Coupling optics for a monolithic array LD module has been developed. High efficient and small sized confocal optics with aspheric lenses based on hyperbolic surfaces has been designed to achieve the uniformity of coupling loss. A small sized (7.2 cc) 4 channel array LD module with this optics was fabricated. This module has low (4.2 dB) and uniform (0.5 dB) coupling loss.

In this paper we propose a formal linearization method which permits us to transform nonlinear systems into linear systems by means of the Chebyshev interpolation. Nonlinear systems are usually represented by nonlinear differential equations. We introduce a linearizing function that consists of a sequence of the Chebyshev polynomials. The nonlinear equations are approximated by the method of Chebyshev interpolation and linearized with respect to the linearizing function. The excellent characteristics of this method are as follows: high accuracy of the approximation, convenient design, simple operation, easy usage of computer, etc. The coefficients of the resulting linear system are obtained by recurrence formula. The paper also have error bounds of this linearization which show that the accuracy of the approximation by the linearization increases as the order of the Chebyshev polynomials increases. A nonlinear filter is synthesized as an application of this method. Numerical computer experiments show that the proposed method is able to linearize a given nonlinear system properly.

This letter presents the results of an analysis concerning the global, dynamical structure of a second order phase–locked loop (PLL) in the presence of the continuous wave (CW) interference. The invariant manifolds of the PLL equation are focused and analyzed as to how they are extended from the hyperbolic periodic orbits. Using the Melnikov integral which evaluates the distance between the stable manifolds and the unstable manifolds, the transversal intersection of these manifolds is proven to occur under some conditions on the power of the interference and the angular frequency difference between the signal and the interference. Numerical computations were performed to confirm the transversal intersection of the system–generated invariant manifolds for a practical set of parameters.

The analytic structure of the governing equation for a 2nd order Phase–Locked Loops (PLL) is studied in the complex time plane. By a local reduction of the PLL equation to the Ricatti equation, the PLL equation is analytically shown to have singularities which form a fractal structure in the complex time plane. Such a fractal structure of complex time singularities is known to be characteristic for nonintegrable, especially chaotic systems. On the other hand, a direct numerical detection of the complex time singularities is performed to verify the fractal structure. The numerical results show the reality of complex time singularities and the fractal structure of singularities on a curve.

In this paper, a novel coupler-type wavelength demultiplexer composed of the K+-ion diffused waveguides, which has an adjustment function for optimizing the diffusion depth, is proposed to achieve reliably the high extinction ratio. The optimization in the diffusion depth is made by repeating the K+-ion diffusion and extinction-ratio measurement alternatively, and the high extinction ratios more than 20 dB are measured reliably at both operation wavelengths of 0.6328 and 0.83 µm. Experimental results on the polarization dependence in the extinction-ratio adjustment are also reported.

We present a novel algorithm to reconstruct the refractive-index profile of a circularly symmetric object from measurements of the electromagnetic field scattered when the object is illuminated by a plane wave. The reconstruction algorithm is besed on an iterative procedure of matching the scattered field calculated from a certain refractive-index distribution with the measured scattered field on the boundary of the object. In order to estimate the convergence of the reconstruction, the mean square error between the calculated and measured scattered fields is introduced. It is shown through reconstructing several examples of lossy dielectric cylinders that the algorithm is quite stable and is applicable to high-contrasty models in situations where the Born approximation is not valid.

For the purpose of detecting the intracranial vascular deformations noninvasively, transducer for bruit sound emanated from diseased lesion and analyzing system were developed and applied clinically. Several aspects of the bruit signals were clarified and the possibility of early diagnosis was increased.

A MOS VCO which has improved linearity of oscillation frequency versus control voltage and has no 1/2 divider is studied. The improved VCO characteristic has been obtained by the use of only two additional transistors, one of which has a role of a load and another of which has a role of a control current source in a differential type delay cell.

We prove that the maximum likelihood estimator for estimating 3-D motion from noisy optical flow is not optimal", i.e., there is an unbiased estimator whose covariance matrix is smaller than that of the maximum likelihood estimator when a Gaussian noise distribution is assumed for a sufficiently large number of observed points. Since Gaussian assumption for the noise is given, the maximum likelihood estimator minimizes the mean square error of the observed optical flow. Though the maximum likehood estimator's covariance matrix usually reaches the Cramér-Rao lower bound in many statistical problems when the number of observed points is infinitely large, we show that the maximum likelihood estimator's covariance matrix does not reach the Cramér-Rao lower bound for the estimation of 3-D motion from noisy optical flow under such conditions. We formulate a superior estimator, whose covariance matrix is smaller than that of the maximum likelihood estimator, when the variance of the Gaussian noise is not very small.

A three-dimensional analysis of Whispering-Gallery modes (W. G. modes) in a coaxial dielectric resonator is proposed and presented. The coaxial dielectric resonator is constructed from a lossy dielectric disk and ring which have diameters of several tens times as large as wavelength. Eigenvalue equations of the W. G. modes are derived rigorously from field expressions and boundary conditions. The resonant frequencies, unloaded Q values and field distributions are calculated numerically from the eigenvalue equations. These calculated results are in good agreement with experimental ones for an X band model. As a result, it is shown that a considerable quantity of modal energy can be confined in a loss-less gap between the disk and ring, and then the unloaded Q value is higher than that of a conventional dielectric disk and ring resonator.

