• 한국안전학회지
• /
• 제37권5호
• /
• pp.42-55
• /
• 2022

In this study, an accident at an in-house maintenance subcontractor of a manufacturing company was analyzed using representative systemic analysis methods, and the results were compared to determine the socio-technical and organizational structure causal factors. Systemic accident analyses were performed using AcciMap, STAMP-CAST, and a method that utilizes work processing procedures. The causal factors derived from the three methods were classified according to HFACS classification criteria. AcciMap and STAMP-CAST analyses were able to derive legal problems and defects in organizational structure between the company and the subcontractors. The method that utilized the work processing procedures drew the most causal factors of the three methods but showed some limitations in deriving legal and facility-related problems. Most of the causal factors identified through the systemic methods could be classified according to the HFACS classification criteria, except for the legal and organizational structure matters. Socio-technical and organizational problems with a holistic perspective of the company and subcontractors could be found using systemic analysis methods. However, it is necessary to conduct analysis using various methods in order to derive more comprehensive measures to prevent accidents because each analysis method showed some limitations in the derivation or expression of some causal factors. The results of this study can be helpful in selecting and using an appropriate method for accident analysis.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제53권12호
• /
• pp.804-810
• /
• 2004

There are two important things for the general purpose neural network processor. The first is a capability to build various structures of neural network, and the second is to be able to support suitable learning method for that neural network. Some way to process various learning algorithms is required for on-chip learning, because the more neural network types are to be handled, the more learning methods need to be built into. In this paper, an improved hardware structure is proposed to compute various kinds of learning algorithms flexibly. The hardware structure is based on the existing modular neural network structure. It doesn't need to add a new circuit or a new program for the learning process. It is shown that rearrangements of the existing processing elements can produce several neural network learning modules. The performance and utilization of this module are analyzed by comparing with other neural network chips.

• 한국멀티미디어학회논문지
• /
• 제9권6호
• /
• pp.720-727
• /
• 2006

In this paper, we propose a method which allows the SuperSQL query processor to share as much code as possible among various, generators, each of which is responsible for the output of a certain medium. SuperSQL is an enhanced query-processing system that converts database records into a variety of formats such as XML, HTML, PDF and etc. However, the existing SuperSQL media generator would require creation of a different processor for each medium, causing duplicated development cost. This research makes three main contributions: First, it analyzes the structures of various media, examining any possibility of integration based on their common structure. Second, it also facilitates the addition of a new output media generator by separating constructors and decorators from each medium. Last, it provides an integrated user interface to each media by method of the Media Abstraction Table Concept. We also show the performance and feasibility of our system using experimental results.

• 한국컴퓨터정보학회논문지
• /
• 제21권12호
• /
• pp.107-114
• /
• 2016

Web system helps high-performance processing for big-data analysis and practical use to make various information using IT resources. The government have started the RPS system in 2012. The system invigorates the electricity production as using renewable energy equipment. The government operates system gathered big-data with various related information system data and the system users are distributed geographically. The companies have to fulfill the system, are available to purchase the REC to other electricity generation company sellers to procure REC for their duty volumes. The REC market operates single auction methods with users a competitive price. But the price have the large variation with various user trading strategy and sellers situations. This papler proposed RPS system modeling and simulation in web environment that is modeled in geographically distributed computing environment for web user with DEVS W/S. Web simulation system base on web service helps to analysis correlation and variables that act on trading price and volume within RPS big-data and the analysis can be forecast REC price.

• 소성∙가공
• /
• 제28권2호
• /
• pp.71-76
• /
• 2019

Corrosion resistance is one of the essential properties required in steel bolts. The various types of coatings are used to improve the corrosion resistance of steel bolts. But, roll formed, subsequently Zn alloy electrodeposited and top coated steel bolt easily takes place the white storage stain or white rust under high humidity condition. To investigate the corrosion resistance of roll formed and subsequently the various types of coated steel bolts, their polarization curves were measured by potentio-dynamic tester in this study. Based on the measured polarization curves, the more times of chromate and top coating on roll formed steel bolt, the higher corrosion resistance was shown. The roll formed steel bolt, which was Zn-Ni electro-deposited, two times chromated, one time inorganic top coated, one time organic top coated and annealed, showed the best corrosion resistance.

