• 로봇학회논문지
• /
• 제17권3호
• /
• pp.273-281
• /
• 2022

In this paper, we propose a robot sound recognition system that detects various sound events. The proposed system is designed to detect various sound events in real-time by using a microphone on a robot. To get real-time performance, we use a VGG11 model which includes several convolutional neural networks with real-time normalization scheme. The VGG11 model is trained on augmented DB through 24 kinds of various environments (12 reverberation times and 2 signal to noise ratios). Additionally, based on random forest algorithm, a decision logic is also designed to generate event signals for robot applications. This logic can be used for specific classes of acoustic events with better performance than just using outputs of network model. With some experimental results, the performance of proposed sound recognition system is shown on real-time device for robots.

• Journal of Information Processing Systems
• /
• 제8권2호
• /
• pp.213-240
• /
• 2012

Safety critical systems, real time systems, and event-based systems have a complex set of events and their own interdependency, which makes them difficult to test ma Safety critic Safety critical systems, real time systems, and event-based systems have a complex set of events and their own interdependency, which makes them difficult to test manually. In order to cut down on costs, save time, and increase reliability, the model based testing approach is the best solution. Such an approach does not require applications or codes prior to generating test cases, so it leads to the early detection of faults, which helps in reducing the development time. Several model-based testing approaches have used different UML models but very few works have been reported to show the generation of test cases that use events. Test cases that use events are an apt choice for these types of systems. However, these works have considered events that happen at a user interface level in a system while other events that happen in a system are not considered. Such works have limited applications in testing the GUI of a system. In this paper, a novel model-based testing approach is presented using business events, state events, and control events that have been captured directly from requirement specifications. The proposed approach documents events in event templates and then builds an event-flow model and a fault model for a system. Test coverage criterion and an algorithm are designed using these models to generate event sequence based test scenarios and test cases. Unlike other event based approaches, our approach is able to detect the proposed faults in a system. A prototype tool is developed to automate and evaluate the applicability of the entire process. Results have shown that the proposed approach and supportive tool is able to successfully derive test scenarios and test cases from the requirement specifications of safety critical systems, real time systems, and event based systems.

• 한국지진공학회:학술대회논문집
• /
• 한국지진공학회 2003년도 춘계 학술발표회논문집
• /
• pp.12-15
• /
• 2003

Korea Institute of Geoscience and Mineral Resources(KIGAM) is operating Wonju Korea Seismic Research Station(KSRS), 29 regional seismic research stations and 5 Korea-China joint seismic stations in China. Also KIGAM is operating Korea Earthquake Monitoring System (KEMS) to archive the real-time data stream and to determine event parameters (epicenter, origin time, and magnitude) by the automatic processing and analyst review. To do this, KEMS used KIGAM's regional seismic network and other institute's network in a near real-time base. From Dec. 1, 2001 to Nov. 30, 2002, 3,827 seismic events were analyzed in a automatic processing procedure and finally 3,437 events were analyzed by analyst and archived. But problem is this event catalog includes not only natural earthquake, but also artificial events produced by the blast. More than 80 % events were concentrated in daytime and many events were concentrated in the confirmed blast sites, Pyeongyang, Pocheon, Yeongjong-do, Donghae city, etc. Because these artificial events are a major potential cause of error when estimating the seismicity of a specific region, discrimination procedure has to be developed in the first place.

• 한국인터넷방송통신학회논문지
• /
• 제12권4호
• /
• pp.57-64
• /
• 2012

게임 서버는 명시된 시간 내에 많은 클라이언트들의 요청을 처리해야 하기 때문에 다중 사용자용 실시간 게임들은 일종의 연성 실시간 시스템이다. 클라이언트 이벤트들은 게임 세계의 본질에 따라 상이한 시간 요건과 일관성 요건을 가지고 있다. 이러한 요건들은 CPU 처리시 상이한 우선순위를 유발하게 되며 이벤트들은 일관성과 우선순위 정도에 따라 여러 그룹으로 분류될 수 있다. 우선순위가 상이한 이벤트들의 시간적 요건을 충족하기 위하여 본 논문에서는 우선순위 기반 그룹 태스크 스케쥴링 정책이 제시된다. 클라이언트의 수나 클라이언트가 발생시키는 이벤트들의 수는 일시적으로 증가할 수 있다. 일시적인 과부하가 발생한 경우에 게임 서버는 우선순위가 높은 이벤트들을 우선적으로 처리하기 위하여 더 많은 CPU 대역폭을 할당할 필요가 있다. 제시된 스케쥴링 정책은 우선순위가 높은 이벤트일수록 종료시한내에 성공적으로 종료되는 수를 최대화함으로써 전체 시스템의 실시간적 성능을 향상시킬 수 있다. 이 정책의 성능은 다양한 실험을 통하여 평가되었다.

