摘要:

AbstractPeripheral vision is important in the human visual function and permits observation of the objects apart from the line of focus. Most of the peripheral vision functions are still unknown. This paper reports on the detectability of a stimulus pattern at the center of the visual field and at the periphery of 23° from the center, which is affected by the coexistence of another pattern. It is found that the isolated stimulus pattern is easily detected at the periphery, in contrast to the center. It is also found that when stimulus patterns form a cluster the cluster as a whole is more easily detected, but detection of the individual patterns forming the cluster is difficult if they are small

ISSN:8756-6621    

DOI:10.1002/ecja.4410690201

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractNonlinear feature extraction methods for pattern recognition have been proposed. Fukunaga [9] and Otsu [7]have shown that thea posterioriprobability function is optimum as the nonlinear feature extraction function for the discriminant analysis or least mean‐square error estimation. However, it is a theoretical function which is considered as a kind of goal to be achieved. As an alternative, this paper proposes a distance function F(x, ω) which behaves similarly to thea posterioriprobability function. The value of F corresponds to the distance between the pattern x and the pattern class w. The number of classes to be recognized is 2. In the method, the multidimensional pattern x is transformed into a two‐dimensional vector y = (y1, y2) whose components are the distance to the classes ωi, i = 1, 2, i.e., yi = F(x, ωi), i = 1, 2. The interactive display is fully utilized in the design and evaluation of F. Once F is designed based on the design pattern sets, the design patterns are all transformed into y for evaluation of F, and the result is displayed on the screen. The adequateness of F is determined by the observation of y on the screen and the value indicating the separability of the two classes. If the choice of F is not adequate, the class representation is modified, and F is redesigned through the interactive processes. Through such interactive procedures, F is improved gradually, eventually arriving at the optimal

ISSN:8756-6621    

DOI:10.1002/ecja.4410690202

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractA model to estimate the opinion scores for overall satisfaction level of speech quality is an effective tool in designing telephone networks and equipment. This paper proposes a telephone transmission performance evaluation model named OPINE. It deals with fundamental factors (loss, noise, room noise, frequency band limitation) that degrade telephone transmission performance. Three psychological factors affecting performance are selected: 1) impairment of speech for band limitation; 2) effective loudness loss or excess of speech; and 3) noisiness during speech interval and non‐speech interval. After finding PI (Performance Index) that expresses the degradation degree of performance for each factor from the spectra of sound and noise in the ear, MOS (Mean Opinion Score) is predicted. The model was applied to the experimental data. As a result, it is shown that the predicted MOS agrees fairly well with the subjectively evaluated MOS, where the standard deviation between these is 0.2

ISSN:8756-6621    

DOI:10.1002/ecja.4410690203

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractThis paper proposes a numerical method of solution for the transient phenomena on transmission line terminated by simple load, including a nonlinear element. The method is based on the discrete numerical Laplace transform and its inverse, utilizing the discrete Fourier transform. Problems and the solutions in the method are discussed. The feature of the method is that the voltage, current and surge impedance matrix of the transmission line are specified on the complex frequency (s) plane, while the boundary condition is given on the time (t) domain. Numerical solutions in the two regions are combined by Laplace transform and its inverse. The Laplace transform and its inverse by discrete Fourier transform have a drawback in that the accuracy of the computation deteriorates at t = 0 for stepwise change of the waveform. A method to solve this problem is described. For the latter half of the sampling point sequence in the Laplace and inverse transforms, the computation error is increased, which effectively halves the sampling points in the computation of the reflected wave. For this problem, a method is presented by which the accuracy of the computation is retained for each calculation of the reflected wave, keeping constant the number of sampling points. The convolution required in the calculation of the boundary condition is time‐consuming, and a solution for this is proposed. As an example of the solution, the transient phenomenon in the multiconductor transmission‐line terminated by the surge arrester is calculated, taking the skin effect into considerat

ISSN:8756-6621    

DOI:10.1002/ecja.4410690204

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractTo improve the accuracy of solutions by the finite‐element method (FEN), it is necessary to refine the element division or use higher‐order elements. However, in both approaches, large‐scale/long‐time computation is necessary because the dimension of the matrix equation to be solved increases. Especially, in vectorial or 3‐dimensional analysis, this is a serious problem with respect to computational cost. By the error analysis theory of FEM, the asymptotic property of errors is known when the interpolation polynomial of certain order is used. In light of this, we use the so‐called deferred approach to the limit and derive the extrapolation formula making use of the asymptotic property of errors. This leads to the more efficient use of FEM, which is one of our objectives. Also, an algorithm using the iterative extrapolation method is proposed to obtain more accurate solutions. The theory is tested numerically for the eigenvalue problem of a dielectric‐loaded rectangular waveguide. The estimate of the exact eigenvalue is obtained by using the extrapolation formula described above. Hence, it is shown that very accurate results are obtained by merely repeating small‐sca

