• 전자공학회논문지B
• /
• 제30B권4호
• /
• pp.1-10
• /
• 1993

In this paper, a new method of phoneme segmentation of handwritten Korean characters using the local graph pattern is proposed. At first, thinning was performed before extracting features. End-point, inflexion-point, branch-point and cross-point were extracted as features. Using these features and the angular relations between these features, local graph pattern was made. When local graph pattern is made, the of strokes is investigated on contacting point. From this process, pattern is simplified as contacting pattern of the basic form and the contacting form we must take into account can be restricted within fixed region, 4therefore phoneme segmentation not influenced by characters form and any other contact in a single character is performed as matching this local graph pattern with base patterns searched ahead. This experiments with 540 characters have been conducted. From the result of this experiment, it is shown that phoneme segmentation is independent of characters form and other contact in a single character to obtain a correct segmentation rate of 95%, manages it efficiently to reduce the time spent in lock operation when the lock.

• 정보처리학회논문지D
• /
• 제10D권2호
• /
• pp.195-210
• /
• 2003

이 논문에서는 다차원 색인구조의 동시성을 최대화하는 동시성제어 알고리즘과 이를 위한 회복기법을 제안한다. 다차원 색인구조에서 동시성을 저하는 가장 큰 요인은 MBR 변경연산과 분할 연산이다. 제안하는 알고리즘은 먼저 MBR 변경연산으로 인한 동시성 저하를 최소화하기 위해서 PLC(Partial Lock Coupling) 기법을 제안한다. 이 기법은 기존 방법에 비해 잠금결합을 사용하는 회수를 크게 줄여 동시성을 높인다. 또한, MBR 변경의 수행 중에도 탐색자들이 해당 노드를 접근할 수 있도록 하는 MBR 변경 방법을 제안한다. 분할로 인한 동시성 저하를 해결하기 위해서 노드 분할로 인한 탐색자의 지연 시간을 최소화 할 수 있는 새로운 분할방법을 제안한다. 제안하는 알고리즘을 BADA-4 DBMS의 저장시스템인 MiDAS-3에서 구현하여 성능평가를 수행한다. 다양한 실험을 통해 제안하는 방법이 기존 방법보다 우수함을 보인다. 마지막으로, 이 논문에서는 제안하는 동시성제어 방법에 적절한 회복기법을 제안한다. 회복기법은 동시성을 최대한 보장할 수 있도록 설계되었으며 빠른 회복시간을 보장한다.

• 한국컴퓨터정보학회논문지
• /
• 제26권9호
• /
• pp.1-12
• /
• 2021

본 논문은 GPU 프로그램의 메모리의 상태 및 접근사건과 배리어 위치 정보를 감시하고, 실제 메모리 쓰기를 배리어 영역 종료 직전에 발생시키거나 배리어 영역을 재수행시켜 원자성 위배를 수행 중에 수리하는 도구인 ARCAV (Atomatic Recovery of CUDA Atomicity violation)를 제시한다. 기존의 연구들은 Lock과 Sleep 명령어를 사용하여 원자성 위배를 진단 및 수리하도록 구현되었기 때문에 지원되는 명령어와 동기화 기법이 CPU (Central Processing Unit) 프로그램과 다른 GPU (Graphics Processing Unit) 프로그램에는 적용될 수 없었고, GPU 프로그램에서는 원자성 위배의 탐지에 대한 연구만 수행되었다. 제안하는 ARCAV는 GPU 프로그램의 실행모델에서 수행될 수 있도록 설계되어 스레드 구성과 메모리 계층에 무관하게 실세계에서 발생한 원자성 위배를 대표하는 네 가지 패턴의 원자성 위배를 실시간으로 탐지하고 수리할 수 있다. 실험 결과 동시에 실행되는 스레드 개수와 구성에 무관하게 일정한 오버헤드를 보였다. 원자성 위배를 프로그램 수행 중에 실시간으로 탐지하고 수리하기 위해 소요되는 오버헤드는 네 개의 실세계 GPU 커널에서 실험되었고, 원본 프로그램 대비 평균 2.1배의 수행시간으로 동작하였다.