• Nuclear Engineering and Technology
• /
• 제49권2호
• /
• pp.342-348
• /
• 2017

Precursor analysis is widely used in the nuclear industry to judge the significance of events relevant to safety. However, in case of events that may damage equipment through effects that are not ordinary functional dependencies, the analysis may not always fully appreciate the potential for further evolution of the event. For fires, which are one class of such events, this paper discusses modelling challenges that need to be overcome when performing a probabilistic precursor analysis. The events used to analyze are selected from the Organisation for Economic Cooperation and Development (OECD) Fire Incidents Records Exchange (FIRE) Database.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2008년도 하계종합학술대회
• /
• pp.669-670
• /
• 2008

In this paper, we present a real-time system to detect abnormal events on gas pipes, based on the signals which are observed through the audio sensors attached on them. First, features are extracted from these signals so that they are robust to noise and invariant to the distance between a sensor and a spot at which an abnormal event like an attack on the gas pipes occurs. Then, a classifier is constructed to detect abnormal events using neural networks. It is a combination of two neural network models, a Gaussian mixture model and a multi-layer perceptron, for the reduction of miss and false alarms. The former works for miss alarm prevention and the latter for false alarm prevention. The experimental result with real data from the actual gas system shows that the proposed system is effective in detecting the dangerous events in real-time with an accuracy of 92.9%.

• 천문학회지
• /
• 제26권2호
• /
• pp.135-139
• /
• 1993

Events of eclipses as well as other major astronomical events observable in the eastern sector of Asian continent are computed and checked with astronomical records of antiquity. Particular attention was given to two types of the events recorded in remaining records of Dangun Chosun Period (DCP): (1) concentration of major planets near the constellation of Nu-Sung $(\beta\;Aries)$ and (2) a large ebb-tide. We find them most likely to have occurred in real time. i.e., when the positions of the sun, moon, and planets happen to be aligned in the most appropriate position. For solar eclipses data, however, we find among 10 solar eclipse events recorded, only 6 of them are correct up to months, implying its statistical significance is no less insignificant. We therefore conclude that the remaining history books of DCP indeed contains important astronomical records, thereby the real antiquity of the records of DCP cannot be disproved.

• 한국수자원학회논문집
• /
• 제45권6호
• /
• pp.531-544
• /
• 2012

본 연구에서는 주어진 호우사상을 베타분포로 적절히 표현하기 위한 매개변수의 결정 방법을 제시하였다. 이를 위해 첨두에 대한 고려를 추가한 경우와 그렇지 않은 경우, 실제 우량주상도 및 첨두를 중심으로 재배열된 수정 우량주상도를 사용하는 경우 등을 고려하여 그 특성을 비교 검토하였다. 본 연구는 서울 지점의 2010년 독립 호우사상과 1961~2010년 사이의 연 최대치 독립 호우사상을 대상으로 베타분포를 적용하고, 그 결과를 분석하였다. 그 결과를 정리하면 다음과 같다. 첫째, 베타분포를 이용하여 실제 호우사상의 우량주상도 형태를 유사하게 표현하기 위해서는 첨두에 대한 추가적인 고려가 필요한 것으로 확인되었다. 둘째, 첨두를 중심으로 시간강우를 좌우 순차적으로 재배열한 수정 우량주 상도를 이용하는 경우가 실제 우량주상도를 이용하는 경우보다 실제 강우의 시간분포 특성에 보다 가까운 베타분포를 유도함을 확인하였다.

• 아동학회지
• /
• 제24권1호
• /
• pp.61-76
• /
• 2003

This study investigated children's cognition of physically impossible televised events by age and types of tasks. The 67 subjects were 3-, 5-, and 7-year-old children from 1 day care center and 1 elementary school in Seoul. Children's responses were taped and transcribed. Cognition about the events were measured by recognition of impossibility, possibility of replication, and reasons for the possibility of replication of the events. Data were analysed by ANOVA(repeated measure) and $Scheff{\acute{e}}$ test. Children's cognitions about the events varied by age groups. Three-year-old children's recognition of the 3 tasks was different from that of the 5-and 7-year-old children; while 3-year-olds interpreted the events on TV as real, 5- and 7-year-olds considered them to be unreal.

• 한국컴퓨터정보학회논문지
• /
• 제21권5호
• /
• pp.91-98
• /
• 2016

We propose a method to train a model that can predict the probability of a crime being committed. CCTV data by matching criminal events are required to train the crime prediction model. However, collecting CCTV data appropriate for training is difficult. Thus, we collected actual criminal records and converted them to an appropriate format using variables by considering a crime prediction environment and the availability of real-time data collection from CCTV. In addition, we identified new specific crime types according to the characteristics of criminal events and trained and tested the prediction model by applying neural network partial least squares for each crime type. Results show a level of predictive accuracy sufficiently significant to demonstrate the applicability of CCTV to real-time crime prediction.