트랜스퓨터를 근간으로 하는 병렬컴퓨터 "KAPAC(Kaist PArallel Computer)"을 설계하고 구현하였다. KAPAC의 목적은 복잡하거나 많은 계산이 요구되는 일을 병렬로 처리하여 속도 향상을 시킴으로써 실시간 처리및 고성능 처리를 하는 많은 응용분야에 대한 계산능력을 제공하기 위함이다. KAPAC은 UNIX 컴퓨터를 Host로 하고 VME bus에 연결할 수 있는 후위 컴퓨터로 구현하였다. 구현된 병렬 컴퓨터는 32개의 처리소자를 가지고 있는 메세지 패싱 타입의 컴퓨터이며 크로스바 스위치를 사용하여 프로그램에 의해 쉽게 연결망 형태를 구성 할 수 있도록 하였다. 구현된 병렬 컴퓨터 "KAPAC"의 재구성 특성을 보기 위하여 구성할 수 있는 다양한 연결망들을 소개했으며 몇개의 응용 프로그램들이 각기 다른 상호 연결 위상에서 수행되었다.

Many real-time supporting features are used to execute real-time process in a time-constraints given. In this paper, The real-tiem resources management mechanism based time constraints which order resources supporting real-time processes blocked by the processes was proposed. This mechanism was all implemented and tested on the UNIX operating system.

A redundant control scheme which can maintain its tracking capability in the case of a controller failure is proposed for the industrial applications which need high reliability with fault-tolerance. It consists of two identical controllers and a switching mechanism which includes failure detection and reconfiguration algorithm. The new detection method against controller failure using fuzzy logic enables the detection of controller failures without failure assumptions through the instability of the failed controller. The failed controller is smoothly removed from the control loop by reducing time weight of the failed controller.

In this paper, a set of TM(template matching) mask using the AM(associative mapping) technique was generated and the edge detection algorithm for a SAR image was proposed. And also, the performance of the proposed edge detection algorithm was tested with the conventional edge detection techniques. The proposed edge detection algorithm created an edge image which was more accurate and clear than the conventional edge detection techniques and the performance of the proposed detection technique was not deteriorated for low intensity area in the image because the uncertainly thresholded value genetated by the conventional detection methods was requested. Also, the number of masks and the detection time were reduced by adjusting resolution of edge detection and the consideration for the threshold value extracting the edge was very intuitive.

This paper addresses the estimation of the shift vector from aerial image sequences. The conventional feature-based and area-based matching methods are simulated for determining the suitable image registration scheme. Computer simulations show that the feature-based matching schemes based on the co-occurrence matrix, autoregressive model, and edge information do not give a reliable matching for aerial image sequences which do not have a suitable statistical model or significant features. In area-based matching methods we try various similarity functions for a matching measure and discuss the factors determining the matching accuracy. To reduce the estimation error of the shift vector we propose the reference window selection scheme. We also discuss the performance of the proposed algorithm based on the simulation results.

In this paper new back propagation algorithm which adds neurons to hidden layer is proposed. this proposed algorithm is applied to the pattern recognition of written number coupled with back propagation algorithm through omitting redundant learning. Learning rate and recognition rate of the proposed algorithm are compared with those of the conventional back propagation algorithm and the back propagation through omitting redundant learning. The learning rate of proposed algorithm is 4 times as fast as the conventional back propagation algorithm and 2 times as fast as the back propagation through omitting redundant learning. The recognition rate is 96.2% in case of the conventional back propagation algorithm, 96.5% in case of the back propagation through omitting redundant learning and 97.4% in the proposed algorithm.