas E., 1962-
Topology optimization has become an increasingly important field in materials engineering for determining optimal material distribution given certain loading conditions. As the desire for stronger, lightweight structural parts grows as well as the influence of additive manufacturing in the composites industry, there is a need for determining optimal structural orientation of Fused Filament Fabrication (FFF)-produced parts, especially those that have been improved with the addition of carbon fiber to the polymer feedstock. In this work, variations of the Solid Isotropic Material with Penalization (SIMP) method for topology optimization are proposed and modified, using the Adjoint Variable Method for design sensitivity analysis, a sensitivity filter to prevent checkerboarding, and a transversely isotropic material model. Using this technique, optimizations are performed for varying x- and y- direction orientation, as well as a variable material angle. Tested FFF-produced samples show agreement with computational results.

Browse

Browsing School of Engineering & Computer Science by Title