Digital Commons Network^™

Effects Of N>M Clos Networks On Edge Coloring And Matrix Decomposition Algorithms, Ming-Hsing Lin

Theses

Considered in this thesis are n > m Clos networks on edge coloring and matrix decomposition algorithms. This work uses a two-fold approach to examine the effects of n > m Clos networks on edge coloring and matrix decomposition algorithms. First, edge coloring and matching on bipartite multigraphs are applied to routing on Clos networks. It is demonstrated that edge coloring algorithms which find a minimum edge coloring set up the routing for n > m Clos networks with redundancy. The results presented here indicate that the Euler coloring algorithm has applications on n > m Clos networks because it has fast time complexity …

Optimization Of Throughput And Stability For Csma/Cd Channel, Liangshyng Jou

Theses

Multiple access technologies plays an important role in designing distributed computer communication system. Here, CSMA/CD has attracted researchers' attentions in the last ten years. In this paper, based on attributions of the slotted, nonpersistent CSMA/CD system, performance of CSMA/CD channel is analyzed. Adaptive Retransmission Technique is adopted to obtain the maximum throughput. Also, equilibrium contours are derived to discuss the stability of the system. And use FET to measure the stability of CSMA/CD channel that FET gives an approximate measure of the average uptime of an unstable channel.

On Implicational Dependency Families Possessing Finite Armstrong Relations, Kazem Taghva

Electrical & Computer Engineering Faculty Research

Let X ≠ 0 be a finite collection of nonempty relations over the relation scheme R(A1, A2 , ... , A,.); then the closure of X under embedding and direct product (up to isomorphism) is a finitely generated Implicational Dependency family (ID-family) generated by X. In this paper, we show that the class of finitely generated ID-families is identical to the class of those ID-families which possess a finite Armstrong relation.

Design Of Robotic Lan Controller And Map Interface, Jayesh M. Joshi

Theses

An application specific integrated circuit has been designed for use as a gateway between two local area networks. This silicon chip provides the gateway function to link the NJIT robotics network and the MAP IEEE-standardized network. This IC utilizes the available Motorola MC 68824 Token Bus Controller for the MAP data path. A special IO design compatible with an optical communication link is used to interface into the NJIT RLAN network.

The gateway IC serves as an asynchronous buffer between the 5Mbps MAP link and the 500 bps local optical link. The control protocol provides the master function for polling …

Fault-Tolerant Interconnection Networks For Multiprocessor Systems, Hamed Mohamed Nassar

Dissertations

Interconnection networks represent the backbone of multiprocessor systems. A failure in the network, therefore, could seriously degrade the system performance. For this reason, fault tolerance has been regarded as a major consideration in interconnection network design. This thesis presents two novel techniques to provide fault tolerance capabilities to three major networks: the Baseline network, the Benes network and the Clos network.

First, the Simple Fault Tolerance Technique (SFT) is presented. The SFT technique is in fact the result of merging two widely known interconnection mechanisms: a normal interconnection network and a shared bus. This technique is most suitable for networks …

An Intelligent Electrocardiogram Tutoring Package Ekgchallenger, Santosh K. Ananthraman

Theses

This thesis presents the development and implementation of an intelligent electrocardiogram tutoring package, namely, EKGCHALLENGER which is meant for use in medical schools to teach medical students a method of diagnosing electro cardiograms.

EKGCHALLENGER is an IBM - PC based, user friendly package. It works in two basic modes. The first is a tutorial mode wherein the user is taught the step-by-step diagnosis of electrocardiograms ( EKGs ) which are presented in order of increasing difficulty. The second is the challenger mode wherein the user is asked to diagnose EKGs which are randomly chosen from an EKG database. The diagnosis …

The Applicability Of Parametric Inverse Scattering For Optical Waveguides With Imperfect Boundaries, Hyoung Jin Choi

Theses

In this thesis, the applicability of an inverse scattering technique for an optical waveguide with imperfect boundaries is investigated. The inversion of imperfect boundaries of waveguides is different from usual inverse scattering algorithms because only eigenmodes can propagate in a waveguide. We suggest a parametric inversion method based on the first-order coupled mode equation. The validity and usefulness of the method are demonstrated using examples of periodic and almost periodic imperfect boundaries.

