In this study under the distributed system for efficient distribution resource completely symmetic global scheduler was designed and implemented to obtain the general global scheduler, that is load balancing as sharing objectives. To balance the system's load efficiently each node must be designed to get right decision-making. Thus we considered computing time to estimate fault such as delay on communication network, communication period and other decision-making. Load balancing mechanism which suggested in this study was implemented in the distributed system which IBM PC/AT linked to and composed with Ethernet. The target operating system was composed of IBM PC/AT as a basic construction in which proper type of UNIX operating system were ported and communication layer chose communication type implemented from Amoeba. The method of IPC employing layered multilevel access method to avoid inefficient protocol using direct communication mode guarantees rapid response due to short ready time for IPC.

A natural language processing system usually has the demerit that its processing time is relatively long. If an interactive system makes its user kept waiting long, it can't be said to be practical. In this paper, the on-line natural language parser in which its processing coincides with the sentence's inputting is designed. Since the greater part of morpholgical and syntatic semantic analysis is already performed during the keyboard input, user can get a prompt response. Moreover, the Korean parser is implemented in multitasking environment, and it is compared with an off-line parser. The on-line parser can be considered to be efficient for its real time processing.

In this paper we develop the software simulator written in a C language for a frequency modulation synthesis and the approximate range of parameters, for a musically satisfactory timbre, obtained by using the software simulator will be applied to develop an algorithm for parameter extraction. For a frequency modulation synthesis, we also develop an algorithm for parameter extraction through waveform analysis in the time domain as well as spectrum analysis using a FFT in the frequency domain. To verify the validity of the developed algorithm as well as software simulator experimentally, we extract parameters for the several music instruments using the suggested algorithm and analyze the synthesized sound by applying the parameters to the software simulator. The evaluation of the synthesized sound is first done by listening the sound directly as a subjective testing. Secondly, to evaluate the synthesized sound objectively with an engineering sense, we compare the synthesized sound with an original one in a time domain and a frequency domain.

In this paper, three types of parameter update law which can be used in model reference adaptive control are suggested based on speed-gradient algorithm which was introduced by Fradkov. It is shown that the parameter update law which was proposed by Narendra is a special from of these laws and that proposed parameter update laws can insure the global stability under some conditions such as attainability and convexity. We also comment that the transfer function of reference model shoud be positive real for the realization of parameter update law.

In industrial multivariable plants, it is often the case that all of plant outputs are measured not simultaneously, but seriallu. While Reference 1 proposed a special type of observer (referred to as "serial-sampling" type obserrversL. which was provento be very effective in this situation, it alsopointed out that the observer may have some minor problems in case of practical applications. In this paper, it isshown that these kind of problems can be resolved by proving the existence of the solution of the simultaneous equations (35) of Reference 1.

A new fully differential sample-and-hold circuit which can effectively compensate the offset voltage of an operational amplifier and the charge injection of a MOS switch is presented. The proposed circuit shows a true sample-and-hold function without a reset period or an input-track period. The prototype fabricated using a 1.2$\mu$m double-polysilicon CMOS process occupies an area of 550$\mu$m$\times$288$\mu$m and the error of the sampled ouput is 0.056% on average for 3V input at DC.

In this paper, the multiplicative algorithm of two polynomials over finite field GF(($p^{m}$) is presented. Using the presented algorithm, the multiple-valued multiplier of the serial input-output modular structure by CCD is constructed. This multiple-valued multiplier on CCD is consisted of three operation units: the multiplicative operation unit, the modular operation unit, and the primitive irreducible polynomial operation unit. The multiplicative operation unit and the primitive irreducible operation unit are composed of the overflow gate, the inhibit gate and mod(p) adder on CCD. The modular operation unit is constructed by two mod(p) adders which are composed of the addition gate, overflow gate and the inhibit gate on CCD. The multiple-valued multiplier on CCD presented here, is simple and regular for wire routing and possesses the property of modularity. Also. it is expansible for the multiplication of two elements on finite field increasing the degree mand suitable for VLSI implementation.

DCT image coding has been well known for its compression efficiency. Its nature makes it also suitable for efficient progressive transmission and reconstruction since low frequency transform coefficients contain most of the energy of image signals. In this paper, we propose a progressive transmission technique of DCT image by non-sequential bit ordering using Lloyd-Max quantizer. And then, we will show that the Lloyd-Max quantizer can be embedded quantization. In simulation results, the proposed method show better subjective picture and MSE of the reconstructed image than the conventional zig-zag scanning transmission of transform coefficients.

In this paper, a high-resolution algorithm for detecting the orientation and position of the SMD and an algorithm for compensating the position and skew angle of the PCB are proposed. The proposed algorithm for the first topic consists of two parts. Its first part is a preprocessing step. in which corner points of the SMD are detected and they are grouped. Then the coarse angle of the principal axis is obtained by line fitting. The second part is a main processing step, in which the fuzzy Hough transform over the limited range of angles is applied to the corner points to detect precisely the orientation of the SMD. The position of the SMD is determined by using its four corner points. The proposed algorithm for the second topic is the one which detects a rotation angle and translation parameters of the PCB using a template matching method. The computer simulation shows that the parametes obtained by proposed algorithms are more precise than those by the several conventional methods considered. The proposed algorithms can be applied to the fast and accurate automatic inspection systems.

