We discussed the cutoff characteristics of slab waveguide, and showed some numerical examples. We used the approximate dispersion equation which is valid in the frequency region near the cutoff, and is applicable not only to the lower modes but also to the higher modes.

In this letter, we investigated in detail the basic characteristics of W type slab waveguide in the vicinity of the cutoff frequency. Concerning with the leaky mode, we obtained some interesting results.

Scattering problem of an electromagnetic beam wave by a rectangular pit on a perfectly conducting substrate is rigorously analyzed. The pit is the canonical model of an optical disk pit. The numerical results of the far field patterns are presented for the various pit depths.

An end-coupled stripline filter is analyzed by the current normal mode expansion method. It is shown that the frequency characteristics are well described with finite number of current normal modes which are existed in the frequency range of interest. This fact intimates the possibility of synthesizing the circuit by Foster-type equivalent circuit.

A new optical limiter system including a viscoelastic polymer waveguide has been studied. The transmittance of this waveguide is changed by the feedback voltage proportional to the optical output power and hence the power of the optical output beam is suppressed to a constant value. Typical characteristics are demonstrated for a polyvinyl-alcohol thin film waveguide with a prism coupling.

As new nonlinear optical materials, several organic crystals such as N, N-dimethyl-p-nitroaniline and 5-methyl-2-nitrobenzoic acid have been grown in the form of crystal fiber from the melts. These crystal fibers with a 100µm core diameter and a 10 mm length generated an efficient SHG of Nd : glass laser. The observed SHG intensities were comparable to that of NBD-cloride crystal fiber.

A new fabrication method for a laser-to-single-mode-fiber coupling module has been developed, where the first spherical (or ball) lens of a double confocal lens system is affixed by a new UV (ultraviolet light) curable adhesive. Temperature characteristics and preliminary reliability test results confirmed that the module is sufficiently stable for usage in fiber-optic communication systems and other applications.

The diffraction losses of transverse modes of a surface emitting laser are studied by beam propagation method and the condition of single mode operation that when the diameter of one side mirror is about 5µm, fundamental transverse mode operation can be achieved, is given.

Single shot picosecond optical pulse waveform measurement has been investigated by an electro-optic sampling technique. An optical gate driven by an ultrafast photoconductive switch is employed as an optical sampler. Discrete sampled signals are coupled into an optical fiber array consisting of different length fibers to control the time delay. The output of the optical fiber array is detected by a photodetector and displayed on an oscilloscope with an expanded time scale. For measurement of 100 ps duration optical pulse, the time resolution of the system was estimated to be 14 ps.

We describe coherence multiplexing of ring interferometric fiber optic sensors using a short coherence length source. It is shown that the high finesse of the two ring interferometers with almost the same cavity length in the tandem arrangement may be exploited to yield a multiplexed system of increased sensitivity.

An optical amplifier repeater using two-stage semiconductor laser amplifiers with automatic power control (APC) system was inplemented for the first time. The APC system was realized by the direct monitoring of the amplified signal power. The assessment of the optical amplifier repeater confirmed the usefulness of the APC syster.

2.4 Gbit/s transmission experiment over 301 km employing five semiconductor-laser optical amplifiers in cascade has been demonstrated. The results present the feasibility of the long span optical fiber transmission using semiconductor-laser optical amplifiers.

Using an interferometric detection technique with AC modulation for searching the interference fringe signal, highly accurate polarization mode dispersion (PMD) measurement with about 0.0028 ps/m resolution has been developed, and the PMD measurement in a 60 km-long installed optical submarine cabled fiber is presented.

The thin aluminum oxide film formed by ECR plasma anodization was studied by x-ray photoelectron spectroscopy. It was observed that the composition ratio of the oxygen atom to the aluminum atom is actually smaller than the stoichiometric binding ratio. This result means that oxygen atoms are not easy to be taken into the film during plasma anodization. In other words, the anodization is performed without enough supply of oxygen atoms generated in the plasma. This fact suggests that the oxidation rate is limited by the generation rate of oxygen atom playing an important role in the process of the oxide growth.

A quantitative observation of damage on the electrical contact surface is very difficult so far because of no proper measuring system. Therefore, developement of a new measuring system for this purpose is needed for performance estimation of characteristics of electrical contacts. A digital image signal processing technique was applied for this purpose and the experimental results show that the developed system is very useful for quantitative observation of contact surface.

