In this paper, an algorithm for implementing a high-speed FFT processor and the memory structure relating with it are mainly discussed. The problem to be considered when implementing a hard-wired high-speed FFT processor will be as follows: (1) How to simplify the decision of data addresses to be accessed. (2) How to avoid the memory access conflict when introducing the parallelization of memory. New method presented in this paper gives a drastic solution to them. Namely, utilizing the proposed memory structure allows any number of butterfly computations as well as single one to be carried out substantially in only one memory cycle. Furthermore, there is no theoretical restriction in any level of parallelization. Therefore, it is easy to implement the FFT processor having a reasonable throughput for particular application. The discussion in this paper is primarily proceeded from the hardware standpoint of implementation. However, there is no essential disparity between the proposed organization and that of conventional computers. The only difference between them is a conceptual one with respect to memory structure. Thus the presented method is useful enough for implementation of the algorithm on a conventional computer system such as multi-processors, and for its implementation by hardware as well.

Realizability criteria for infinite dimensional linear dynamical systems are studied. We set up three types of systems, i.e., regular system, balanced system and Helton's system, in which state spaces are locally convex spaces. We obtain necessary and sufficient conditions for given weighting patterns or given transfer functions to be realizable with the above three types of systems, respectively. Moreover, we characterize the classes of weighting patterns and transfer functions which admit realizations by means of systems whose state spaces are either locally convex spaces, Banach spaces or Hilbert spaces, respectively, and clarify the relationship between these classes of functions.

The reliability of the FE method and the CV method for determining the localized state distribution N(E) in GD a-Si are evaluated by computer simulation. It is shown that calculation by the CV method is almost exact and the FE method is less reliable. But the FE method may be more reliable than the CV method about the accuracy of measured values. The new calculation method combining both methods and the iteration method are suggested for obtaining almost true distribution N(E). And the effect of surface states on calculation of N(E) is evaluated; the allowable surface state density is shown to be below 1011cm-2eV-1. The effect of the Fermi-Dirac distribution function F(E) at room temperature is estimated. The results indicate that correction by room temperature F(E) is necessary.

The authors have previously reported the algorithm of a lowpass-type variable cosine equalizer in the case of relatively rough input data. As a continuation of this work, in this paper we will propose a new algorithm applicable to the bandpass-type input data.

Several long (11.0-30.4 km) graded-index fibers with good transmission properties in the wavelength bands of 1.3 and 1.55 µm have been fabricated by the VAD method. Special attention was paid during the deposition process to improve the longitudinal uniformity in transmission characteristics of the fibers. The transmission loss uniformities of the VAD fibers along axial direction have been measured by the back scattering technique with use of Nd-YAG laser at 1.06 µm. The bandwidth uniformities of these fibers were also ascertained to be satisfactory. Preliminary transmission experiments on spliced 65.1 km and 116.5 km fiber systems were also made by use of 1.3 µm and 1.55 µm laser diodes.

The filter characteristics of electromagnetic TM waves in a dielectric slab waveguide having a sinusoidal variation of its permittivity in the propagation direction are investigated. With the aid of a singular perturbation procedure, the coupled-mode equations governing the nature of TM waves interactions are deduced for both symmetric and asymmetric first-order Bragg interactions. Omitting the end effects, an expression describing the characteristics of reflection of TM wave is derived from the boundary value problem in the propagation direction. Experiments on reflection characteristics have been performed in the millimeter wave region where the dielectric slab created by arranging an Aflon slab in periodic fashion in a Teflon slab is tested as a permittivity-modulated slab waveguide. Return loss data and observed stopband responses are reasonably consistent with theory.

The authors have studied the Penning effect and direct ionization of charged gases, Ne and Ar, in a gas tube protector, which is equipped with the trigger of carbon line and (BaSrCa)O-applied electrode, instead of the radioisotope, and have successfully realized a quite high-performance protector. The present study, which has been conducted on the mixing ratio of Ne and Ar as well as on the effects of gas pressure on the sparkover voltage for the impulse voltage at various ramp speeds and on the impulse life characteristics, has clarified the following points. (1) The impulse sparkover voltage can be held low, if the Ar concentration is limited to from 5 to 30% and the gas pressure is set to about 100 Torr. The increase of capacitance and rise of sparkover voltage is infinitesimal, even if the 10 1,000 µs, 100A impulse current has been applied 100 times. (2) Holdover voltage is approximately 80 V, when the 10 1,000 µs, 100 A impulse current and 350 mA DC current are applied. (3) Compared with a conventional radioisotope-charged protector, the newly developed protector has a low sparkover voltage for a high ramp speed impulse voltage, and its glow discharge time is extremely short.

Asynchronous sequential circuits, obtained by generalizing the astable multivibrator and monostable one as sequential circuits having state-transitions among any number of stable states and quasi-stable states, are called sequential circuits having quasistable states (SCQ's). Moreover, a mathematical model of the SCQ is called the sequential machine having quasi-stable states (SMQ). This paper is devoted mainly to the state reduction problem of the SMQ. There are two types of the SMQ's (celled type I and type II). The difference between the types is the specifications of each quasi-stable duration. Thus this paper describes state reduction methods for type I and type II, where the latter utilizes the former. The systematic state reduction methods described in this paper become considerably more complicated than that for conventional sequential machines, because of the particularity of the quasistable state. Of course, this method is an extension of the state reduction method for the conventional asynchronous sequential machine. This paper includes a definition and representations of the SMQ, definitions of concepts available for the SMQ's, and their systematic state reduction methods.

