Electrical and Electronics Commons^™

Function Prediction Using Recurrent Neural Networks, Randall L. Lindsey

Theses and Dissertations

A fully recurrent neural network was applied to the function prediction problem. The real-time recurrent learning (RTRL) algorithm was modified and tested for use as a viable function predictor. The modification gave the algorithm a variable learning rate and a linear/sigmoidal output selection. Verifying the networks ability to temporally learn both the classic exclusive-OR (XOR) problem and the internal state problem, the network was then used to simulate the frequency response of a second order IIR lowpass Butterworth filter. The recurrent network was then applied to two problems: head position tracking, and voice date reconstruction. The accuracy at which the …

Gabor Segmentation Of High Resolution Synthetic Aperture Radar Imagery, Michael A. Hazlett

Theses and Dissertations

This thesis investigates the use of Gabor filters and a radial basis function (RBF) network for segmentation of high resolution (1 foot by 1 foot) synthetic aperture radar (SAR) imagery. Processing involved correlation between the SAR imagery and Gabor functions. Two methods for selecting the optimal Gabor filters are presented. This research used complex Gabor functions and operated on single polarization HH complex data. Following the selection f the proper Gabor filters, correlation coefficients for each image were calculated and used as features for the RBF network. Provided are results demonstrating how Gabor processing and a RBF network provide image …

Face Recognition With The Karhunen-Loeve Transform, Pedro F. Suarez

Theses and Dissertations

The major goal of this research was to investigate machine recognition of faces. The approach taken to achieve this goal was to investigate the use of Karhunen-Loe've Transform (KLT) by implementing flexible and practical code. The KLT utilizes the eigenvectors of the covariance matrix as a basis set. Faces were projected onto the eigenvectors, called eigenfaces, and the resulting projection coefficients were used as features. Face recognition accuracies for the KLT coefficients were superior to Fourier based techniques. Additionally, this thesis demonstrated the image compression and reconstruction capabilities of the KLT. This theses also developed the use of the KLT …

An Investigation Of The Application Of Artificial Neural Networks To Adaptive Optics Imaging Systems, Andrew H. Suzuki

Theses and Dissertations

Recurrent and feedforward artificial neural networks are developed as wavefront reconstructors. The recurrent neural network studied is the Hopfield neural network and the feedforward neural network studied is the single layer perceptron artificial neural network. The recurrent artificial neural network input features are the wavefront sensor slope outputs and neighboring actuator feedback commands. The feedforward artificial neural network input features are just the wavefront sensor slope outputs. Both artificial neural networks use their inputs to calculate deformable mirror actuator commands. The effects of training are examined.

Design And Testing Of A Lightweight, Planar Microwave Lens, Paul M. Proudfoot

Theses and Dissertations

This thesis documents the design, fabrication, and testing of a lightweight, planar microwave lens. The lens consists of two planar arrays of lightweight printed circuit antennas interconnected by microstrip lines whose length varies as a function of position on the lens. Microwave energy is coupled from one face of the lens to the other by means of a slot etched in the ground plane. At 8.0 GHz the lens is ten wavelengths across, is less than 1/8 inch thick, and weighs less than 4 pounds. The lens can scan a beam over a +/- 30 degree region in both the …

Delay Performance Of Cebus And Modified Cebus Schemes, Altaf Hussain

Theses

In this thesis the performance of the CEBUS (Consumer Electronics Bus), a local area network proposed by the Electronics Industries Association for the 'Smart House,' is analyzed under varying network load and traffic patterns that would occur in a typical home environment. Reduced complexity versions of the CEBUS scheme are compared with the standard CEBUS scheme.

It was found from simulation analysis that the delay in the CEBUS network, at a given network bandwidth, is dependent upon the message priority and the traffic pattern in the network. It was also observed that the performance of a modified CEBUS scheme was …

Broadband Isdn Control Using Neural Networks, Stavros Zervoudakis

Theses

Broadband Integrated Services Digital Networks (B-ISDN) are considered to become the standard communications networks in the near future. Major research efford is being applied in determining the appropriate mode of operation for B-ISDN systems. Two proposals, the Synchronous Transfer Mode (STM) and the Asynchronous Transfer Mode (ATM) claim to best fit the B-ISDN transmission requirements individually. A new method for B-ISDN traffic control is proposed in this paper. This method assigns priority levels to incoming traffic while integrating STM with ATM. A neural network is used to decide how the multiplexing of the incoming packets will be performed. Several traffic …

Direct Recovering Of Surface Structure Characterized By An Nth Degree Polynomial Equation Using The Uoff Approach, Mazen A. Salhi

Theses

There are two different approaches for estimation of structure and/or motion of objects in the computer vision community today. One is the feature correspondence method, and the other is the optical flow method [1]. There are many difficulties and limitations encountered with the feature correspondence method, while the optical flow method is more feasible, but requires a substantial amount of extra calculations if the optical flow is to be computed as an intermediate step.