For the analysis of textured image, it requires large storage space and computation time to calculate the matrix features such as SGLDM(Spatial Gray Level Dependence Matrix). NGLDM(Neighboring Gray Level Dependence Matrix). NSGLDM(Neighboring Spatial Gray Level Dependence Matrix) and GLRLM(Gray Level Run Length Matrix). In spite of a large amount of information that each matrix contains, a set of several correlated scalar features calculated from the matrix is not sufficient to approximate it. In this paper, we propose a new classifier for textured images based on these matrices in which the projected vectors of each matrix on the meaningful directions are used as features. In the proposed method, an unknown image is classified to the class of a known image that gives the maximum similarity between the projected model vector from the known image and the vector from the unknown image. In the experiment to classify images of agricultural products, the proposed method shows good performance as much as 85-95% of correct classification ratio.

This paper proposes an implementation of DCT (Discrete Cosine Transform) and IDCT (Inverse DCT) using a fast DCT algorithm with shift and addition operations instead of multiplications Based on the proposed algorithm, a new VLSI architecture for the DCT and the IDCT is proposed. It shows modularity , regularity and capability for multiprocessing. Its performance is also simulated by a simulation software, "Compass". The results of the simulation provide the quality of decompression images, the increase in processing speed, representing the superiority of the proposed architecture.

The problems in back-propagation(BP) generally are learning speed and misclassification due to lacal minimum. In this paper, to solve these problems, the classical modified methods of BP are reviewed and an extension of the BP to compensate the sigmoide-prime function around the extremity where the actual output of a unit is close to zero or one is proposed. The proposed method is not onlu faster than the conventional methods in learning speed but has an advantage of setting variables easily because it shows good classification results over the vast and uncharted space about the variations of learning rate, etc.. And it is simple for hardware implementation.

This paper describes on the analysis of morphemes and it's synthesis being constituted of Korean word phrases. To analyze morphemes, we propose the introduction of "morph" for morpheme features in lexicon and the usage of chart data structures. it controls over the generation of unnecessary morpheme, and extracts every possible morpheme unit in a word phrase which minimized lexicon investigation by using heuristic information. Moreover, to synthesize morphemes, it is composed of every possible analyzed morphemes in word phrases to take advantage of speech and union information which can be obtained for program. Therefore, the systhesis of analyzed morphemes were designed to aid a syntactic analysis next step of natural language processing. This system for analyzing and systhesizing morpheme was to generate a word phrase by unifying syntactic and semantic features of analyzed morphemes in lexicon, and then established by C language of the personal computer.

온라인 필기 문자 인식 시스템 'SCRIPT(Symbol/Character Recognition In Pen-based Technology)'는 조합 가능한 모든 한글과 영어 대문자, 숫자, 그리고 키보드 부호 등 자연스럽게 필기되는 정자체 문자를 인식하기 위한 알고리듬이다. 필기 문자는 동일인이 쓰더라도 형태의 변화가 다양해서 정보의 불확실성을 지니게 된다. 그런데 기존의 결정 트리(decision tree)를 이용한 특징 분석 방법(feature analysis approach)은 효율적이지만 필기의 변형에 약하여 잘못된 선택을 하기 쉽기 때문에, 이러한 단점을 보완할 수 있는 방법이 필요하다. 이 논문에서는 패턴의 계층적(hierarchical)특성에 맞추어 획 자체의 형태와 획간의 위치 관계를 파악하기 위한 두 단계의 퍼지 결정 트리(fuzzy decision trees)를 사용하여 문자 패턴의 특징을 분석하는 방법을 제안한다. 이러한 방법은 다양한 가능성을 저장함으로써 형태의 변형에 강하고 이전의 잘못된 선택을 수정하기 쉬우며, 특히 하위 후보 패턴들에 의한 상위 패턴의 인식률 상승 효과가 매우 크다. 실헌 결과, 한글은 약 91%의 인식률과 약 0.33초의 인식 속도를 나타냈으며, 영어 및 기타 문자는 약 95%의 인식률과 약 0.08초의 인식 속도를 보였다. 이는 퍼지 결정 트리를 적용하지 않은 겨우에 비하여 인식률이 8~18% 정도 향상된 것이다.

In this paper, we could find optimum quantum sequence(OQS) to minimize the output ripple voltage of the quantum series resonant converter(QSRC). This sequence control is so general that it is irrelevant to the voltage gain so far as it is operating in the continuous conduction mode(CCM). Further more the dynamic range of QSRC is much extended by the optimum quantum sequence control(OQSC). Througuout the time-domain analysis, the solution of steady state and the boundary condition between continuous and discontinuous mode is QSRC is obtained. This feature is verified by simulations and experiments with good agreements.