This thesis introduces a method by which low-frequency analog waveforms may be generated using delta-sigma modulation. The technique centers around a delta-sigma based oscillator which, with the exception of a continuous-time low-pass filter, is entirely digital and provides precise control over the oscillation frequency, amplitude, and phase. The incorporation of a delta-sigma modulator inside the resonator loop leads to an efficient implementation requiring 4 multi-bit adders, 4 delay elements, and a 2-input multiplexer. Two additional circuits, which generate multi-tone and piece-wise linear waveforms, are presented as extensions of the original single-tone design. / Prototypes of the proposed designs have been assembled using Field-Programmable Gate Array, and BiCMOS technologies. The test results have successfully verified the validity of the proposed concepts indicating dynamic ranges exceeding 80 dB and 60 dB for the single and multi-tone generators respectively.

This thesis focuses on the routing, power control and channel allocation of Packet Radio Network (PRnet) with multiple channels. The Nearest Forward with Progress (NFP) routing policy is known to achieve higher throughput in the single channel setting than the Most Forward with Progress (MFP) strategy. This observation motivates our work on the variant of the Shacham/King model in which MFR is replaced by NFP in a multi-channel case. The resulting network shows that the superiority of NFP relative to MFP persists. Another way to improve network performance is to control the network topology by restricting the set of possible paths. Pairwise Channel Assignment (PCA), which assigns the different channels to the restricted paths, is modified so as to expand its applicability and reduce its computational burden without significant loss of throughput. The performance evaluations are based on a combination of analysis and simulation.

A key problem in the interpretation of visual form is the partitioning of a shape into principal components that correspond to the parts of an object. This thesis presents a method for partitioning a set of surface estimates obtained with a laser range finding system into subsets corresponding to such parts. The strategy employed makes use of two complementary representations for surfaces: one that describes local structures in terms of differential properties (e.g. edges, lines, contours) and the other that represents the surface as a collection of smooth patches at different scales. It is shown that by enforcing a consistent interpretation between these two representations, it is possible to derive a partitioning algorithm that is both efficient and robust. Examples of its performance on a set of range images are presented.

A decision feedback (DFB) equalizer is used to correct for the effects of inter-symbol interference in digital communications systems. The order of the DFB filter is reduced to a bare minimum and studied when it is insufficient to equalize the channel, i.e., when the filter does not have enough poles to cancel all the zeroes of the channel. The error surfaces produced by the DFB filter in the symbol-by-symbol, frame-by-frame, and aggregate sense are investigated. A genetic algorithm is then applied to the problem of adapting the DFB filter coefficients. The performance of the genetic algorithm is compared to that of the conventional gradient search algorithm for both the sufficient and insufficient order cases with varying levels of noise. It is found that the genetic algorithm outperforms the gradient algorithm in the insufficient cases.

This thesis presents a new approach to fault diagnosis in telecommunications. The principles of fuzzy theory and fuzzy expert systems will develop a framework for the diagnosis and analysis of results produced by the circuit measurement instruments of the telephone lines. The thesis is a case study in which a fault is isolated and then, membership functions are empirically produced to represent fault parameter behavior. Fuzzy system rules are then constructed and the whole module system is tuned and put to testing. The results of the tests are thoroughly discussed, concluding with recommendations for extending the case module study in for the purpose of constructing a complete fuzzy diagnostic system.

This thesis deals with the use of planar near-field measurements for characterizing and modeling general electromagnetic radiators, particularly non-directive ones. / Relationships between the measured field on a plane and the equivalent currents on the radiator plane were established. These were used to determine the resolution with which these currents can be mapped. The equivalent currents then serve as a model from which the radiator external field can be calculated. It was shown that significant filtering of measurement errors results from the utilization of such a model. The consequent reduction in errors was estimated. / The behavior of the measurement probe was studied in depth. The two general probe types, wire and aperture, were analyzed, and field extraction formulation and procedure were established. / The techniques developed were applied to the experimental study of three different radiators leading to new information about their behavior. Results show close agreement between the calculated field values from the model and those measured directly.

