An equilateral triangular reflector is inserted into an E-plane symmetrical T-junction in order to improve the wave transmission characteristics from the truncated-guide to the through-guide. The reflector is installed in contact with the wall of the through-guide facing the junction plane. In the case that the truncated-guide width b is equal to the through-guide width c, we find that the larger the top angle of the triangle is made, the better the wave transmission becomes. In particular, we can attain a remarkable improvement in obtuse triangles of the top angle θ100 and the height h0.6b. For instance, in the triangle of θ120 and h0.6b, at least 99.5 percent of the energy fed from the truncated-guide transmits evenly into the other two arms, and 0.5 percent at most reflects over the frequency range of 0.31b/λg0.37. The λg is the guided wavelength.

We have studied the oxidation state of Al/Hf and Ta/Hf bi-layered films by AES/XPS analysis. The anion transport process through the upper oxidized layer which is oxygen supply medium for Hf oxidation is examined. It is found that the oxidation reaction of Hf occurring with sufficient anion transport through the upper oxidized layer corresponds to the excellent properties as capacitors.

This paper presents a methos of 3D measurement using a newly developed device that acquires 3D range data and surface color data at the same time. Roughly speaking, for the recognition and synthesis of various objects, two kinds of data are used: 3D shape (range) data and texture (color) data. Usually, these two data types are measured separately, using different kinds of acquisition systems. Typically, range finders are used to measure the 3D shape of objects, and color television cameras are popular to acquire the surface colore. Due to the delay between the two measurements, it is sometimes difficult to match the two data sets, especially in the case of changeable objects like human faces. Moreover, camera angles and lighting conditions may differ in each measurement. Though adjustment may be possible to some extent, acquiring fully synchronized, consistent data for 3D objects having various shapes and colors by existing methods is practically impossible. The authors soleve the problem by combining existing stable technologies to measure the shape and surface color of objects a the same time, resulting in the first cylindrical scanner in the world that acquires 3D range data and color data in synchronization. In the proposed method, the cylindrical range data is measured by a laser light source and a CCD sensor with a resolution of 512 vertical scan lines, 256 points per scan line. The color data is acquired as a cylindrical projection imge having 512 by 256 pixels, 24 bits/pixel (8 bits each for red, green, and blue). The scanner is successfully applied to the measurement of various 3D objects including human heads.

In a cellular mobile system, assigning a channel for a call in a cell so as to achieve high spectral efficiency is an important problem. In usual channel assignment for a cellular mobile system, a channel can be simultaneously assigned to some cells with a constant separation distance. This usual model of a cellular mobile system has been formulated using a graph and it is known that the channel assignment problem is equivalent to the coloring problem of graph theory. Recently, a new channel assignment scheme has been proposed. This scheme takes the degree of interference into consideration. In the scheme, a channel is simultaneously assigned if the CIR (carrier-to-interference ratio) is more than the desired value. In this paper, we formulate this new model using a network and a new coloring problem of networks. The new coloring problem of networks is a generalization of the usual coloring problem of graphs. One of the merits of this formulation is that the degree of cochannel interference between cells can be represented. In the usual formulation using a graph, the degree of cochannel interference between cells can not be represented. Therefore, spectral efficiency in the formulation using a network is higher than spectral efficiency in the formulation using a graph. In this paper, we show that the new coloring problem is an NP-hard problem. Subsequently, we rewrite the new coloring problem of networks to a coloring problem of graphs on some assumptions and consider the relation between the results on the new coloring and the results on the usual coloring.

The radiation pattern of a car antenna is evaluated with the power received by its antenna in mobile communication environments. In the evaluation, the horizontal radiation pattern is assumed to be omnidirectional; the vertical one is expressed in terms of the main beam direction θ1 and the half-power beamwidth θBW. The received power is calculated from such a radiation pattern and an angular probability density distribution of wave arrival. The distribution is estimated using a propagation model to simulate radio propagation paths and surroundings of them in urban areas. The tendency of the estimated distribution is in good agreement with that of the measured one. It is clarified from the estimated results that the angular probability density distribution of wave arrival strongly depends on the street angle ψ. Since the received power calculated from the distribution also depends on ψ, the minimum value (minimum received power) in the received power for ψ=5, 10, 15, , 85 is used to evaluate the radiation pattern. From the view that the radiation pattern for which the minimum received power is highest is suitable for mobile communications, it is revealed that θ1 and θBW in the suitable radiation pattern of the car antenna are roughly 20 and 15, respectively.