• 제어로봇시스템학회논문지
• /
• 제4권4호
• /
• pp.525-533
• /
• 1998

A conventional contact type brake system which uses a hydraulic system has mny Problems such as time delay response due to pressure build-up, brake pad wear due to contact movement, bulky size, and low braking performance in high speed region. As vehicle speed increases, a more powerful brake system is required to ensure vehicle safety and reliability. In this work, a contactless brake system of an eddy current type is proposed to overcome problems. Optimal torque control which minimizes a braking distance is investigated with a scaled-down model of an eddy current type brake. It is possible to realize optimal torque control when a maximum friction coefficient (or desired slip ratio) corresponding to road condition is maintained. Braking force analysis for a scaled-down model is done theoretically and experimentally compensated. To accomplish optimal torque control of an eddy current type brake system, a sliding mode control technique which is, one of the robust nonlinear control technique is developed. Robustness of the sliding mode controller is verified by investigating the braking performance when friction coefficient is varied. Simulation and experimental results will be presented to show that it has superior performance compared to the conventional method.

• 한국정보과학회논문지:데이타베이스
• /
• 제28권3호
• /
• pp.512-527
• /
• 2001

클라이언트-서버 공간 데이터베이스에서는 대용량의 공간 데이터를 네트웍을 통해 접근하는 과부하를 피하기 위해 캐쉬를 사용한다. 이 논문은 클라이언트에서 지도를 수정하는 변경트랜잭션들의 동시성 지원과 일관성 제어 문제를 다룬다. 지도를 수정하는 클라이언트 트랜잭션은 화면을 통해 대화식으로 진행되는 긴 트랜잭션이며, 변경 대상인 공간 객체는 공간 관련성에 의한 종속성을 가지는 특징이 있다. 또한, 캐쉬의 동적 중복에 대한 일관성을 제어하기 위해서는 변경 후 서버 뿐만 아니라 다른 클라이언트들로 전파가 필요하며, 이러한 변경 전파로 인한 통신 부하는 캐쉬의 이점을 잃지 않도록 최소화 되어야한다. 이 논문은 CR 잠금과 CX 잠금을 이용한 캐쉬 영역 잠금법을 제시하여 공간 관련성에 의한 변경 종속성을 해결한다. 또한, CS 잠금 및 COD 잠금을 제시하여 낙관적인 탐지기반 기법 으로 일관성 제어를 지원함으로써 클라이언트들의 메시지 부하 최소화를 지원한다. 그리고, 이러한 확장된 잠금을 적용한 공간 관련성 기반 2PC 프로토콜을 통하여 긴 트랜잭션들 간 의 정확한 지도 수정을 보장한다.

• 산업경영시스템학회지
• /
• 제31권1호
• /
• pp.74-82
• /
• 2008

The objective of this paper is to remove noises of image based on the heuristic noises filter and to extract a tumor region by using morphology techniques in breast ultrasound image. Similar objective studies have been conducted based on ultrasound image of high resolution. As a result, efficiency of noise removal is not fine enough for low resolution image. Moreover, when ultrasound image has multiple tumors, the extraction of ROI (Region Of Interest) is not accomplished or processed by a manual selection. In this paper, our method is done 4 kinds of process for noises removal and the extraction of ROI for solving problems of restrictive automated segmentation. First process is that pixel value is acquired as matrix type. Second process is a image preprocessing phase that is aimed to maximize a contrast of image and prevent a leak of personal information. In next process, the heuristic noise filter that is based on opinion of medical specialist is applied to remove noises. The last process is to extract a tumor region by using morphology techniques. As a result, the noise is effectively eliminated in all images and a extraction of tumor regions is possible though one ultrasound image has several tumors.