B(P)SCCO films were prepared on MgO substrate by a double cathode dc sputtering with a mosaic Bi_χSrCaCu_yO_z and (Bi_0.7Pb_0.3)_χ SrCaCu_yO_z target. The films were deposited at 200 and 550, and then annealed. We have obtained the B(P)SCCO films with T_c of 90-100 K under the condition of the deposition at 550 and the in-situ annealing in 200 Torr O₂ and the post-deposition annealing at 860.

An analysis is presented of the performance of the optimum detection for NRZ magnetic recording. The error rate performance of the optimum detection is obtained by theoretical analysis. And the recording density with the optimum detection is compared with that with some other well-known detecting methods.

The objective of this article is to investigate whether or not a network containing one capacitor, one inductor, and three resistors is able to realize a given biquadratic function. This information should be sought before any synthesis procedure is attempted. It is hoped that the technique may be extended to higher order networks.

For stable solution of a linear equation, the modified relaxation-based algorithms have been proposed in both fields of circuit simulation and digital signal processing. This letter describes the analogy between these modified iterative methods proposed independently in two fields.

This letter proposes a mechanical single-mode fiber-optic switch which interconnects a pair of probe fibers and pairs of line fibers to perform fault testing. The 16-channel switch array is fabricated and the switching performance is confirmed to be excellent.

A theoretical prediction of the spectral width for 1.55µm wavelength surface emitting lasers with GaInAsP/InP multilayer distributed Bragg reflectors is presented. Our calculation shows that a considerable reduction of the linewidth to less that 1 MHz can be achieved by an appropriate design of the laser.

A configuration of multi/demultiplexer using a 4-port optical circulator and interference filters is presented. Owing to the circulator, interference filters for the signal separation can be set normal to the optical signal propagating direction. Demultiplexing operation for 3 wavelength-multiplexed signals, of which wavelength spacings are 1.8 and 4.9 nm, is experimentally confirmed.

This letter proposes a new-type electrostatically-driven fiber-optic micromechanical switch by fabrication of a 32-element switch array and examines its switching performances. Low-loss and high-extinction-ratio characteristics are confirmed.

A simple adaptive scheme is proposed for controlling the length of time-steps for numerical simulation of Hopfield's networks.

Simple feed-forward type neural networks have been investigated to obtain those learning abilities and to obtain effects of initial values of network weights by back-propagation (BP) algorithm. Only restricted types of solutions are obtained by the BP algorithm out of many of them.

This paper discusses a new simultaneous design method of both magnitude and phase of IIR digital filters. It is inherently a nonlinear problem to approximate the magnitude and phase of an IIR digital filter simultaneously. In this paper, however, such a nonlinear problem is converted into a linear one by vector rotation method and solved by the linear programming (LP) technique. As a result our method requires no initial guesses for any type of filter specifications. A stable digital filter is designed which approximately satisfies the linearity of the phase in the pass band and the given attenuation in the stop band. The design examples given in this paper shows the usefulness of the proposing method.

This is a sequel of my previous paper. There, the mathematical definition of the stationary Markov renewal process (MRP) was given with a complete proof that the defined process is really a stationary point process on the whole time axis. As a consequence, various stochastic processes appearing in practice which are derived from MRP can easily be shown to be stationary. The author has been interested in the calculation of the power spectrum of MRP generated signals in electrical or nervous systems. For this purpose, it was necessary to construct a stationary process generated by MRP and hence it was inevitable to introduce a stationary MRP on the whole time axis. In this paper, the ergodicity of the one-parameter group of measure-preserving transformations associated with the time shift of the stationary MRP is discussed. An irreducible MRP is either aperiodic or periodic with some period. Here, a proof is given that an aperiodic MRP is strongly mixing, from which ergodicity follows. On the contrary, as a periodic MRP is only ergodic without being mixing, it is directly proved ergodic. Once the ergodicity of MRP is established, various stationary processes derived thereof are easily shown to be ergodic as well. This means that the characteristics, among which the power spectrum of each sample process is our concern, are sample-wise invariant, which will be useful in many applications. Recently, as MRP and related stochatsic processes are widely used in many fields, e.g. physics, OR, communication theory, and bioscience, our results will find large utilities in both the theory and the application of stochastic processes.