This paper proposes a new technology mapping procedure for minimizing the area and/or delay of complex logic gates using the so-called Path-Sensitized Covering Graph (PSCG) constructed from the given Boolean network and library cells. In the earlier tree covering algorithms, the original DAG (Directed Acyclic Graph) covering problem is only approximately solved by reducing the problem size and the search space based on the partitioning of the whole circuit at the multiple fan-out node, which makes the resultant solution locally optimal. Compared to this, our algorithm is more general in that it does not arbitrarily partition the DAG into a set of trees. This is necessary not to exclude the possibility of mapping into some library cell a portion of a given circuit across branching points (fan-out nodes). Three different problems, i.e., area minimization, delay minimization considering the capacitive loading, and area minimization under timing constraint are formulated as a nonlinear 0-1 integer programming problem for which a recently developed MFA (Mean Field Annealing) algorithm was successfully applied. Results on the performance of this procedure using a set of MCNC Benchmark examples have shown a notable improvement compared to the earlier tree covering algorithm for the case of area minimization.

VLSI technology has matured to the extent that hundreds of thousands or even millions of transistors can be integrated in a single chip. Layout data for VLSIs consist of around ten mask layers, and the total number of polygons on the layers is about ten times larger than the number of circuit elements. In order to deal with such a large number of polygons, algorithms for mask pattern processing are usually based on the underlying assumption that the whole data is accommodated in the secondary storage and that only those patterns within a few consecutive thin slits of the plane can be processed in the main memory (work list). This is called the work-list method. A linear time geometrical design rule verification algorithm is presented in this paper. This algorithm is based on the work-list method. Content Addressable Memory (CAM) is introduced to implement the work list, so as to make the algorithm run in linear time with linear memory space. Data in RAM can be accessed only by its address, whereas data in CAM is accessed not only by address but also by a content which matches a given referential data. This function is called equivalence search. An equivalence search is executed in constant time independent of the amount of data. The advantages of CAM for our algorithm are summarized as follows. (1) It provides the flexibility to deal with a variety of geometrical search problems for VLSI design. (2) Each geometrical search is done in constant time. (3) Complicated coding for sophisticated data structures depending on problems is not necessary, unlike software implementation.

This study represents the fault tolerance analysis and recovery of multilayer perceptron model used for the pattern recognition. In this paper, we have investigated the faults that may occur in hardware implementations of the model. The effect of two kinds of faults on network performance is simulated and analyzed by modeling the faults and the neural network in software. We will consider two different types of faults-1. stuck-at-faults, 2. faults due to damaged connections between neurons. In case of stuck-at-faults, we considered three kinds of fault-1. stuck-at-0, 2. stuck-at-0.5, and 3. stuck-at-1. In case of faults due to damaged connections between neurons, we considered two kinds of faults-1. reduced connection weights, 2. zero connection weights. We have investigated the output layer neurons' output affected by faults. We found that the output is related with the connection weights with positive sign and those with negative sign. And we found that the damaged neuron can be recovered by magnifying both connection weights with positive sign and those with negative sign.

The one-dimensional map represented by the Bernoulli shift is a candidate of true random number generators. However, no computer with finite-word-length can precisely realize the Bernoulli shift. The M-sequence is shown to be one of finite-word-length approximations to the Bernoulli shift in the sense that the many-to-many map generated by M-sequence is the same as the finite-word-length realization of the iterated Bernoulli shift.

We present a method that uses optical flow to estimate facial muscle actions which can then be recognized as facial expressions. Facial expressions are the result of facial muscle actions which are triggered by the nerve impulses generated by emotions. The muscle actions cause the movement and deformation of facial skin and facial features such as eyes, mouth and nose. Since facial skin has the texture of a fine-grained organ, which helps in extracting the optical flow, we can extract muscle actions from external appearance. We are thus able to construct a facial expression recognition system based on optical flow data. We investigate the recogniton method in two ways. First, the optical-flow fields of skin movement is evaluated in muscle winsows, each of which defines one primary direction of muscle contraction to correctly extract muscle movement. Second, a fifteen dimensional feature vector is used to represent the most active points in terms of the flow variance through time and local spatial areas. The expression recognition system uses the feature vector to categorize the image sequences into several classes of facial expression. Preliminary experiments indicate an accuracy of approximately 80% when recognizing four types expressions: happiness, anger, disgust, and surprise.

A new system for high speed and continuous 3-D measurement is presented. It is based on slit-ray projection method. Remarkable features of our system--high speed and continuous measurement--mainly come from its image plane constructed by PSD array which is horizontally non-divided and linear, whereas vertically divided in numbers. Each row PSD element is attached to respective analog signal processor, A/D converter and memory element. By the virtue of this configuration of the image plane, we can store the positional information of slit-like image on the image plane into memory elements in real time, without waiting for one frame interval as is required in the conventional methods with scanning type image grabbers. As the result, we can scan the slit-ray at a constant high angular velocity and calculate the deflection angle of the slit-ray from scanning time of it, which is set on the address bus of memory elements in our system. Thus, basic datum for 3-D measurement are acquired during only one scanning of slit-ray at high speed in the form of addresses of memory elements and datum stored in them. Moreover, if we use large capacity and/or dual port memories, we may continuously obtain 3-D datum of as many scenes as we want. Trial system has verified our method showing data acquisition time per scene within a few milliseconds and enabling us to obtain 3-D datum continuously at a rate of hundreds scenes per second.

