Variation of power distribution caused by splices is calculated as a function of numbers of splice points and misalignment. The calculated power distribution is verified by experiment. Calculation shows that mode conversion caused by splice makes power distribution approach to steady-state. The numbers of splice points required to make power distribution stationary are calculated as a function of misalignment. Pulse broadenning in the fiber transmission line is calculated for the case that fibers are mode conversion free. Pulse broadenning rate becomes proportional to the square root of transmission distance after optical pulse passes many splice points. These calculated results show that mode conversion caused by splice has similar tendency as that caused by longitudinal inhomogeneity of waveguide.

The statistical model for urban radio propagation has been proposed to describe the characteristic of radio pulse propagation in urban area. A modified Poisson process was used to model the clustering phenomena of the pulse arrival time sequence. In this paper, the model analysis is extended for more flexibility in data fitting, and generalized using a Markov chain to get a clear view of the statistical model. Some numerical results are shown for different values of the clustering parameter (Δ). Referring to the numerical results, the effect of Δ and the another clustering parameter (K) on the pulse number distribution is discussed. Next, the relation between the empirical pulse occupancy rate and the underlying occupancy rate of the model is derived for Δ1 using a Markov chain. By this extension and generalization, the statistical model of the pulse arrival times will be applied to a wide range of the clustering phenomena.

We consider an adaptive PCM system, in which the input samples are expanded or compressed by a constant factor c or 1/c each time before quantization. Assuming a stationary Gaussian input with a rational power spectral density, we derive an integral equation for the joint distribution of the input and the state of the system. Its solution provides us with a feasible way to numerical computation. The mean-squared error are computed for the Gauss-Markov input in terms of the constant c, the sampling interval T, bound parameters for the scaler and the number of quantization levels. The numerical results show good performance in comparison with regular PCM.

By the term peak frequency" of a short-time power spectrum, we mean a frequency at which the power spectrum takes a maximal value with respect to frequency. Peak frequencies are recognized to be important feature parameters for machine recognition of sound signals. We propose a new method for extracting the peak frequencies from a pitched signal like speech sounds, musical instrument tones, and so on. Our method makes use of the first and the second frequency derivatives of a short-time power spectrum, so that it is called Power Spectrum Derivative Method (PSDM)". PSDM has such an advantage that exact peak frequencies having close relevance to the resonance and harmonic components can be extracted from a short duration signal. PSDM was applied to the analysis of Japanese vowels and guitar sounds with good results.

We formulate an improved lower bound on the maximum number of prime implicants of n-variable Boolean functions. It is given as n!/(n/3!(n1)/3!(n2)/3!)(n, 0, (n1)/32)(n, O, (n2)/32), where (n, 0, r) is evaluated by the following recursive procedure: (n, 0, r)0 for r0, (n, 0, 0)1 and (n, 0, r)n!/(r/2!(nr)!(r1)/2!)(n, 0, (r1)/2 2) for 1rn. The total logarithm computing time cost and the total uniform computing time cost of this lower bound by a random access machine are O (n (log₂n)²) and O (n), respectively. The ratio of this new lower bound to the old lower bound is bounded by 1c (1/2)^n/3, where c is a constant independent of n.

This paper is concerned with reparing time evaluation for a multi-subequipment system. By Convolution application, a renewal of the system is shown to depend on the repair executing sequence, and to be accomplished merely by replacing 0.264 times located cards.

First, it is shown that pole sensitivity with variation of gain-bandwidth product is dominant in an RC active circuit realized by an internal phase-compensated operational amplifier. Second, a method is described, by which a single amplifier 2nd-order RC active circuit with prescribed pole sensitivity with variation of gain-bandwidth product can be realized.