• 한국통신학회논문지
• /
• 제34권3C호
• /
• pp.329-340
• /
• 2009

본 논문에서는 이진천이반송파 (binary offset carrier, BOC) 변조된 확산신호의 부호동기추적에 적합한 기법을 제안한다. 이진천이반송파 변조된 확산신호의 상관함수는 정확한 상관옵셋에서 나타나는 주첨두와 여러 부첨두를 갖고 있다. 다중 경로 채널환경에서 가시신호에 비해 다중경로신호는 항상 지연되어 수신되기 때문에 여러 부침두 가운데 정확한 상관옵셋보다 이른 상관옵셋 영역에 있는 상관함수는 늦은 상관옵셋영역에 있는 상관함수에 비해 덜 왜곡된다. 본 논문에서는 이 특징을 활용해 BOC(n,n) 변조된 확산신호의 부호추적에 적합한 이른영역상관(advanced region correlation, ARC) 기법을 제안한다. 그리고 모의실험을 통해 안정된 다중경로 (static multipath) 채널과 지상이동체위성시스템 (land mobile satellite system) 채널에서 기존과 제안한 기법의 부호동기 추적성능을 비교한다. 또한 BOC(pn,n) 변조된 확산신호에 적합한 이른영역상관 기법의 확장 형태인 일반화된 이른영역상관기법 (generalized-ARC, G-ARC) 제안하고, 모의실험을 통해 그 성능을 보인다.

• 한국운동역학회지
• /
• 제30권1호
• /
• pp.7-16
• /
• 2020

Objective: It is to find factors related to stability through analysis of plantar pressure factors according to the level of instability when performing Snatch. Method: Foot pressure analysis was performed while 10 weightlifters performed 80% of the highest level of Snatch, and motion was classified and analyzed in 3 grades according to the level of instability. Results: First, in Bad Motion, the movement distance of the pressure center in the direction of ML and AP was larger significantly in Phase 2. Second, in Phase 2, the number of zero-crossing in the AP direction was larger statistically significantly in Good Motion. Third, in the bad motion in Phase 3, the number of zero-crossing in the ML direction showed a significantly larger value. Fourth, in Phase 4, it was found that the more stable the lock out motion, the greater the activity of foot controlling in the left and right directions. Fifth, Phase 3, the greater the Maximum/Mean foot pressure value, the more stable the pulling action. Sixth, in Phase 2, the foot pressure was concentrated with a wide distribution in the midfoot and rearfoot. Seventh, the triggering number of the forefoot region was small in the last pull phase. Eighth, the number of triggers in the toe area was significantly higher during Good Motion in Phase 4. Conclusion: Summarizing the factors of instability in Snatch, there was no significant difference in Phase 1 for each condition. In order to enhance the stability in Phase 2, the sensory control ability in the AP direction is required, and focusing the foot pressing motion with a wide distribution in the middle and rear parts increases the instability. In Phase 3, it was found that the more unstable, the more sensory control activity was performed in the ML direction, the stronger the forefoot pressing action should be performed for a stable Snatch. In Phase 4, It is important that the feet sensory control activity in ML directions and the control ability of the toes in order to have stable Lock out motion.

• 기술혁신연구
• /
• 제13권2호
• /
• pp.71-95
• /
• 2005

This study aims to explore and understand the role of return migrants in the regional innovation system of a transition economy (China) by analyzing the activities of returning entrepreneurs in two emerging high-tech industries in Shanghai. The empirical analysis is based on in-depth interviews with founders of high-tech companies and experts in Shanghai. The results of the analysis reveal that return migrants are a significant factor for the Shanghai innovation system, which is presently in a transition from a former manufacturing site to a metropolitan region comprising a range of industries (including high-tech) and services. First of all, return migrants are important for the Shanghai RIS in terms of numbers. Second, they engage in activities in the medium range of high-tech which reflects prevailing weaknesses of the framework conditions for innovation in Shanghai. However, due to their international background, returning entrepreneurs are able to overcome these weaknesses, and thus contribute to the development of high-tech industries in Shanghai and to a reduction of the technological lock-in.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.1958-1961
• /
• 2004

Ring Laser Gyroscopes used as navigational sensors inherently experience a lock-in region, where very low rotational rates are not measurable. Most RLG manufacturers use a mechanical dither motor that applies a small oscillatory rotational motion larger than this region to resolve this problem. Any input acceleration that bends this dithering axis causes flexure error, which is a noncommutative error that can not be compensated by simply using integrated gyro sensor output. This paper introduces noncommutative error equations that define attitude errors caused by flexure errors. In this paper, flexure error is classified as sensor level error if the sensing axis coincides with the dithering axis and as system level error if the two axes do not coincide. The relationship between gyro output and the rotation vector is introduced and is used to define the coordinate transformation matrix and angular motion. Equations are derived for both sensor level and system level flexure error analysis. These equations show that RLG based INS attitude error caused by flexure is directly proportional to time, amount of input acceleration and the dynamic frequency of the vehicle.