Direct methods have been developed [2-4], that use the optical flow approach, but avoid computing the full optical flow field as an intermediate step for recovering …

Canmol : Expert System In Medicine, Kulin Hemani

Theses

The intriguing words for the nineteen nineties are "Expert Systems", and "Artificial Intelligence". To a layman these terms conjure up thoughts of spaceships and futuristic beings. In reality they exemplify man's capability to harness technology.

Expert systems are computer software programs that make use of previously accumulated knowledge to assist in solving complex problems.

In recent years the practice of medicine has changed dramatically from a fundamentally individually deductive approach to an Artificial Intelligence based science. The modern physician no longer draws conclusions simply by examining the patient. Today the physician employs sophisticated computer based diagnostic tools to generate a …

Radiation From The Cylindrical Dielectric Taper, Tor-Odd Hoydal

Theses

The radiation problem of the cylindrical dielectric taper, excited by the HE₁₁ dominant mode, is studied. A theoretical model and a numerical investigation into the radiation characteristics of such a tapered structure are presented. The model is based upon the combination of the exact modal field solutions for a step profile cylindrical dielectric waveguide, Local Mode Theory, and the Equivalence Principle for the determination of equivalent current densities.

The far-zone radiation field is considered to be the summation of the fields produced by two radiative components in the dielectric rod/conical structure: The "aperture" of the uniform dielectric rod at …

Knowledge Based Designing Of Broadband Coaxial Cable Networks, Tauqir Hameed

Theses

The objective of this research was to build a broadband coaxial cable network. One of the basic requirement of a network is the physical link transmission medium. For the purposes of this research, broadband coaxial cable was chosen as a physical medium of data communication. The approach taken in this research was to divide network design into three main steps. In the first step, the decision on topology of the network was determined. Both Star and Minimum Spanning Tree topologies were used and then the Esau-Williams Algorithm was employed to minimize the cable length. In the second step, various signal …

Gaseous Component Identification With Polymeric Film Sensor, Edward S. Kolesar Jr.

AFIT Patents

A sensor having an interdigitated gate electrode field effect transistor (IGEFET) coupled to an electron beam evaporated copper phthalocyanine thin film is used to selectively detect parts-per-billion concentration levels of atmosphere contaminants such as nitrogen dioxide (NO₂) and diisopropyl methylphosphonate (DIMP). The sensor is excited with a voltage pulse, and its time- and frequency-domain response are examined. The envelopes of the magnitude of the normalized difference frequency spectrums reveal features which unambiguously distinguish the NO₂ and DIMP exposures.

Performance Of The Extended Token Bus Network With Asymmetric Load Distribution, Hong Lin

Theses

The Token Bus, IEEE 802.4, provides high priority message options. The delay of high priority messages becomes an important criterion in the design and management of computer communication networks. A goal of the protocol is to allocate sufficient bandwidth and frequency of access to minimize of the high priority message delay. In this study, a simulation model of a priority based network, consisting of two token bus networks interconnected by a bridge is implemented. The load distribution among stations is asymmetric and the messages are divided into two priorities, high and low. Both fixed and variable token token holding times …

The Performance Evaluation Of A New Neural Network Based Traffic Management Scheme For A Satellite Communication Network, Dequan Liu

Theses

The thesis describes a new neural network based traffic management scheme for a satellite communication network. The scheme consists of two levels of management. The front end of the scheme is a derivation of the Kohonen's self-organization model to configure maps for the satellite communication network dynamically, The model consists of three stages. The first stage is the pattern recognition task which selects an exemplar map that best meets the current network requirements. The second stage analyzes the discrepancy between the chosen exemplar map and the state of the network, and adaptively modifying the chosen exemplar map to conform closely …

Graded Buried Channel Charge Coupled Device Design, Zengjing Wu

Theses

A process and device simulation study has been made to optimize a new type of buried channel CCD (BCCD) structure to be referred to as Graded BCCD (GBCCD) whose buried channel is formed by multiple implantations for achieving higher charge transfer efficiency. The 2-D process simulation tool SUPREM4 and 2-D device simulation tool PISCES2 were used in this research. The simulation results of the GBCCD with double channel implants are compared with the previously obtained experimental results.