• 한국멀티미디어학회논문지
• /
• 제20권8호
• /
• pp.1379-1386
• /
• 2017

As the Internet of Things evolves, various IoT services are provided. IFTTT is an abbreviation for If This Then That and refers to a service that links different web-based services. This paper proposes a system that generates and manages rules that combine the possibility of IFTTT service and the real-time event processing according to the concept of IoT service. Conventional database-based data processing methods are burdened to process a lot of data of IoT devices coming in real-time. The IoT device's data can be classified into formal data such as the amount of power, temperature value and position information, and informal data such as video or image data. Thus, this system classifies the data stream of IoT devices coming in real-time using the CEP engine Esper into a file signature table, classifies the formal/informal data, and shows the condition of the device data defined by the user and the service to be provided by applying the service.

• 한국IT서비스학회지
• /
• 제13권4호
• /
• pp.231-243
• /
• 2014

Of the various data that corporations can approach, web log data are important data that correspond to data analysis to implement customer relations management strategies. As the volume of approachable data has increased exponentially due to the Internet and popularization of smart phone, web log data have also increased a lot. As a result, it has become difficult to expand storage to process large amounts of web logs data flexibly and extremely hard to implement a system capable of categorizing, analyzing, and processing web log data accumulated over a long period of time. This study thus set out to apply Hadoop, a distributed processing system that had recently come into the spotlight for its capacity of processing large volumes of data, and propose an efficient analysis plan for large amounts of web log. The study checked the forms of web log by the effective web log collection methods and the web log levels by using Hadoop and proposed analysis techniques and Hadoop organization designs accordingly. The present study resolved the difficulty with processing large amounts of web log data and proposed the activity patterns of users through web log analysis, thus demonstrating its advantages as a new means of marketing.

• 한국산학기술학회논문지
• /
• 제11권12호
• /
• pp.5042-5048
• /
• 2010

유비쿼터스 환경에서 다양한 센서로부터 생성되는 데이터 스트림에 대해 사용자가 요구하는 위치 기반 서비스를 효과적으로 지원하기 위해 연산자 네트워크와 공유 등을 통한 연속 질의 처리 기법이 활용되고 있다. 위치 정보 기반의 연속 질의 처리는 특정 영역을 중심으로 유사 질의가 집중적으로 발생할 수 있으므로, 본 논문에서는 유사한 조건을 갖는 공간 연산을 공유하기 위해 그리드 영역 분할을 이용한 공간 연산 공유 기법을 제안한다. 제안기법은 공간 연산을 직접 공유하지 않고 그리드 셀 별로 이동객체를 공유하기 때문에 유사한 조건을 갖는 공간 연산의 공유가 가능하여 공간 연산의 실행 빈도수를 크게 감소시켜 질의 처리 속도의 향상과 함께 메모리 이용률을 향상시킨다.

• 한국정보처리학회논문지
• /
• 제5권3호
• /
• pp.824-832
• /
• 1998

본 논문에서는 영상의 코너점을 추출하기 위한 방법을 제안한다. 먼저 윤곽선을 따라 경계화소의 방향각을 lookup table의 참조에 의해 구하고, 기준치 이상의 화소를 후보점으로 선정한 뒤, 면적편차를 이용하여 코너점을 추출한다. 본 방법은 실시간 처리를 위해 처리시간을 줄이는 동시에 디지털 영상을 표현하는데 있어 라운드 오프 오차에 의해 발생되는 한 화소차를 극복하는 장점을 가진다. 제안된 방법의 검증을 위해 각종 실험 패턴을 이용하여 기존의 방법과 비교분석을 하였다.

• 융합신호처리학회논문지
• /
• 제16권1호
• /
• pp.1-8
• /
• 2015

In many image processing tasks, connected component labeling (CCL) is performed to extract regions of interest. CCL was usually done in a sequential fashion when image resolution was relatively low and there are small number of input channels. As image resolution gets higher up to HD or Full HD and as the number of input channels increases, sequential CCL is too time-consuming to be used in real time applications. To cope with this situation, parallel CCL framework was introduced where multiple cores are utilized simultaneously. Several parallel CCL methods have been proposed in the literature. Among them are NSZ label equivalence (NSZ-LE) method[1], modified 8 directional label selection (M8DLS) method[2], and HYBRID1 method[3]. Soh [3] showed that HYBRID1 outperforms NSZ-LE and M8DLS, and argued that HYBRID1 is by far the best. In this paper we propose an improved hybrid parallel CCL algorithm termed as HYBRID2 that hybridizes M8DLS with label backtracking (LB) and show that it runs around 20% faster than HYBRID1 for various kinds of images.