This letter discusses a relation between reliability and network topology of starred polygon with redundant nodes. Each node of starred polygon is augmented by a spare node, and the degree of survivability is newly defined to evaluate reliability. The main result presents three special topologies of starred polygon with redundant nodes that realize an optimal degree of survivability.

A new approach for topological routing is proposed by W-graph. We employ a W-graph G_w(V, E, W) for indicating all nets which will be assigned to two-layer, where V is a set of all terminals, E is a set of edges corresponding to two-terminal nets and W is a set of wild components corresponding to multi-terminal nets. Such that the topological routing problem can be considered as: Given a circle H containing V in the sequence corresponding to terminals on the boundary of routing region, then drawing H G_w on a plane with minimum number of created vertices (crossing points on H).

In this letter, a constructive simplified Newton method is presented for calculating solutions of infinite dimensional nonlinear equations, which uses a projection scheme from an infinite dimensional space to finite dimensional subspaces. A convergence theorem of the method is shown based on Urabe's theorem.

Multi Step Diakoptics (MSD) is a parallel computation method for solving linear equations. It is used for device simulation, being it executed on a parallel computer. The results show the efficiency of the simulation when the number of nodes that are defined by the five point discretization of semiconductor equations, is over 2⁹.

This note presents a new global optimization method and derives a learning schema based on the method for multilayer artificial neural networks. The schema consists of (1) pasting" the admissible region in Rⁿ to a n-D torus Tⁿ and smoothly connecting the potential function at the boundary; (2) global searching along the flow of a nonvanishing vector field on the compact smooth manifold Tⁿ. This flow is featured by the ability of automatically passing through distinct local minima one after another along the negative/positive gradient field. It has also a unit norm everywhere on the Tⁿ, so the searching speed will not slow down in the neighborhood of critical points.

This letter considers the problem of modeling a nonstationary Gaussian ARMA process. The corresponding approximation problem is formulated on the basis of the theory of canonical representation of the Gaussian process. Further, it will be shown that the solution can be obtained by solving a set of linear equations, as an extension of the Padé approximation established in the stationary case.

In this letter, a method to analyze the steady-state characteristics of multirale IIR adaptive notch filter is proposed. A closed-form expression for the steady-state mean square error of variable coefficient has been obtained. The results of computer simulation are shown which confirm the theoretical prediction.

A user-friendly simulator PANDA (Program for the Analysis of Networks with Digital Arithmetic) is introduced, which can analyze the frequency responses, coefficient sensitivities and roundoff noise of given linear shift-invariant digital networks from their structural descriptions.

This letter presents, as an extension of a previous work of the present authors, a design method of IIR digital filters, which is guaranteed in principle to be stable, by using the interpolation of the characteristic function of the filter for a set of prescribed values of the function and, in addition, its higher derivatives. As a numerical example, an IIR digital filter having the characteristics such that it has a maximal flatness in the passband and prescribed steepness in the transition band is designed to show the effectiveness of the proposed method.

An architecture for radix-2, decimation in frequency fast Fourier transform butterfly computation unit is proposed. It operates on 16 bit inputs and uses 'merged core' type of multipliers which reduce hardware components and hence promise smaller chip area. This reduction in area is achieved with negligible increase in computation delay.

The phase-modulated optical fiber gyroscope signal was analyzed in a time-domain. The rotation rate detected by the gyroscope optics was extracted from the gyroscope signal as time-domain characteristics at the maximum sensitivity.

This letter describes a new thresholding method for recognizing pattern regions in noisy gray-scale images. Many previously developed thresholding techniques occasionally lead to unreasonable results for noisy images with small pattern occupancies. In the proposed method, the threshold is evaluated on the basis of the difference function of a pair of local histograms for two windows set in the image. The zero-cross point of the function is independent of both the noise spectrum and the pattern ratio, and it coincides with the true threshold. This method is a powerful tool for thresholding noisy gray-scale images that have unknown pattern ratios.

An errors-and-erasures decoding procedure for t-error-correcting Reed-Solomon codes with designed distance 2t1 is proposed which corrects some patterns of erasures and errors beyond the BCH bound. In the procedure, a decoding algorithm beyond the BCH bound is applied only when the number of erasures is odd.

In this letter we present an efficient root finding technique for a polynomial over GF(2^m) when m is even number. The solutions that we have ever known are made to accelerate by our technique.

In order to investigate the dynamic balance function, we have studied about the tracking motion in standing posture. We have analyzed the motion using the multiple regression analysis method. It is found that the regression coeffcients indicate the characteristics of the motion.

With the use of the Fourier-Laplace integral, the impulse response for a two-dimensional model of the Cherenkov-type oscillator is obtained. From the numerical analysis of the impulse response, the temporal evolution of the response is found to have two different phases ini which it grows linearly and exponentially with the distance from the origin where an impulsive excitation is applied.

In this letter, an analysis on transmission-characteristics of a new type broadside-coupled strip transmission-line is presented which consists of two center strips placed respectively on the upper- and lower-side surface of a rectangular dielectric substrate with a large dielectric constant and two ground electrodes placed respectively on the other surfaces of the rectangular substrate. Analytical expressions for characteristics of this transmission-line are given througn a conformal mapping technique and numerical results are given to facilitate the design works. Using this new strip-line, microwave miniature multi-stage broadside-coupled filters are easily realized.