Implementation Of The Sliding Dictionary Data Compression Algorithm On An At&T Dsp 32 Digital Signal Processor, Shu Fang

Theses

Compared with other data compression algorithms, the sliding dictionary is more practical for implementation in hardware. After a comprehensive study of the algorithm, an on line data compression scheme which uses AT&T DSP 32 Digital Signal Processor has been developed. Assembly language programs for encoder and decoder have been written and tested. The whole compression system has been simulated in a real-time environment and results have been obtained. A quantized coding scheme has been used to increase compression ratio. Separate assembly language programs have been written to implement the quantized coding scheme and tested. Results obtained from real-time simulation are …

Adaptive Subband Video Coding With Motion Compensation, Manjunatha S. Kadur

Theses

An efficient scheme for video sequence coding is presented here. The interframe redundancy present in the video sequence is exploited by motion compensation using the efficient independent orthogonal search technique, which requires least number of computations in the worst case. The motion compensated frame difference signal is encoded with a new technique — subband adaptive vector quantization.

Subband coding like transform coding results in signals of unequal energy patterns, which is efficiently utilized for bit allocation. In addition, the subband approach offers means for efficient (but suboptimal) Vector Quantization (VQ) by splitting the signal into several signals. By rate distortion …

Csma/Cd Lan Performance : Modeling And Simulation In Csim, Nanda K. Kanuri

Theses

A simulation model to analyze the performance of CSMA/CD protocol is developed. CSIM, a process oriented simulation language, is used. Performance measures such as throughput, average delay and maximum delay per packet are presented. System performance is analyzed for different loads, number of stations, and network lengths, with packet length as parameter. By trading off the throughput and delay time, an optium packet length can be selected from simulated performance curves. A general scheme for MAC protocols modeling is provided.

Distributed Data Collection System Based On Hierarchical Network, Harish Mahindrakar

Theses

On line data collection systems are essential in order to satisfy the requirements for real-time applications. Typical applications are computer integrated manufacturing, security control, hourly wage calculations, human resources management. On line data collection system makes it possible to closely monitor and track different operations with more accuracy. On line data collection systems eliminate redundant work, save time and money.

If the distributed system is based on a multicontroller, it becomes really cost effective solution to build such a system.

This effort is the design of a distributed multicontroller system, required in a data collection system. A hierarchical network is …

Study Of System Dynamics In A Primitive Identification/Counter Model, Zofia Monczka

Theses

Previous study of a simplified identification/counter model (1), (2) indicated bifurcations in its behavior. In order to study this topic, graphical representations of the system dynamics are derived. Analysis of the structure of the identification/counter model reveals the reasons why the graphs assume their given shapes.

The graphical representations of the system are used to characterize the system behavior and thus explain the bifurcations previously encountered. These representations are then used as the basis for deriving a general model for system behavior. Analysis of this model for a specific range of parameter values indicates that the system contains numerous additional …

A Dynamic Cmos Programmable Logic Device, Praveen K. Parvathala

Theses

This thesis presents the design of a Dynamic CMOS Programmable Logic Device chip (NJIT PLD). The main feature of the chip is its reconfigurability and the provision of a processor interface which makes the chip capable of performing coprocessor like functions. The chip has been designed to be used with the Intel 8088 microprocessor, but may also find stand-alone applications. Provision is made for dynamic reconfiguration of the chip. The configuration data for the chip comes from a serial PROM, to which the chip sends a synchronizing clock. The NJIT PLD is a high performance, digital integrated circuit. It has …

Trellis Encoded M-Ary Partial-Response Schemes, Raju P. Patel

Theses

Convolutional codes can be used to improve the reliability of partial-response channels in a Gaussian noise environment. Multilevel partial-response schemes can be used to increase the information rate in a given bandwidth compared to binary partial-response schemes. This thesis deals with the theory and construction of convolutionally encoded 4-ary partial- response schemes. The free Euclidean distance of the combined trellis of rate-1/2 convolutional code and 4-ary partial-response scheme is found through a computer search, using the Viterbi-algorithm.

The matched encoding concept is also applied to find the conditions for matched encoding for combined trellis of rate-1/2 convolutional codes and 4-ary …

Hardware And Software Designs For Integrated Electrical Engineering Experiments, Sujin Pramojaney

Theses

The interfacing between the various electrical engineering fields is not emphasized in the electrical engineering laboratory; the traditional aim of a laboratory course is to complement a single theoretical course. The object of this thesis is to provide means, hardware and software, that can be used in different system configurations to perform useful industrial tasks. The chosen task is to control electrical machines using a microcomputer. Experiments can be written to be performed by students, using these systems. The emphasize of these experiments can be on the characteristics of a power electronics device, but the device will be seen as …