ISSN:8756-6621    

DOI:10.1002/ecja.4410690205

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractThis paper defines that globally there exists a state equation when a dynamical system of time‐invariant nonlinear network can be described only by capacitor voltages and inductor currents. A sufficient condition is presented for the state equation to exist globally. A sufficient condition is also shown for the state equation to exist globally, if the capacitor voltage and the inductor currents are replaced by capacitor currents and inductor voltages. Then the effect of the parasitic resistance or parasitic reactance on the existence of the state equation is discussed. When only parasitic resistances are added, the existence of the global state equation is preserved. However, the existence of the global state equation based on the capacitor currents and the inductor voltages is related to the eventual passivity of the original network. The latter half of this paper discusses the effect of perturbation of the branch characteristics constituting the network on the existence of the global state equation. Especially, it is shown that when the element is strictly locally passive, both the existence of the global state equation and the strictly local passivity are not affected by the perturbatio

ISSN:8756-6621    

DOI:10.1002/ecja.4410690206

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractThe fundamental equation for the voltage and current in the transmission‐line with distributed sources can be converted into the state equation in the operator domain by using the Laplace transform. The general. solution of this equation contains the initial condition, which can be specified arbitrarily. This paper presents a method in which the arbitrary condition is specified by numerically solving the system of equations derived from the boundary condition of the transmission line. By applying the inverse transform to the solution for the voltage and current of the transmission line obtained in the operator domain, the steady‐state and the transient solutions in the time‐domain can be determined. As the inverse transform, ac analysis is applied for steady state, and the numerical Laplace transform is used for transient analysis. As an application example, the steady‐state and the transient analysis are performed for the transmission line being affected by the external electromagnetic field. The effectiveness of the proposed method is verified by comparing the result with the result of other

ISSN:8756-6621    

DOI:10.1002/ecja.4410690207

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractPassive bus user‐network interface which connects multiple terminals to the network terminating equipment (NT) is promising as ISDN user network interface. Problems in realizing this interface are the discontinuity of phase and/or amplitude between channels and the effect of parallel operation of terminals on the transmitted or received waveform. This paper proposes a generalized DPLL performance analysis method based on a time‐variant Markov chain model with special reference to the passive bus user network interface with multiplexed channels whose mean phases are different. A new waveform simulation method that allows plural drivers to transmit is also proposed in designing a transformer. The effectiveness of these methods is demonstrated experimentally. It is also shown experimentally that the sampling margin derived from the DPLL jitter and the waveform distortion determines the maximum transmission distance in the passive bus system without a line equali

ISSN:8756-6621    

DOI:10.1002/ecja.4410690208

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractMultiprocessor controlled electronic switching systems suffer from performance degradation due to common memory access contention. Therefore, there have been many studies on the common memory access contention. However, most of them deal with the case of equal access rate. This paper considers common memory access contention in multiprocessor controlled switching systems, in which one of the processors, called MCP, has a different common memory access rate than the other processors, called CPs. In this multiprocessor system, common memory access requests are served according to non‐preemptive priority discipline based on the processor type. This common memory access contention is modeled as a finite source queueing system with nonpreemptive priority discipline. The model is analyzed and the effect of nonpreemptive priority discipline on system performance (number of instruction steps executed) is evaluated, compared with the ordinary FIFO disciplin

ISSN:8756-6621    

DOI:10.1002/ecja.4410690209

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY

摘要:

AbstractAn optimum design of alternative routing systems using the equivalent random method was proposed by Akimaru et al. However, it has a problem in that an individual call loss differs depending on call types, because of the design which places the average call loss rate of the alternative route within a predetermined value. In the present paper, we propose alternative routing systems with partial overflow scheme as a method to control the individual call losses. We describe the optimum design which minimizes the system cost under a given point‐to‐point call loss rate between origin and destination, and apply GI approximation, which we proposed before, to evaluate the individual call losses. The design method is described on a basic triangle alternative routing model, and the approximation is compared with an exact solution via numerical examples. It is shown that the design method has sufficient accuracy in less computational time compared with the exact solution. Further, it is clarified that the method prevents excessive investment with reasonable service and that a significant saving in system cost is attained compared with the conventional alternative routing sys

ISSN:8756-6621    

DOI:10.1002/ecja.4410690210

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1986

数据来源: WILEY