Triple channel implants under the conditions of various process parameters were also simulated and the optimized case was chosen on the base of …

Mobile Manipulator Path Planning By A Genetic Algorithm, Min Zhao

Theses

The thesis addresses the path planning problem for a mobile manipulator system which is used to perform a sequence of tasks specified by locations and minimum oriented force capabilities. The problem is to find an optimal sequence of base positions and manipulator configurations for performing a sequence of tasks given a series of task specifications. The formulation of the problem is nonlinear. The feasible regions for the problem are nonconvex and unconnected. Genetic algorithms applied to such problems appear to be very promising while traditional optimization methods cause difficulties. Computer simulation is carried out on a three degrees-of-freedom manipulator mounted …

Mobile Robot Path Planning Based On Hierarchical Hexagonal Decomposition, Dan Zheng

Theses

In this thesis, a new algorithm based on hierarchical cell decomposition and artificial potential field is developed to solve the find-path problem for a mobile robot. The basic cell structure used for decomposition is a hexagon. The artificial potential field is based on an attractive force from the goal position and repelling forces from the obstacles. Computer simulations of the algorithm for various obstacle scenarios are also presented.

Measurement Of The Electrical And Optical Properties Of Thermid Polymide, Zhixiong Zhu

Theses

Electrical and optical properties of spin-coated THERMID polyimide EL-5010 have been studied and evaluated in this thesis. The D.C. bulk resistivities of 0.96^-2.56 μm thick samples were measured to be 10¹⁶ Ω-cm over the 27-160°C range using a test system consisting of Keithley 236 Source-Measure Unit. The bulk resistivity decreased to 2 x 10¹⁴Ω-cm at 200?C. The surface resistivity, 10¹⁵Ω/0 , and breakdown voltage, 350 V/pm, were also measured. The fixed charge concentration and mobile ion concentration in the THERMID polyimide EL-5010 insulator of a MIS capacitor structure were measured using a high …

Mapping Hierarchical Structures Onto Hypercubes, Sandip C. Patel

Theses

Hierarchically structured arrays of processors are appropriate for low-level and intermediate-level image processing and computer vision. The most widely used structure is the pyramid. Neverthless, pyramid machines do not support the efficient implementation of the majority of scientific algorithms and their cost may become prohibitively high. In contrast, hypercube machines efficiently implement a wide variety of important algorithms due to the versatility of the underlying network. In fact, hypercube-based systems have become commercially available. The hypercube network supports both regular and irregular communications and it can efficiently simulate a variety of other structures like rings, trees, meshes, pyramids, etc. Thus, …

Studies Of Electrical Properties Of Sio2 /Inp System, Milind S. Patel

Theses

Analyses of Electrical characteristics, of TilAu/SiO₂1InP structures, are presented here. The (MIS) Metal-Insulator- Semiconductor capacitors were fabricated with oxide thickness of 195nm, 420nm and 610nm respectively. The SiO₂ films were deposited on n-type Indium Phosphide(/nP) using Rapid Thermal Chemical Vapor Deposition (RTCVD). The characterization techniques employed in these experiments were (I-V) Current Voltage and High Frequency Capacitance Voltage (C-V) techniques. The electrical parameters obtained from the I-V and C-V measurements have been discussed here.

Electrical Characterization Of Preoxide Growth Fluorine Incorporated Thin Sio2 Films, Nilay J. Patel

Theses

Analyses of electrical characterization, of thin SiO₂ films grown on virgin and etched silicon surfaces, are presented here. The characterization techniques employed in these experiments are High Frequency Capacitance-Voltage (C-V), Capacitance-Time (C-T) & MOS Doping Profile and Static & Dynamic Current-Voltage (I-V) measurements. The structures, with areas ranging from 0.001-0.05 cm², used in this study are Al/Poly/SiO₂/Si/Al capacitors fabricated on virgin and C₂F₆ CHF₃ + He plasma exposed n-Si substrates. In this experiment, we have studied the effect of sacrificial oxidation. In this step, the oxide is thermally grown on the …

A Microengineered Beam Diaphragm Structure For The High Performance Pressure Sensor, Alkesh Shah

Theses

The thesis details the design for a novel microengineered beam diaphragm structure for piezoresistive pressure sensor applications. The beam diaphragm structure is used as a pressure sensing device, within a Wheatstone bridge circuit. The single crystal silicon with [100] orientation is used for the beam diaphragm structure.

It has the advantages of high sensitivity(2.56 v/v/mm of Hg), good linearity(3Kpa to 1.4 Mpa) and overpressure protection with the help of the bosses. It's protection range is about 400 times the standard pressure range. The beam diaphragm structure has two rectangular bosses which are formed by anisotropic etching. Two pairs of diffused …

Efficient Multiprocessor Scheduling Based On Genetic Algorithms, Hong Ren

Theses

The problem of multiprocessor scheduling can be stated as finding a schedule for a general task graph to be executed on a multiprocessor system so that the schedule length can be minimized. This scheduling problem in known to be NP-hard and methods based on heuristic search have been proposed to obtain optimal and sub-optimal solutions to the problem. Genetic algorithms have recently received much attention as robust stochastic searching algorithms for various optimization problems. In this thesis, we propose an efficient method based on genetic algorithms to solve the multiprocessor scheduling problem. The representation of the search node will be …

