An empirical Bayesian method was used to obtain a point estimator for the reliability function of a bivariate exponential distribution associated to a two-component parallel electronic system.

A discharge fusion splicing technique for fluoride optical fibers has been investigated. Discharge heating conditions were experimentally examined by making a fiber offset from the electrode axis. An average splice loss of 0.14 dB was achieved for 50µm-core fibers.

An observation of a neutral IBW (ion Bernstein wave) modified by a pure IBW in a large space chamber was represented comparing with the theory. The results for the dispersion relation, the excitation coefficient and the maximum damping rate were discussed.

This paper presents an antenna for the UHF-TV reception consisting of skeleton slot array installed on a reflector. We propose a dipole model of the skeleton-slot elements to calculate its radiation pattern. The prototype antenna has 13 dBi antenna gain, 30% frequency band width and a function of beam tilting, which are sufficient characteristics for the UHF-TV reception.

The authors measured the temperature dependence of the level of the 1/f noise in carbon-black graft-polymer resistors. A distinct dip is observed in the noise level at a temperature at which the temperature coefficient of the resistance is zero. This is in agreement with our temperature fluctuation model on the 1/f noise in resistors which include a large number of small conducting spots.

The single-row single-layer routing is considered for net lists with density5 when backward moves are allowed. A heuristic routing algorithm is given and tested. Comparing with a conventional router, better realizations were found for 38% of randomly generated net lists.

Our primary purpose of this paper is to study attractors and bifurcation phenomenon appearing in an eventually bounded circuit. The secondary purpose is to demonstrate the results obtained from the instruments which display Lorenz maps, Poincaré maps and cross sections of attractors on a synchroscope. The above mentioned circuit contains two non-linear resistors and is written as 3 first order differential equations. Experimental observations show:
(a) maximum number of the attractors in three.
(b) maximum number of the stable limit cycles is three.
(c) maximum number of the chaotic attractor is two.
Observed bifurcations are:
(a) Hopf bifurcation,
(b) period doubling bifurcation,
(c) periodic window.
Furthermore, we examined the Feigenbaum constant δ experimentally and estimated it at 4.67.

Let γ and n be positive integers. An integer z with gcd(z, n)1 is called a γ^th-residue mod n if there exists an integer x such that zx^γ (mode n), or a γ^th-nonresidue mod n if there doesn't exist such an x. Denote by Z^*_n the set of integers relatively prime to n between 0 and n. The problem of determining whether or not a randomly selected element zZ^*_n is a γ^th-residue mod n is called the γ^th-Residuosity Problem (γ^th-RP), and appears to be intractable when n is a composite integer whose factorization is unknown. In this paper, we explore some important properties of γ^th-RP for the case where γ is an odd integer greater than 2, and discuss its applications to cryptography. Based on the difficulty or γ^th-RP, we generalize the Goldwasser-Micali bit-by-bit probabilistic encryption to a block-by-block probabilistic one, and propose a direct protocol for the dice casting problem over a network. This problem is a general one which includes the well-studied coin flipping problem.

A general analysis of the effect of the power-law envelope detector by the use of output harmonic components on signal-to-noise ratio (SNR) is performed. By using multiplication and filtering, the additional zero-frequency components are obtained from output harmonic components of the power-law envelope detector. An general expression of the improvement factor on SNR is obtained which is applicable to non-Gaussian noise. It is found that the output SNR can be improved by using the additional zero-frequency components from the output harmonic components of the power-law envelope detector.

For the quantum state control communication, one of the most serious problems is the effect of transmission loss which degrades the advantage of the quantum state controlled signal. In this paper, we investigate loss effect for the squeezed state signal, which is one of typical quantum state controlled signals. First, it is shown that when the squeezed state signal optimized regardless of the effect of loss is used as transmitter state in the communication system with energy loss, the signal to noise ratio of the received signal is higher than that of the coherent state signal with the same transmitted photon number when and only when the loss is less than 3 dB. Then this paper gives an optimum condition of the squeezed state signal for lossy channel. This optimum condition provides higher signal to noise ratio of the received signal than that of coherent state signal for any degree of loss. Furthermore, we compare the performance of the quantum coherent communication systems using the optimum squeezed state signal with that of the photon communication using the photon number state signal taking the effect of loss into account.

A thermal model of a TO-5 package in steady state has been derived. The model describes analytically the heat conduction, the heat transfer by free convection and the thermal radiation, taking account of a complicated structure of the package. It has been applied to a study on the negative dynamic resistance by self-heating which is observed in MOSFETs at high current level. The sublinear dependence of the temperature of the cap on the dissipated power in the MOSFETs are explained by the temperature dependence of the heat transfer by free convection.

A design method of multilevel logic networks, in single-rail input logic, with MOS cells for parity functions of n variables is presented. The parity functions x₁x₂x_n in a two level network requires 2^n-1+1 gates, whereas the number of NOR gates in a multilevel network for this function is proportional to n. Since MOS cells are playing a major role in LSI and can represent more complex functions than a NOR or a NAND gate, it is possible to obtain smaller networks with complex MOS cells than with NOR or NAND gates. However, if MOS cells become excessively complex, the propagation delay time of MOS cells is too large and the normal operation of the whole MOS network cannot be expected. The design procedure given in this paper is systematic and very easy to apply. In the beginning, the formulation of MOS cells and MOS networks has been presented. In this formulation, the number of series and parallel transistors and the number of MOS cells and levels can be defined and enumerated for MOS networks. Without the restrictions the maximum number of series transistors is the number of input variables, n. With the estriction on the number of series transistors to R, the number of parallel transistors is less than or equal to 2^R-1. Furthermore, the number of MOS cells and the number of levels of the MOS network are proportional to n and log_Rn, respectively. Since the designed network is similar to a complete tree, it would also be useful to process a series of combinations of input variable values.

Imaging the acoustical nonlinear parameter using second-harmonic and difference-frequency waves produced from finite-amplitude sound propagation is described. Experiments for reconstructing nonlinear-parameter tomograms were performed with gelatine phantoms. The tomograms were not of high quality because we used a small-aperture receiver to detect the substantially refracted waves. We show that by laterally translating the receiver to the position of maximum intensity for each measurement of a point on a projection, we can improve the quality of nonlinear parameter tomograms as well as that of conventional attenuation tomograms. We analyze the effects of attenuation and diffraction for a narrow transmitted beam in nonlinear parameter imaging. Using second-harmonic waves is advantageous for attenuation compensation when the attenuation in the object is proportional to the frequency, such as is the case for soft biological tissue. On the other hand, using difference-frequency waves has certain advantages where attenuation is not a problem. With these waves there is less phase variation.