Nonlinear Variants Of The Least Mean Square (Lms) Algorithm, John George Van Saders

Theses

Adaptive systems using the LMS algorithm or its variants usually require bilinear correlation multipliers with high dynamic range and large bandwidth. The class of LMS-type algorithms utilizing zero-memory limiters (on one or both correlation multiplier inputs) to reduce dynamic range are examined. The correlator limiters may be included to simplify the hardware implementation or may arise due to limited dynamic range of the loop components. The application of a theorem due to Price provides a technique for simplifying the analysis, which otherwise might be extremely difficult. The analysis provides closed-form expressions for the mean weight convergence behavior, permitting examination of …

Image Coding With Quantization Noise Feedback, Jen-Chieh Wang

Theses

A progressive image transmission based on Laplacian pyramid structure is used for the scheme of quantization noise feedback. Because a reduction in transmission rate inevitably increases the quantization noise and degrades the image quality, the designer of the quantizer laces the dilemma of balancing the transmission rate and the quantization noise. Using the algorithm of noise feedback will reduce the quantization error without appreciably increasing transmission rate. Uniform quantization and nonuniform quantization rate used to compare the result with or without noise feedback. In the transmitter end, the original image can be reduced by a certain procedure to form a …

Efficient Design Of Trellis Coded Cpfsk, Hsinhsyong Yang

Theses

In this thesis convolutional coded continuous-phase frequency shift keying (CPFSK) is studied.

CPFSK which is a subclass of continuous phase modulation ( CPM ) is a constant amplitude digital modulation scheme with very good spectral characteristics.

A new method, called matched encoding, that combines convolutional codes and CPFSK is studied. With the maximum free Euclidean distance as the optimization criterion, the best combined coded schemes were found for the same rate of the codes and given total number of states in the optimum decoder.

A different method which was suggested by Rimoldi[7], called the decomposition method, is also described in …

Integrated Voice/Data Services On Multihop Radio Networks, Chialun Lu

Theses

Integrated voice/data is the first step toward Integrated Services Digital Networks (ISDN). Various techniques have been proposed and analyzed for multiplexing voice and data over a large bandwidth channel. Network efficiency is what we are most concerned about. In this paper, we address the problem of determining optimal access policies for a tandem ISDN network by using Dynamic Programming. Computer simulation shows that optimal policy minimizes the data delay.

Rays And Beams Near The Shadow Boundary, Weixing Gao

Theses

Asymptotic techniques in wave propagation problems exhibit attractive features in computational aspects compared to numerical methods. However, they experience problems in complicated environments due to the transition regions associated with them. The transition region involving a shadow boundary in the downward refractive index profile has been investigated. The ray paths were determined solving the ray equation analytically. The Green's function due to a line source is constructed and evaluated using numerical integration (reference solution) and asymptotic techniques. Then, the field produced by the source, is synthesized, using superposition of beams generated from complex source point representation. The floating parameters such …

Reduced Complexity Receivers For Trellis Coded Modulations Via Punctured Trellis Codes, Bonchul Koo

Theses

We introduce a new concept, called matched punctured trellis encoding, that simplifies the complexity of Maximum Likelihood Sequence Estimation (MLSE) receivers for combined trellis encoding and modulations with memory. Matched punctured trellis encoding is applied to tamed frequency modulation (TFM) which is a bandwidth efficient correlative - FM scheme. TFM finds applications in satellite, microwave radio, and mobile communications.

Our approach is based on puncturing a rate - 1/2 matched convolutional code to obtain a rate - 2/3 mismatched code. A matched code is one that produces trellis coded modulations of minimum complexity. Puncturing these codes to obtain mismatched codes …

Computer Controlled Oxygen Bioreactor, Dinesh Sachdeva

Theses

A computer controlled oxygen bioreactor has been designed to obtain more accurate values for oxygen bioreactor requirements and to optimize oxygen productivity with respect to oxygen concentration. This study solves a specific problem relating to oxygen concentration measurement and bioreactor control. The proposed reactor system is computer controlled to insure a desired oxygen concentration supply to an oxygen consuming reaction or organism. This thesis also explains and illustrates the bioreactor and its computer control system.

The computer controller incorporates oxygen sensing, oxygen measurement and temperature measurement, Electrical circuits have been designed and built for proper signal conditioning, data acquisition, data …

A Decomposition Algorithm For The State Transition Lattices Of Certain Non-Product Queueing Protocols, Cheng-Kuo Hong

Theses

In this thesis, an algorithm is developed to decompose the state transition lattice of a class of non-product form queueing models. Here only the lattices of regular type A structure are considered. By applying the algorithm, the lattice is decomposed into solvable subsets which can be solved sequentially.

