• 대한안전경영과학회지
• /
• 제12권2호
• /
• pp.161-166
• /
• 2010

EPC Information Services (EPCIS) is an EPCglobal standard designed to enable EPC-related data sharing within and across supply chain. The EPCIS standard defines standard interfaces to enable EPC-related data to be captured and subsequently to be queried using a set of service operations and an associated data model. There are two kinds of EPCIS data: event data and master data. Event data is created in the process of carrying out business processes. Traceability of goods across supply chain is based on event data. Therefore, each company must have an event data. This study compared he difference between MS-SQL(DRDBMS) and ALTIBASE(MMDBMS) for data storage. We compared the difference between two database management in many respects such as insert time and select time. We come to a conclusion that ALTIBASE is more efficient than MS-SQL.

• 한국ITS학회 논문지
• /
• 제14권6호
• /
• pp.125-132
• /
• 2015

본 논문은 무선 채널을 노드에게 할당하여 보장하는 GTS 기반의 무선 센서 네트워크에서, 고정적으로 할당하는 채널을 공유하고, 이벤트 발생 시 특정 노드가 보장 채널을 독점적으로 활용하기 위한 방안을 제시한다. 평상시에는 무선 자원이 충분하지 않은 조건에서 다수의 노드에게 특정 채널을 같이 할당하고, 할당한 채널을 노드들이 공유하면서 데이터를 전송한다. 이벤트 발생 시 특정 노드에게 채널을 독점적으로 활용하게 하게한다. 이러한 방안은 데이터 전송 시 백오프 값을 적절히 활용하고, 이벤트가 발생한 노드와, 이 노드로부터 데이터를 수집하는 루트 노드로의 경로상의 노드에게 우선권을 주는 방법으로 동작한다. 시뮬레이션을 통해 트리기반 GTS 채널을 사용하는 제안 방안이 효과적으로 백오프 값을 조절하여, 이벤트의 전달노드에게 확실한 채널을 보장하면서, 필요한 이벤트 데이터가 효과적으로 수집될 수 있음을 확인할 수 있었다.

• 한국정보통신설비학회:학술대회논문집
• /
• 한국정보통신설비학회 2009년도 정보통신설비 학술대회
• /
• pp.157-161
• /
• 2009

In recent years, there is a remarkable progress in ICTs (Information and Communication Technologies), and then many attempts to apply ICTs to other industries are being made. In the field of disaster managements, ICTs such as RFID (Radio Frequency IDentification) and USN (Ubiquitous Sensor Network) are used to provide safe environments. Actually, various types of early warning systems using USN are now widely used to monitor natural disasters such as floods, landslides and earthquakes, and also to detect human-caused disasters such as fires, explosions and collapses. These early warning systems issue alarms rapidly when a disaster is detected or an event exceeds prescribed thresholds, and furthermore deliver alarm messages to disaster managers and citizens. In general, these systems consist of a number of various sensors and measure real-time stream data, which requires an efficient and rapid data processing technique. In this study, an event-driven architecture (EDA) is presented to collect event effectively and to provide an alert rapidly. A publish/subscribe event processing method to process simple event is introduced. Additionally, a complex event processing (CEP) technique is introduced to process complex data from various sensors and to provide prompt and reasonable decision supports when many disasters happen simultaneously. A basic concept of CEP technique is presented and the advantages of the technique in disaster management are also discussed. Then, how the main processing methods of CEP such as aggregation, correlation, and filtering can be applied to disaster management is considered. Finally, an example of flood forecasting and early alarm system in which CEP is incorporated is presented It is found that the CEP based on the EDA will provide an efficient early warning method when disaster happens.

• 정보처리학회논문지D
• /
• 제14D권5호
• /
• pp.483-490
• /
• 2007

최근에 데이터 스트림을 분석하여 잠재되어 있는 지식을 발견하기 위한 마이닝 방법에 대한 연구가 진행되고 있다. 그러나 대부분의 지지도 기반의 마이닝 방법들은 일정 주기 동안에 미리 정의된 지지도 이상의 발생 빈도를 갖는 이벤트만을 고려함으로써 발생 빈도에 비하여 중요도가 높은 이벤트를 간과하는 문제점을 가지고 있다. 본 논문에서는 이상 이벤트에 대한 연관 규칙을 탐사할 수 있는 SM-AF 방법을 제안한다. SM-AF 방법은 이상 이벤트가 감지된 윈도우만 고려하여 연관 정보를 탐사함으로써 자주 발생하지 않더라도 중요도가 높은 이벤트에 대한 연관 정보를 탐사할 수 있다. 또한 SM-AF 방법은 이상 이벤트에 대한 의미 있는 희소 항목 집합과 주기적인 이벤트 집합도 탐사한다. 그리고 다양한 실험을 통하여 SM-AF 방법이 기존의 연관 규칙 방법들에 비하여 우수함을 확인하였다.

• 한국CDE학회논문집
• /
• 제17권2호
• /
• pp.97-110
• /
• 2012

High performance sensors and modern data logging technology with real-time telemetry facilitate system fault diagnosis in a very precise manner. Fault detection, isolation and identification in fault diagnosis systems are typical steps to analyze the root cause of failures. This systematic failure analysis provides not only useful clues to rectify the abnormal behaviors of a system, but also key information to redesign the current system for retrofit. The main barriers to effective failure analysis are: (i) the gathered data (event) logs are too large in general, and further (ii) they usually contain noise and redundant data that make precise analysis difficult. This paper therefore applies suitable pre-processing techniques to data reduction and feature extraction, and then converts the reduced data log into a new format of event sequence information. Finally the event sequence information is decoded to investigate the correlation between specific event patterns and various system faults. The efficiency of the developed pre-filtering procedure is examined with a terminal box data log of a marine diesel engine.