Testing VLSI circuits is a complex task that requires enormous amounts of resources. To decrease testing costs, testing issues are considered earlier in the design process. This is known as "design for testability" (DFT). Built-in Self Test (BIST) is one proposed DFT approach. BIST generally consists of incorporating additional circuitry on the chip to generate test patterns and compact the response of the circuit under test (CUT) into a reference signature. Compaction implies an information loss, introducing the possibility that a faulty circuit declares itself as good. Such errors are known as aliasing errors. Several BIST schemes have been proposed, and each have a particular performance in regard to aliasing. However, the schemes are often evaluated and compared with ill-defined measures for which the underlying assumptions are either not stated or understood clearly. Here, a novel classification for the measures of aliasing is proposed. By providing clear definitions of different possible measures, the proposed classification augments the understanding of the aliasing problem. / This dissertation focuses on the popular BIST scheme that consists of applying pseudorandom test patterns to a CUT and compacting the latter's response by a signature analysis register (LFSR). Assessing the quality of such a scheme in regard to fault coverage is crucial. Fault coverage can be established by full fault simulation. However, high costs may preclude this approach. Other techniques, probabilistic in nature, have been proposed, but a lack of computationally feasible techniques for analyzing the aliasing problem under a reasonable model has left them elusive. Here, new and computationally feasible techniques are developed. More specifically, closed-form expressions for the probability of aliasing are derived for a certain type of LFSRs. Upper bounds are derived for LFSRs characterized by primitive polynomials. An iterative technique is developed for computing the exact probability of aliasing for LFSRs characterized by any feedback polynomial, and for any test sequence length. These new techniques enable better assessments of the quality of BIST schemes that use signature analysis for response compaction. In turn, they are useful for making important design decisions, e.g., determining the number of test patterns that should be applied to a CUT to achieve a certain test confidence; alternatively, deciding how long the signature analyzer should be, and what type of feedback it should possess to achieve a certain desired test confidence. / The techniques developed for computing the probability of aliasing in BIST are also useful in the context of coding theory. The iterative technique developed for computing the probability of aliasing may be used as an efficient technique for computing the probability of an undetected error for shortened versions of cyclic codes.

With an ever-decreasing minimum feature size, integrated circuits have more transistors, run faster and are larger than ever before. As a result, problems such as heat dissipation, clock generation and clock distribution are at the forefront of challenges facing chip designers today. A Globally Asynchronous, Locally Synchronous (GALS) system combined with dynamic voltage and frequency scaling is an architecture that can combat many of these issues while allowing for high performance operation. In this thesis, we investigate three distinct circuit designs compatible with, but not limited to, such a system. The first uses a novel bi-directional asynchronous FIFO to communicate between independently-clocked synchronous blocks. The second is an All-Digital Dynamic Clock Generator designed to glitchlessly switch between frequencies with very low latency. The third is a Digitally-Controlled Oscillator that can either be used stand-alone or as part of an all-digital PLL (ADPLL) to generate the global fixed frequency clocks required by the All-Digital Dynamic Clock Generator. These designs have been designed, simulated and shown to perform all the tasks required to implement a Globally Asynchronous, Locally Dynamic System (GALDS) in either a traditional ASIC design or a newer System-on-Chip (SoC).

Recently, Huang and Carrington introduced a preconditioned inexact spectral transform (PIST) method for calculating energy levels and wave functions in quantum mechanics. Mathematically, the PIST method is a new iterative technique for solving eigenvalue problems. It gives both eigenvalues and eigenfunctions within a prescribed eigenvalue spectrum. The method requires storing only a small number of vectors but allows one to calculate eigenvalues and wavefunctions with far fewer matrix-vector products than the conventional Lanczos method used for solving eigenvalue problems. We present here a numerical solution of the scalar Helmholtz wave equation using the PIST technique in solving the associated eigenvalue problem. / In this work, we first convert the scalar Helmholtz wave equation to an eigenvalue problem: we use the basis function expansion method, i.e. by expanding the unknown electrical field function in a complete set of orthogonal basis functions, the differential equation can be converted to a matrix eigenvalue equation. The propagation constants and mode field distributions can be obtained from the solution of the corresponding matrix eigenvalue problem. / The method can handle any type of eigenvalue spectrum, in particular if the eigenvalue spectrum contains only one guided mode. Furthermore the algorithm can give distinct orthogonal eigenfunctions associated with degenerate propagation constants. / We have written a user-friendly and modularized computer program. The program can handle arbitrary index profiles and geometrical shapes by merely changing the input files that describe the index profile. The method proves to be efficient, versatile, and requires less memory than popular finite difference and finite element methods. / We show examples of its application for optical waveguide modeling in finding the propagation constants and mode field distributions of various waveguides and couplers. For couplers, the method can give the global evolution of the electrical field.

Vertical cavity surface emitting lasers (VCSELs) are increasingly being used in various photonic applications. In order to design an optical system which will transmit the light emitted by a VCSEL it is necessary to model the intensity profile. The purpose of this thesis is to apply a method of extracting the modal content from an intensity profile assumed to be composed of Hermite-Gauss modes. This will be done for both the simulated output of a VCSEL and for experimentally measured intensity profiles. It will be demonstrated that the method will produce an accurate model for a VCSEL output which is close to being ideally Hermite-Gauss. Two experimental setups used to measure the intensity profiles will be presented. The first uses a scanning near-field optical microscope (SNOM) to measure the intensity near the surface of the VCSEL. In the second setup the intensity is measured at the output of a two-lens system used to image the beam waist.