We propose and demonstrate an excellent linearly frequency-swept laser diode (LD) for sensing system utilizing frequency-moduleted continuous-wave (FMCW) technique. In order to linearly sweep the optical frequency, we adopt a reference interferometer and an electric phase comparator. The interference beat signal of the reference interferometer is phase-compared with an external reference rectangular signal having a fixed frequency near the interference beat signal frequency by a lock-in amplifier. The error signal from the lock-in amplifier is fed back to the modulating signal of the injection current of the LD. Thus, a phase-locked loop composed of optical and electric circuits can be established, and the beat signal frequency is locked to the frequency of the reference signal. The optical frequency of the LD is, therefore, excellently linearly swept in time. In order to experimentally confirm the linearlity of the proposed method, we apply this light source to the FMCW reflectometry. Resultingly, the improvement of the linearity is estimated to be about 10 dB. And the theoretically limited spatial resolution of the FMCW reflectometry is achieved.

Telecommunication services are becoming more and more personalized, integrated, and refined. Advanced personal and mobile telecommunication services, intelligent networks, and network management operations require cooperative network-wide distributed processing on a very large scale. Telecommunication programs must support these services with great flexibility, efficiency, and reliability. This paper proposes a new call processing model that improves the availability and flexibility of telecommunication programs. It first points out requirements, outlines the distributed processing platform PLATINA, and discusses several approaches to the enhanced call processing model. Then it explains the call processing program structure, and gives illustrations of mobile and multi- party service control as typical examples. The Caller-Callee decomposition reduces the complexity of the call processing program and enhances the call model; the separation of call and bearer enhances service flexibility and integration; distributed object-oriented techniques meet software evolution requirements. A prototype program has been implemented and has proved the effectiveness of this approach.

Fractal image coding using iterated transformations compresses image data by exploiting the self–similarity of an image. Its compression performance has already been discussed in [2] and several other papers. However the relation between the performance and the self–similarity remains unclear. In this paper, we evaluate fractal coding from the perspective of this relationship.

In this letter we present an electronic circuit based on a neural net to compute the discrete Walsh transform. We show both analytically and by simulation that the circuit is guaranteed to settle into the correct values.

Recently actual use of the OSI standardized protocols has begun on client-server systems of LANs, and reduction of OSI protocol overheads in high-speed networks has become more important. We studied a parallel-processing architecture for Message Handling System (MHS), which requires a large amount of protocol processing and is expected to be used widely. We implemented a prototype MHS server with performance scalable to number of CPUs, by porting an existing MHS software with minimum modification. This paper reports on the parallel processing scheme, hardware and software architecture of the prototype, as well as evaluation of the scheme based on measurement and simulation.

A stepwise refinement method of communications service specifications is proposed to generate communications software that can conform to any network architecture. This method uses a two-layered language; one layer is a service specification description language (STR), and the other layer is a supplementary specification description language for implementing STR description on a communications system (STR/D). STR specifies terminal behaviors that can be recognized from a perspective outside of the communications systems. With STR, a communications service is defined by a set of rules that can be described without detailed knowledge of communications systems or communications network architectures. Each STR rule describes a global state transition of terminals. Supplementary specifications, such as terminal control and network control, are needed to implement communications services specified by STR rules. These supplementary specifications are described by STR/D rules. Communications services, such as UPT (Universal Personal Telecommunication), are standardized so that they can be provided on a given functional model consisting of functional entities. Specifications for each functional entity in a network are obtained from the two kinds of initially described specifications mentioned above. The obtained specifications are described by STR(L) and STR/D(L) rules, which specify local specifications of a functional entity. These specifications for functional entities are then transformed into software specifications, and finally communications software is generated from these software specifications. This stepwise refinement method makes it possible to generate communications software that can conform to any functional model from service specifications.

By applying Wigner distribution, which has high time resolution and high random noise reducing capability, to the acoustic bio–signals, the possibility of early diagnosis in both intracranial vascular deformation and prosthetic cardiac valve malfunction increased. Especially in latter case, 1st–order local moment of the distribution showed its effectiveness.

In this letter we propose an interpolation technique for low-quality fingerprint images for highly reliable feature extraction. To improve the feature extraction rate, we extract fingerprint features by referring to both the interpolated image obtained by using a directional Laplacian filter and the high-contrast image obtained by using histogram equalization. Experimental results show the applicability of our method.

A novel optoelectronic mesoscopic neural device is proposed. This device operates in a neural manner, involving the electron interference and the laser threshold characteristics. The optical output is a 2–dimensional image, and can also be colored, if the light emitting elements are fabricated to form the picture elements in 3–colors, i.e. R, G, and B. The electron waveguiding in the proposed device is analyzed, on the basis of the analogy between the Schrödinger's equation and the Maxwell's wave equation. The nonlinear neural connection is achieved, as a result of the superposition an the interferences among electron waves transported through different waveguides. The sizes of the critical elements of this device are estimated to be within the reach of the present day technology. This device exceeds the conventional VLSI neurochips by many orders of magnitude, in the number of neurons per unit area, as well as in the speed of operation.

In general, message passing multiprocessors suffer from communication overhead and shared memory multiprocessors suffer from memory contention. Also, data I/O overhead limits performance. In particular, computer vision tasks that require massive computation are strongly affected by these disadvantages. This paper proposes new parallel architectures for computer vision, a Flexibly (Tightly/Loosely) Coupled Multiprocessor (FCM) and a Flexibly Coupled Hypercube Multiprocessor (FCHM) to alleviate these problems. FCM and FCHM have a variable address space memory in which a set of neighboring memory modules can be merged into a shared memory by a dynamically partitionable topology. FCM and FCHM are based on two different topologies: reconfigurable bus and hypercube. The proposed architectures are quantitatively analyzed using computational models and parallel vision algorithms are simulated on FCM and FCHM using the Intel's Personal SuperComputer (iPSC), a hypercube multiprocessor, showing significant performance improvements over that of iPSC.