Simple analytical equations, representing power distributions in a propagation direction for a Cerenkov-type SHG, were derived. Using this model, dependence of conversion efficiency on propagation loss has been analyzed. Waveguide loss affects the beam profile of the output second harmonic wave, as well as the conversion efficiency. The power distribution in an external-resonant-cavity configuration can be treated by this model. The waveguide loss effect is more conspicuous for the cavity-type SHG.

The beam-propagation method (BPM) is applied to the evaluation of the loss of S-shaped step-index waveguides. It is demonstrated that the BPM enables us to analyze S-shaped bends which have an offset core and a trench section in the cladding at the outer side of the bend.

This letter describes transformation of an elliptical emission region of a semiconductor laser into a triple mode pattern of a graded-index optical fiber by a holographic filter. As a result of comparison of experiment with computer simulation, it has been shown that this method is effcient for the above transformation.

In this letter, the temperature properties of laser speckle pattern generated from a multimode optical fiber are studied. It is shown that the power spectral width of speckle intensity variation induced by the temperature change is proportional to the fiber length and the rate of temperature change.

The relationship between steady arc duration and contact resistance characteristics is studied on Ag-Pd30% alloy contacts. The experimental results show that longer arc duration doesn't necessarily cause higher contact resistance, which is the same result as obtained on silver contact.

Possibility of high performance P-channel HEMT was investigated. P-channel HEMT structure with an InGaAs strained layer was fabricated and yielded a measured hole mobility of 260-310 cm²/Vs. Numerical simulations based on the experimental results indicate an expected transconductance of 186 mS/mm for the intrinsic part and 82 mS/mm for the extrinsic part (source resistance of 6 Ωmm) for 0.1/µm gate HEMTs.

A new DRAM bitline architecture, called Divided/Pausing Bitline Sensing Scheme (DIPS), is proposed for DRAM core of 64 Mbit level and beyond. This architecture eliminates the inter-bitline coupling noise and realizes a high speed sensing operation.

This letter describes the fan-out optimization method of the SRAM decoder having line capacitance that minimizes the total delay time. It is shown that the total delay time of the SRAM decoder optimized by this mothod is less than that of the equal fan-out condition.

Delay time and power consumption, serious problems in high-density SRAM's are simulated over several generations with the HWD architecture. The optimum grade of hierarchy is obtaind for 64 kbit to 256 Mbit SRAM's.

We analyze the substrate corrent of submicron transistors in the memory cell of ULSI SRAMs in the cases of PMOS and NMOS bit line loads. The lifetime of the transistors is also estimated. The SRAM using an NMOS bit load is found to be better as a hot carrier than that with a PMOS bit load.

This letter proposes a new-type of two-dimensional multichannel optical switch employing polarization control techniques. This switch has small-size and low-driving-power characteristics. It is demonstrated by an 88 optical switch.

A new phenomenon of a slow-fast system, canard, is studied. Some interesting propeties and the existence of canards are shown by using non-standard analysis and singular perturbations. There results explain the strange behavior of a nonlinear circuit with a small parameter.

A numerical design of a coaxial-to-radial line adaptor is presented for the use as a feed in a radial line slot antenna. To realize stable performances in mass production, the reflection from a probe type adaptor in which only the outer conductor of a coaxial line is in contact with the waveguide, is analyzed and suppressed. The tolerance for the change and the errors in the height of the waveguide as well as the bandwidth is highlighted; the advantages of the conical probe over the conventional shorting post and the coax-gap adaptor are emphasized.

A process algebra dACP based on differential equivalence (d-equivalence) and the algebra ACP by Bergstra et al. is proposed. The features of dACP are as follows:
・Real parallelism can be distiguished from action interleaving.
・Action refinement is possible under d-equivalence.
Thus, dACP resolves the difficulties of ACP caused by the atomicity of actions assumed in ACP. After defining dACP, d-equivalence is shown to imply the original equivalence defined in ACP. Next, it is shown that action refinement can be introduced in a natural way under d-equivalence.

We examine the problem of reporting all intersecting pairs in a set of c-oriented polygons, each having at most k edges (for some constant k). Polygons are called c-oriented if the edges of all polygons have a constant number of orientations. The problem arises in many applications such as VLSI design rule checking and architecture databases. We present an asymptotically optimal algorithm that reports all pairs in O(n log nt) time and O(n) space, where n is the number of polygons and t the number of intersecting pairs. Since the optimal algorithm may report an intersecting pair more than once (but at most a constant number of times), we also give a further algorithm which reports each pair exactly once but is not space-optimal, namely, requires O(n log n) space.

A sequence α(a₁, , a_n) is k-sorted if for all 1i, jn, i jk implies a_ia_j. The number of k-sorted permutations of {1, , n} is denoted by R_k(n). We show that 1.4439n and 1.6933n are average case lower bounds on the numbers of comparisons needed to sort 3-sorted and 4-sorted sequences of length n, respectively. A compact form of a characteristic equation of degree k1 for a lower bound on R_k (n)/R_k(n1) is derived. From this characteristic equation, we also derive a good lower bound on the expected number of comparisons needed to sort k-sorted sequences for each k5.