• 산업기술연구
• /
• 제27권B호
• /
• pp.235-243
• /
• 2007

Sensor network is used to obtain sensing data in various area. The interval to sense the events depends on the type of target application and the amounts of data generated by sensor nodes are not constant. Many applications exploit long sensing interval to enhance the life time of network but there are specific applications that requires very short interval to obtain fine-grained, high-precision sensing data. If the number of nodes in the network is increased and the interval to sense data is shortened, the amounts of generated data are greatly increased and this leads to increased amount of packets to transfer to the network. To transfer large amount of packets fast, it is necessary that the delay between successive packet transmissions should be minimized as possible. In Sensor network, since the Operating Systems are worked on the event driven, the Timer Event is used to transfer packets successively. However, since the transferring time of packet completely is varies very much, it is very hard to set appropriate interval. The longer the interval, the higher the delay and the shorter the delay, the larger the fail of transfer request. In this paper, we propose ESTEO which reduces the delay between successive packet transmissions by using SendDone Event which informs that a packet transmission has been completed.In ESTEO, the delay between successive packet transmissions is shortened very much since the transmission of next packet starts at the time when the transmission of previous packet has completed, irrespective of the transmission timee. Therefore ESTEO could provide high packet transmission rate given large amount of packets.

• 스마트미디어저널
• /
• 제6권3호
• /
• pp.41-48
• /
• 2017

Event detection using social media has been widespread since social network services have been an active communication channel for connecting with others, diffusing news message. Especially, the real-time characteristic of social media has created the opportunity for supporting for real-time applications/systems. Social network such as Twitter is the potential data source to explore useful information by mining messages posted by the user community. This paper proposed a novel system for temporal event detection by analyzing social data. As a result, this information can be used by first responders, decision makers, or news agents to gain insight of the situation. The proposed approach takes advantages of deep learning methods that play core techniques on the main tasks including informative data identifying from a noisy environment and temporal event detection. The former is the responsibility of Convolutional Neural Network model trained from labeled Twitter data. The latter is for event detection supported by Recurrent Neural Network module. We demonstrated our approach and experimental results on the case study of earthquake situations. Our system is more adaptive than other systems used traditional methods since deep learning enables to extract the features of data without spending lots of time constructing feature by hand. This benefit makes our approach adaptive to extend to a new context of practice. Moreover, the proposed system promised to respond to acceptable delay within several minutes that will helpful mean for supporting news channel agents or belief plan in case of disaster events.

• 전기학회논문지
• /
• 제64권9호
• /
• pp.1337-1346
• /
• 2015

In this paper, pattern classifier is designed to classify precipitation and non-precipitation events from weather radar data. The proposed classifier is based on Fuzzy Neural Network(FNN) and consists of three FNNs which operate in parallel. In the proposed network, the connection weights of the consequent part of fuzzy rules are expressed as two polynomial types such as constant or linear polynomial function, and their coefficients are learned by using Least Square Estimation(LSE). In addition, parametric as well as structural factors of the proposed classifier are optimized through Differential Evolution(DE) algorithm. After event classification between precipitation and non-precipitation echo, non-precipitation event is to get rid of all echo, while precipitation event including non-precipitation echo is to get rid of non-precipitation echo by classifier that is also based on Fuzzy Neural Network. Weather radar data obtained from meteorological office is to analysis and discuss performance of the proposed event and echo patter classifier, result of echo pattern classifier compare to QC(Quality Control) data obtained from meteorological office.

• 응용통계연구
• /
• 제27권2호
• /
• pp.277-290
• /
• 2014

본 연구에서는 로지스틱 회귀 모형을 이용하여 보은 지방의 산사태 자료를 분석하였다. 5000 지역의 관측치 가운데 단 9개만이 산사태 발생 지역이므로 이 자료는 희귀 사건 자료로 간주될 수 있다. 로지스틱 회귀 분석 모형이 희귀사건 자료에 적용될 때 주요 이슈는 회귀 계수 추정치에 심각한 편의 문제가 생길 수 있다는 것이다. 기존에 두 가지의 편의 수정 방법이 제안되었는데, 본 논문에서는 시뮬레이션을 통해 정량적으로 비교 연구를 진행하였다. Firth(1993)의 방식이 다른 방법에 비해 우수한 성능을 보였으며, 이항 희귀 사건을 분석하는 데 있어서 매우 안정된 결과를 보여주었다.

• 인터넷정보학회논문지
• /
• 제17권6호
• /
• pp.153-158
• /
• 2016

최근 소셜 네트워크 사용자들이 늘어나면서, 각 지역에서 관심 받고 있는 사회적인 이슈나 재해 등과 같은 이벤트에 대한 정보들이 소셜 미디어 사이트를 통해 실시간으로 빠르게 대량으로 게시되고 있으며, 사회적 파급효과도 매우 커지고 있다. 본 논문에서는 지역정보를 가진 트위터 데이터를 이용하여 특정 시간, 지역에 사용자들이 관심을 가지고 있는 이벤트를 탐지하는 방법을 제안하고자 한다. 이를 위해 트위터 스트리밍 API를 이용해 데이터를 수집하고, 트윗의 키워드들의 시간에 따른 빈도수를 분석하여 정상적인 패턴과 다른 패턴을 가진 키워드를 이벤트로 추출하고, 같은 이벤트에 대한 키워드들을 군집화 하기 위해 co-occurrence 그래프를 이용하여 이벤트 감지 시스템을 구현하였다. 그리고 실험을 통해 제안한 기법의 유효성을 검증한다.