Candidate plans for adaptive digital transmission are briefly reviewed. It is suggested that adaptive format transmission with asynchronous multiplexing capability is attractive and promising from the standpoint of practical application. Feasibility and efficiency of this new transmission plan are investigated. A formula for predicting the minimum required transmission rate is derived. Upper and lower bounds of transmission efficiencies are analysed. A signal distributing technique for improving transmission efficiency is also developed. Investigation of memory requirements anticipates the possibility of simple implementations for format conversion and reconversion. Results of simulations prove the reliability of theoretical predictions. Fiber optic transmission is expected to be suitable for realizing this new communications concept.

Filtering techniques on a finite ring are paralleled to the techniques on the usual complex number field. A z-transform is defines over a finite ring which assimilates the complex number field. This z-transform is related to the usual z-transform and shown to have many properties in common. This z-transform can uniquely represent a filter on the finite ring with a given impulse response or frequency sample response function. The z-transform representation enable one to design recursive FIR (Finite Impulse Response) filter on the finite ring by the analogous methods of the usual z-transform. An interesting property of the recursive filter design on the finite ring is that stability problems due to particular design configulations need not to be considered so long as the inputs are limited to a certain value so that the design problem becomes primarily a problem of complexity reduction.

This paper describes an analytical investigation of mode-coupled graded index (GI) fibers. The analysis is based on the Scattering Matrix Method. The mode coupling parameters of the fibers are clarified in calculations and experiments, and the results obtained are used to determine the actual transmission characteristics of long, spliced GI fibers, for the wavelength region near 1.3 µm, as well as for the region near 0.85 µm. It is determined that coupling between adjacent modes is dominant and that a microbending loss of an order of about 0.1 dB/km results in an extremely large increase in the bandwidth of GI fibers larger than 20 km. In this paper the propagation constant for mode groups with the same propagation angle is derived directly from a WKB analysis.

The "relativized Turing machines" are introduced as media executing program schemas. On these machines, it is proved that, for any a, 0a1, there exists an O(na) space and O(n2-a log n) time bounded program schema to sort n elements, and any O(na) space bounded schema to sort n elements requires more than O(n2-a) time. Furthermore, it is shown that there exists a problem which causes a large scale time-space trade off. That is, we show the existence of a problem Q such that there is an O(n2) space and polynomial time bounded schema to compute Q, and there is an O(n log n) space bounded schema to compute Q, and further, any S(n) space bounded schema requires more than exponential time if S(n)n2.

The performance of a pulsed HF chemical laser with an F2/H2 chain reaction system is widely compared to that of an F/H2 non-chain reaction system by using a chemical laser simulation code based on the constant gain method. According to the simulation results, peak power differs little between both systems; laser pulse duration is more than an order of magnitude shorter in the F/H2 system than that in the F2/H2 system; laser pulse energy in the F2/H2 system is greater than that in the F/H2 system; difference in these energies is exponentially proportional to the inverse of the initial fluorine atom concentration; laser pulse energy per mol of HF() produced prior to termination of the lasing differs little between both systems; the thermodynamical efficiency is a decreasing function of the initial fluorine atom concentration in the F2/H2 system while in the F/H2 system it has a maximum value at the certain concentration.

This paper describes designing on required signal to crosstalk power ratio of symmetrical cable for bipolar pulse transmission. Error rate in bipolar pulse transmission is discussed taking into account average signal to far-end crosstalk power ratio, average signal to near-end crosstalk power ratio and their standard deviation. Experiments were conducted with regard to the relation among average signal to crosstalk power ratio, its standard deviation and error rate. Calculated and measured results show good agreements. From these results, designing on required signal to crosstalk power ratio and its standard deviation in dB for required error rate is clarified.

Several analytically solvable queueing models of multiprogramming with different jobs and various scheduling disciplines are investigated. It is shown that the analysis of these models supports the optimality of the already proposed CPU scheduling discipline which assigns the higher processing priority to the more I/O bound jobs. Furthermore, the effects of some endogenous scheduling disciplines such as preemptive-resume last-come-first-served are compared with that of this discipline using some queueing models.

A left identity type automation is defined as a finite automation whose transformation semigroup has a left identity. In this paper, we give a necessary and sufficient condition for a given automaton to be of left identity type in terms of the generators of its transformation semigroup.

In this paper, a partial matching problem among pictures composed of multiple closed regions is described by taking a sequence of animation pictures as an example. Closed regions including holes and islands are analyzed and described by a set of edges specified by angles, positions, trailing and leading vectors. Then, a kind of relaxation methods is used to make partial matchings between the preceding picture (template) and the current picture through global and local matching processings. This method was applied to a sequence of cel animation pictures drawn manually, and as the results, more than half of closed regions included in pictures were identified correctly.

Dependence of GaAs power MESFET performance on the device parameters was experimentally investigated. Power gain and output power were most sensitive to the gate length, giving the degradation rates due to increase of the gate length of 1.7 dB/µm for the power gain and 0.67 dB/µm for the output power. The effect of the gate finger width on the microwave performance was not serious and the performance deterioration at 8 GHz did not begin until the gate finger width of 200 µm. Temperature dependence of the microwave performance of GaAs power MESFET, as well as the dependence of the thermal resistance on the device geometric parameters, was investigated quantitatively. Power gain degraded with increase of the device temperature with the rate of 0.026 dB/.

It is shown that we can simultaneously make both of Zech's logarithm table and trace table in a finite field by using a feedback shift register which generates a pseudo-random sequence. Convenient method for the initial loading of the feedback shift register is given.

The m-center problem asks to place m objects on the plane so that the distance from a client (a point) to the closest object is at most a given number r. This problem is often encountered in locating facilities or resources of a geographically distributed system. This paper shows that this problem is NP-complete. The NP-complenteness indicates its computational intractability, i.e., it is most unlikely that some algorithm can solve it in polynomial time of the problem size.