A low-noise 2-GHz-band VCO for frequency synthesizers used in mobile/portable radio sets is designed and implemented. This frequency band has become increasingly important for mobile/portable radio to alleviate the present frequency congestion problem below 1 GHz. So far, however, low noise VCOs which satisfy the severe requirements of mobile radio systems operating above 1 GHz have not been investigated. The design specifications are summarized, and it is shown that the VCO phase noise reduction requires a rigorous designed oscillator. Oscillator noise characteristics are reviewed, and design principles are clarified. In order to suppress the 1/f noise, a Clapp oscillator circuit and a microstrip resonator are adopted. A Si BJT and GaAs MES FET are compared under optimum bias conditions. The noise level of the Si BJT oscillator is 13 dB less than to that of the GaAs MES FET at a 25-kHz offset frequency from the oscillation frequency, and was thus adopted in our VCO implementation. The following performance levels were obtained: 28 MHz variable frequency range with control voltage from 1 to 5 volts, 112 dB single sideband noise to carrier ratio in 1-Hz bandwidth at 25 kHz offset, and 37.5 dB noise ratio from 5 Hz to 20 kHz with respect to signal.

This paper presents an analytical approach to the error performance of Trellis Coded 8-PSK with Viterbi decoding, over a Hard-Limited (HL) satellite channel. Although an error probability of Trellis Coded Modulation (TCM) depends on the transmitted sequence in general, we treat a sequence independent TCM composed of a convolutional code of rate 2/3 and constraint length 3 combined with 8-PSK. The channel is considerered to be disturbed by up-link and down-link cochannel interference (CCI) apart from both up-link and down-link noise. It is assumed that the system band is sufficiently wide so that the signal and each interference pass without distortion. For the non-Gaussian characteristics of received signal due to the nonlinearity and CCI, the moments of path metric is employed in the analysis. The first event error probability is obtained by the moments with the aid of the Gram-Charlier expansion without the knowledge of the actual probability density function (pdf) of the metric related with the non-Gaussian characteristics. The bit error probability is derived through the transfer function method. Numerical results of coded 8-PSK are presented for both hard-limited satellite channels and linear terrestrial channels with a comparison to uncoded QPSK.

The transmission and reflection characteristics of a H-plane symmetrical T-junction having a reflector near the junction plane are theoretically studied. The reflector is an equilateral triangle post, and it is installed in contact with the wall of the through-guide facing the junction plane to improve the wave propagation from the truncated-guide to the through-guide. In case of equal cross section, we find an optimum reflector, the top angle of which triangle reflector is 0 degree and its height is about half of the guide width. Roughly speaking, the fin, that is, the reflector of 0 degree has the characteristic value of (transmission coefficient |S₁₂| between the truncated-guide and the through-guide) 0.705 over a wide frequency range. This value is nearly equal to the upper limit value of |S₁₂| 1/20.7071. Thus 99.4 percent (0.705²2) of the energy fed from the truncated-guide transmits evenly into the other two arms, and 0.6 percent reflects. The problem is treated by a method of combining the boundary-element method and a hybrid-representation circuit equation.

Two algorithms are presented for a time-domain analysis of a switching converter which is replaceable with a piecewise-linear system. One of them is for the transient state analysis and the other is for the steady-state analysis. Both of them use the eigen-value and the eigen-vector calculations. The analysis based on these algorithms can be carried out more rapidly and accurately than the conventional analysis using standard fixed or variable step-size integration methods. On the uk converter, the results of the proposed method are compared with those of the Hamming method (a variable step integration method) and of SPICE2 (a general-purpose circuit analysis program using variable step-size integration method).

In this paper considered is the problem of scheduling a large number of parallel tasks (tasks having the same release and due times) on parallel processors. It is assumed that tasks all require one unit processing time, and the release and due times are non-negative integers. Task sets each of which consists of parallel tasks are dealt with instead of individual tasks in the conventional manner. The minimum number of processors required to process the given task sets is obtained by increasing stepwise the number of processors. To this end the minimal subset of task sets which cannot be allocated to a certain fixed number of processors is determined by a linear-order breadth-first search using the result of task assignment in two different ways, one top-down and the other bottom-up. The time complexity of the overall procedure is O(min(n, m)(mn log n)) and the space complexity is O(nm), where n and m are the numbers of task sets and time slots respectively.

One of the most important factors indicating the effectiveness of a word clustering method is how commonly it can be applied to different languages. This paper evaluates the applicability of a new word clustering method to the English and Japanese languages using word sets edited from technical summaries. The method employs an iterative clustering routine which increases the number of clustered words. Thus, evaluations are achieved as a function of the number of iterations of the clustering routine from the aspects (a) clustering characteristics determined from the number of clustered words, the number of clusters formed, etc., and (b) performance determined from the average clustering ratio and the average cluster uniformity. Consequently, the applicability of the method to English and Japanese is obtained through evaluations indicating similarities between them for both clustering characteristics and performance. It is also clarified that about fifty percent of the target words can be clustered in less than five iterations of the clustering routine.