음성신호의 진폭이 갖는 학율밀도함수(PDF)를 이용하여 음성신호가 균일한 스탭을 갖는 양자기 입력에 가해질 때 발생하는 잡음을 양자화 잡음 및 포화성잡음으로 분리하여 잡음에 관한 표현식을 유도하였다. 이러한 잡음 표현식에서 빗트수 및 포화점을 결정하는 상수에 관해서 수치해석함으로서 각 빗트마다 SNR을 최대로하기 위한 조건을 제시 하였다.

The optimal selection of Prime Implications for the multi-output switching function is difficult task, as the input variables increase. This paper is concerned with the technique for the minimization of the multi-output switching function using the intersection table. This procedure is applicable to both manual and computer-programmed realization without complexity.

This paper deals with the computer method of finding the minimal sum of products for a switching function by using the MASK method derived from the characteristics of the Boolean algebra. The experiments with the program which is dissimilar to the previous computer programs show that the algorithm presented will be more efficient.

While the standard linear-quadratic-Gaussian problem has fixed horizons, this paper considers the LQG problem with moving horizons. By the separation principle the solution will be given by the kalman filter with the approaching horizon and the LQ regulator with the receding horizon. Sufficient conditions on weighting matrices are derived under which the filter and regulator are asymptotically stable. It wall be shown that the computation method of the moving-horizon LQG regulators is better than that of the standard LQG regulator. The performance measure between the two optimal controls will be compared. A simulation result is given in order to show the usefulness of the moving-horizon LQG regulator.s

A 16-bit computer system, called X77, is designed and constructed using bit-slice micro-slice microprogrammable. A cross-assembler/simulator software has been implemented and used to debug the entire system design including actual microcodes prior to hardware realizations. Prototypes of X77 CPU, I/O processing subsystem and their interconnections have been constructed and are in the process of being debugged.

A linear state equation of the first order differential form relating the load-frequency dynamic characteristics of interconnected power systems was derived for use in computer simulation. A now solution of the algebraic matrix riccati equation for application in quadratic optimal controllor and least-square state estimator dermination was developed. The program for a dynamic state equation for two interconnected control areas was developed. The optimized load-frequency deviation was analysed and a numerical analysis was tried based on the computer simulation. It was shown that the dynamic response of th loed-frequency could be optimized with weighting factors IR and Q. The result was that the load-frequency and the tie-line deviation were visibly reduced.

직류전동기(直流電動機) 제어(制御)를 위한 thyristor chopper 회로(回路)에 대한 연구(硏究)는 여러가지 면(面)에서 많은 학자(學者)들에 의해 활발히 진행(進行)되고 있는 실정(實情)이며 그 설계(設計) 방안(方案)에 있어도 여러 규모(規模)에서 해석(解析)되고 있다. 그러나 지금까지 소개(紹介)되어진 대부분(大部分) 문제점(問題點)들은 전류실패(轉流失敗)로 인(因)한 thyristor chopper 회로(回路)의 신뢰성(信賴性) 향선(向選)을 소홀히 취급하여온 경향(傾向)이 있었다. 그러므로 본(本) 연구(硏究)에서는 전류작용(轉流作用)의 확실(確實)을 가하도록 하는데 주안점(主眼店)을 두어 전류실패(轉流失敗) 요인(要因)을 실험적(實驗的)으로 분석(分析)하여 이를 개선(改善)할 수 있는 설계(設計) 방안(方案)을 제시(提示)하였고 이에 의하여 실험(實驗) 제작시험(製作試驗)한 결과(結果) 바람직한 특성(特性)을 얻을 수 있었다.

The purpose of this paper is to develop an efficient model for the analysis of a John's Chopper Circuit. In the John's Chopper Circuit analysis, the open branches are removed from the associated graph to formulate the modified incidence matrix. An algorithm for the generation of a modified proper tree and fundamental cut set matrix from a network graph is developed, which utilizes much less computer storage space and computation time compared to the classical methods.

전력용(電力用) 병렬(竝列) 콘덴서는 변전소(變電所)에서 설치되어 계통(系統)의 전압조정 및 역률개선에 사용되어 왔으며 콘덴서의 원-용(元-用) 범위를 송전전선(送電電線)의 직렬(直列)보상에까지 활용하게 되어 선로의 운용도(云誘導) reactance를 보상함으로서 등가적으로 선로의 길리을 단축시키게 되어 장거리, UHV 손전선(送電線) 및 배전선(配電線)에 채용(採用)되고 있다. 본(本) 연구는 각국에서 채용되고 있는 직렬(直列)콘덴서 보상방식을 우리나라의 모델 선로에 적용시켰을 경우에 대한 전압(電壓), 전류(電流) 전력(電力) 및 전력손실(電力損失) 등의 제특성에 대하여 고찰한 결과이다.

The Fourier convolution algorithms are used to reconstruct a 3-D density function from the projection data sets. The convolved data are then back projected to obtain a density function. There are several choices of the weighting function for the design of the reconstruction(deblurring) filter. Present paper reviews the design of reconstruction filter considering the problems such as the effects of sampling rate, aliasing, and noise.

Based on minimum-mean-sqare error criterion, a noise filtering algorithm for the reconstruction of an image function from noisy projection data is suggested. The filter is constructed with a few projection data. This algorithm requires less computational time compared with other noise filtering algorithm.

The degradation of image due to the finite size of sensing device has been one of the problems to all of the digital imaging systems. The improvement of the spatial resolution has been attempted by "differential method" with fixed sensor size and finer sampling. The computer simulations were carried out for the cases of PLF system (Parallel Linear Fan-beam) and SR(Stationary Ring) system and the results are presented.

A new detector array scheme for the positron camera for tile improved image resolution and a method for the resolution improvement of the undersampled projection data from ring positron camera are proposed. Computer simulation results are presented and problems ana limitations are discussed.

Two dimensional kernels which reconstruct the tomographic image from the blurred one formed by simple back-projection are investigated and their performances are compared. These kernels are derived from tile point spread function of the tomographic system and have the form of a ramp filter modified by several window functions to suppress ringing in the reconstruction. Computer simulation using a computer generated phantom image data with different correction functions(kernels) has been carried out. In this simulation, filtering in frequency domain by 2-D FFT technique or in space domain by 2-D direct convolution is considered. It is found that the-computation time required for real space convolution technique is much larger than that of Fourier 2-D filtering technique in the pratical situation.

Performances of three computational algorithms for one-dimensional frequency filtering are compared and tradeoffs are studied. If the size of the filter impulse response is small, it is well-known that the conventional convolution is superior than the FFT convolution. If the size of the impulse response is large, it was suggested that the recursive filter might be competitive in terms of speed to the FFT convolution. We, therefore, have developed an computational, algorithm for the recursive filter to compare the speed advantages over the FFT convolution and the results are presented.