Transmission Ellipsometry On Transparent Unbacked Or Embedded Thin Films With Application To Soap Films In Air, R. M.A. Azzam

Electrical Engineering Faculty Publications

The ratio ρ_t = T_p/T_s of the complex amplitude transmission coefficients for the p and s polarizations of a transparent unbacked or embedded thin film is examined as a function of the film thickness-to-wavelength ratio d/λ and the angle of incidence Φ for a given film refractive index N. The maximum value of the differential transmission phase shift (or retardance), Δ_t = argρ_t, is determined, for given N and Φ, by a simple geometrical construction that involves the iso-Φ circle locus of ρ_t in the complex plane. The …

Simulation Study Of Signal Coupling For Monolithic Gaas 8-12 Micron Direct Injection Mqw Image Sensor, Lei Zhang

Theses

The simulation of signal coupling for a single pixel of a monolithic 8-12 micron GaAs/A1GaAs direct injection MQW infrared image sensor with 100% fill factor has been performed by using PISCES-IIB and PSPICE simulators. A vertical structure with the readout circuit built on the back of MQW was adopted for simulations. Both the static and transient characteristics were analyzed by simulation. The simulation was used to verify the doping profiles and device parameters for the n-GaAs MESFETs used in the read-out circuit.

The simulation result is as follow:

(1) A 900A thick p-GaAs (p=10¹⁶/cm3) substrate for GaAs MESFETs …

A Multiple Model Adaptive Estimator Using First And Second-Order Acceleration Models For Use In A Forward-Looking Infrared Tracker, Barbara J. Niblett

Theses and Dissertations

The performance of a multiple model adaptive estimator (MMAE) for an enhanced correlator/forward-looking-infrared tracker for airborne targets is analyzed in order to improve its performance. Performance evaluation is based on elemental filter selection and MMAE estimation error sizes and trends. The elemental filters are based on either first or second-order acceleration models. Improved filter selection is achieved by using acceleration models that separate the frequency content of acceleration power spectral densities into non- overlapping regions with second-order models versus the more traditional overlapping regions with first-order models. A revised tuning method is presented. The maximum a posteriori (MAP) versus the …

Investigation Of Adaptive Controllers For Puma Trajectory Tracking, Daniel J. Sims

Theses and Dissertations

Two robust model-based controllers and two decentralized adaptive controllers are experimentally evaluated. Algorithm evaluation is motivated by the need for controllers with good high speed tracking under varying payload conditions. The test case is a PUMA-560 robotic manipulator operating over the standard test suite. The model-based controllers are made robust by the addition of either an auxiliary input term or an adaptive feedforward compensator based on Lyapunov Theory. The model-based auxiliary input controller (MBAIC) adapts the gain matrices used in computing an additional torque to be combined with model-based feedforward and PD feedback torques. The adaptive model-based controllers adapt the …

High Frequency Scattering Code In A Distributed Processing Environment, Scott Suhr

Theses and Dissertations

Government agencies and academic institutions are very interested programming for concurrent processing to cut computer processing time. However, many of the world's problems have already been coded for conventional serial computers. This research demonstrates the feasibility of modifying existing serial codes for execution in a concurrent processing environment. A electromagnetic scattering prediction code known as NECBSC is incrementally modified to incorporated various levels of concurrent computing. The data processed by the code are completely independent, providing an avenue for data decomposition of the process. Portions of the data set are processed on each mode and the results combined for final …

Development Of A Personal Computer Simulation Program Of The Ln-94 Inertial Navigation System, Thomas E. Flynn

Theses and Dissertations

This thesis develops a simulation that emulates a Litton LN-94 INS. Real time simulation is achieved for a stationary navigation case and a straight-and-level flight trajectory are provided to exercise the simulation dynamics. The user interface is simple and requires little previous knowledge to operate. The communication link with the MILSTD 1553B bus provide a realistic environment in which to collect and analyze data. The simulation is processed on an 80386 IBM compatible personal computer. A 23-state INS error model is implemented and integration is performed by a fifth-order Kutta-Merson routine. The simulation is interfaced to a plotting routine and …

An Alternative Algorithm For Discrete-Time Filtering, Frederick H. Zeitz Iii

Theses and Dissertations

The discrete-time Kalman filter is a conditional mean estimator of the states of a linear stochastic process, conditioned on the previous state and current measurements. It assumes that the states and noise inputs can be represented as jointly Gaussian random variables. The influence diagram is a decision analysis tool. Under certain conditions, it can represent continuous, jointly Gaussian random variables. The conditioning order of the random variables may be changed (using Bayes' rule) so that any random variable can be conditioned on any other subset of random variables in the diagram. Under these conditions, an influence diagram can represent the …