Trellis Coded Partial-Response Channels, Kumari Yalamanchili

Theses

Binary partial-response channels with transfer functions (1+D), (1-D) and (1 — D²) are intersymbol-interference channels which find applications in baseband data communications and magnetic recording systems (tapes, discs, compact discs, etc.). In this thesis we develop a coding technique for improving the reliability of digital transmission over noisy partial-response channels with the above transfer functions. We do so by introducing a new concept, called matched encoding, which arises in the context of trellis coding for channels with memory.

We use inner and outer trellis coding of the partial-response channels. The inner codes called matched inner coset codes, are …

Methods Of Achieving High Speed On Data Flow Machines, Chung-Lin Yu

Theses

Simultaneity is a key to high speed and large computation. Assuming hardware components of a given speed, it is the only remaining consideration in achieving raw speed. Simultaneity can be shackled by dependencies, however, and years of hardware and software work have been devoted to understanding the types of dependencies and how they can be obeyed or removed from a computation.

In this thesis, the principles of the data flow machines and the structure of the data flow languages are illustrated. Various methods of achieving high speed are discussed. These methods include streams, Lenient Cons (I-structure), and dependence-driven computation, which …

Image Coding With Switchable Dynamic Quantizers, Shing-Sun Yu

Theses

In this thesis, an image coding scheme based upon dynamic quantizations and Differential Pulse Code Miodulalton (DPCM) is studied for application in digital image transmission. This scheme achieves a degree of data compression, and at the same time, maintains an acceptable degraded Mean Square Error (MSE) in the reproduced image. The image to be transmitted is first divided into blocks and each block is then quantized before transmission. The dynamic quantizations including DPCM) are dynamically utilized in reducing the transmission rate while maintaining a desirable quantization noise level. Results of computer simulation of the proposed scheme demonstrate that the switchable …

Algorithms And Analysis Of Dialogue Time Patterns, Muhammad Saleem Akhtar

Theses

The characteristics of a conversation between two people can be determined not only what they say to each other, but also by the rhythm of their dialogue. Their moods, attitudes and attributes may be conveyed in part by time patterns of the speech. The programs described herein can be useful in exploring the relationship between dialogue time patterns and other psychological aspects of human behavior. The measurements of these time patterns can be quickly, inexpensively, and reliably accomplished by the use of a PC.

The programs described herein are capable of calculating the variabilities, the time durations and frequencies of …

Unified Study Of Trellis Encoded Correlative-Fm, K. Narendra Bhat

Theses

Trellis encoding of correlative-FM is studied from the point of view of the overall structure, performance, bandwidth occupancy, and complexity. Here, correlative (partial-response) coding is used to improve the bandwidth efficiency of MSK (Minimum-Shift Keying) at the expense of performance. Then, trellis codes are used to improve the error-rate performance of the corresponding correlative-FM scheme with little or no increase in bandwidth occupancy (compared to uncoded MSK). In a sense, this is similar to Ungerboeck's coding for (memoryless) PSK modulation.

A new encoding approach, called matched encoding, is used here that produces trellis coded modulations achieving greater coding gains than …

Spectrum Analysis Of Heart Rate Variability, Xuming Wang

Theses

This research investigated the respiratory influence on heart rate variability. The heart rate spectrum was obtained from the heart rate variability signal which was extracted from the electrocardiogram We then studied the magnitude of the respiration peak in the heart rate spectrum and examined the dependence of the magnitude of this peak on breathing frequency, subject and day.

An experiment designed to simultaneously acquire the electrocardiogram and the respiration signal was conducted on ten people ranging from 22 to 55 years of age (3 females, 7 males). Three tests were performed on each subject. In each test, 2 different postures …

Statistical Multiplexing Of Video Sources For Packet Switching Networks, Krishna Kancherla

Theses

Communication networks are fast evolving towards truly integrated networks handling all types of traffic. They employ integrated switching technologies for voice. video and data. Statistical or asynchronous time division multiplexing of full motion video sources is an initial step towards packetized video networks. The main goal is to utilize the common communication channel efficiently, without loosing quality at the receiver. This work discusses the concept of using statistical multiplexing for packet video communications. The topology of a single internal packet network to support ISDN services has been adopted. Simulations have been carried out to demonstrate the statistical smoothing effect of …