A new circularly polarized microstrip antenna fed by a cross slot using an aperture coupled feed is proposed. The proposed antenna has no 90 degree hybrid coupler for circular polarization and allows easy connection to microwave devices for electronical beam scanning. In this paper, the conditions for circular polarization in this new antenna are derived, and the measured results are also described to verify the authors' proposed conditions for circular polarization. Good return loss and axial ratio characteristics have been obtained.

A successive iteration algorithm for estimating the response of a parallel two-wire transmission line terminated by nonlinear protective devices in an external electromagnetic pulse is developed and implemented for numerical simulation. The validity of the presented algorithm is assured by showing that the numerical result for a particular nonlinear impedance characteristics of the protective devices approaches the rigorous solution which is also derived here analytically.

Short backfire (SBF) antenna developed by H.W. Ehrenspeck is experimentally well-known to have high efficiency radiation characteristics in spite of its small size so far. Therefore this antenna is expected to be one of the candidates of single antennas and an element of array antennas for several types of communications. For example, this antenna has been proposed for shipborne antenna of INMARSAT-M system for maritime satellite communications. Although many efforts to analize and/or optimize this antenna has been made mainly experimentally, rigorous design procedure has not been almost established even now because of its complex configuration. In 1980, an approach to analyze this antenna was tried by Kawakami, H. and et al. which was the first attempt to obtain the theoretical characteristics by more rigorous process, however, from the view points of practical designing, the calculated values in their paper were insufficient because of the lack of strict numerical consideration. Upon the above circumferences, analysis of SBF antenna is performed in this paper by using method of moment with wire-grid model from the view points of establishing the design technique. This paper presents the analytical results of the radiation characteristics of this antenna and also the comparison with the experimental results. From this study, it can be noticed that the radiation characteristics of this antenna can be almost exactly analyzed and also there exists the good agreements between the theoretical and experimental results.

The objective of the Intelligent Network (IN) concept is to rapidly and flexibly add new services without upgrading the complicated switching systems. To realize that concept in the real telecommunication system, the systems must have an environment for the creation of new services and an architecture for execution of those services. This paper proposes an environment for service creation in advanced telecommunication systems such as the IN. The service creation environment provides means to verify the new service specifications and to create executable specification. To facilitate the verification as well as to model the system resources, the service specifications are formally described by means of High-Level nets, Coloured Petri nets (CP-nets). Specifying services in CP-nets gives the capability for modeling many kinds of systems resources over the network. Especially colour plays effective roles to define both the executer of services and resourses clearly. Since IN architecture allows customers to define their own services, it is difficult to verify the services defined by different customers. Therefore, it is important to define the customers services formally at the specification level. The verification method in this paper is based on the pure net analysis technique. It is, however, extended to handle CP-nets. A customized service is verified for the integration with other services and with other customers' services. The verified services are called Service Scenarios. A software architecture in which services can be executed is also discussed. The system consists of Scenario Database, Scenario Interpreter, and Functional Execution. By simulating the service scenarios, the customized services are executed.

This paper describes the improvement of the prior personal computer based karyotyping system of Giemsastained chromosomes that had resolution of 256256 pixels. To upgrade the accuracy of karyotyping the resolution of the image acquisition board was increased to 512512 pixels and the algorithm of finding the centromere of a chromosome was improved. The results of the karyotyping using this improved system with an image of chromosomes taken from the Rana Amurensis had been better than the previous one.

In computer vision and digital image processing it is an important task to estimate an ellipse most fitting a set of points. An ellipse is specified by nonlinearly constrained parameters, which are difficult to be reflected in a criterion function. We show that iterative processes are useful to solve problems in such a case, and propose an iterative procedure to fit an ellipse to a set of points. In the procedure the nonlinear constraints on parameters are represented by a properly defined nonlinear operator. Experimental results on personal computer indicate the effectiveness of the procedure.

Nonlinear operation of an analog frequency divider using a GaAs dual-gate FET has been studied theoretically and experimentally. It is shown that the log-log plot for basic input/output characteristics of the divider shows straight lines in the second, third, and fourth regions adjacent to the first narrow threshold region in the increasing order of input power magnitude. It is suggested from theoretical consideration that universality of the exponent K appearing in the PoutPinK relation holds at least approximately. The value of K in the second region is found to be about 3.0 theoretically and experimentally and is in agreement with the value reported in an earlier paper. The strict but obvious universality of K holds in the third and fourth region with K=1 and 0, respectively.