A monomial-based method for solving systems of non-linear algebraic equations

A monomial-based method for solving systems of algebraic non-linear equations is presented. The method uses the arithmetic-geometric mean inequality to construct a system of monomial equations that approximates the system of non-linear equations. A change of variables transforms the monomial system into a system of linear equations, which is readily solved. Special properties of the monomial method are identified and their significance is discussed. Invariance properties of the monomial method produce a built-in, self-adjusting scaling of the variables and equilibration of the equations of the linear system. Other special properties can lead to useful bounds on, and invariances of, the conditioning of the linear system. An invariance to uniform scaling is responsible for extremely rapid convergence to the equation surfaces in the initial iterations. An invariance to multiplication of the algebraic equations by a certain class of functions leads to a useful insensitivity to form of the algebraic system. Insensitivity of the monomial method to solutions with negative components avoids meaningless solutiuons of the algebraic system that appear as undesirable by-products of the formulation. A practical engineering design problem is solved to demonstrate the special properties of the monomial method.     

Use of a doubly polarized laser for remote powering of anintegrated optical modulator

Stable operation of an integrated optic modulator is demonstrated using a 1.3 μm doubly polarized laser as a depolarized source in conjunction with a long run of ordinary fiber. The laser is found to be unusually susceptible to feedback due to gain competition between the polarization modes. The resulting low-frequency polarization noise is significantly reduced by the addition of fiber isolators to the system     

Using neural networks to select wavelet features for breast cancerdiagnosis

This study focuses on improving microcalcification classification by establishing an efficient computer-aided diagnosis system that extracts Daubechies-4 and biorthogonal wavelet features. These wavelets were chosen because they have been used in military target recognition and fingerprint recognition research with images characterized by low contrast, similar to mammography. Feature selection techniques are employed to further increase classification performance. The artificial neural network feature selection techniques are complemented by a conventional decision boundary-based feature selection method. The results using the wavelet features are compared to more conventional measures of image texture, angular second moment, and Karhunen Loeve coefficients. The use of alternative signal processing to compare wavelet and neural techniques allows for a measure of the problem difficulty. It is concluded that advances and contributions have been made with the introduction of two novel feature extraction methods for breast cancer diagnosis, wavelets and eigenmasses. Additionally, feature selection techniques are demonstrated, compared, and validated, transforming adequate discrimination power into promising classification results     

Growth of part-through thickness fatigue cracks in sheet polymethylmethacrylate

Fatigue growth data has been obtained for part-through thickness cracks in sheet specimens of polymethylmethacrylate subjected to cyclic tension. Statistical techniques are used to examine the usefulness of linear elastic fracture mechanics for correlating this part-through thickness crack growth data with throughthickness data published previously.     

Euclidean controllable realizations of linear hereditary systems

A realization problem for a class of linear hereditary systems is formulated and state space representations are constructed from inputoutput relations. The resulting hereditary models are shown to be Euclidean controllable, and comparisons are made between these hereditary realizations and other recently developed theories.This research was supported in part by the Air Force Flight Dynamics Laboratory under grant AFOSR-77-3221.     

Picosecond pulse generation with ultralow jitter in 1.5-μmmulticontact MQW lasers using Q-switching

High-peak-power picosecond pulses with ultralow jitter of 65 fs (50-500 Hz) have been generated by Q-switching multi-quantum-well (MQW) distributed-feedback (DFB) and Fabry-Perot (FP) multicontact lasers. Peak pulse powers of 80 mW have been measured for pulses having durations of 26 ps and repetition rates of 1.25 GHz     

Placement and routing tools for the Triptych FPGA

Field-programmable gate arrays (FPGAs) are becoming an increasingly important implementation medium for digital logic. One of the most important keys to using FPGAs effectively is a complete, automated software system for mapping onto the FPGA architecture. Unfortunately, many of the tools necessary require different techniques than traditional circuit implementation options, and these techniques are often developed specifically for only a single FPGA architecture. In this paper we describe automatic mapping tools for Triptych, an FPGA architecture with improved logic density and performance over commercial FPGAs. These tools include a simulated-annealing placement algorithm that handles the routability issues of fine-grained FPGAs, and an architecture-adaptive routing algorithm that can easily be retargeted to other FPGAs. We also describe extensions to these algorithms for mapping asynchronous circuits to Montage, the first FPGA architecture to completely support asynchronous and synchronous interface applications