http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
An Electricity Energy and Water Consumption Model for Korean Style Apartment Buildings
Dongjun Suh,Yoon-Sik Yoo,Il-Woo Lee,Seongju Chang 제어로봇시스템학회 2012 제어로봇시스템학회 국제학술대회 논문집 Vol.2012 No.10
This research focuses on building-information related energy consumption estimation modeling by analyzing domestic electricity energy and water consumption data. We identify the elements of apartment buildings which are related to electricity and water use and demonstrate energy consumption forecast based on the proposed model. To verify the patterns of electricity energy and water consumption in relation with apartment buildings, we collect data for one of the apartment complexes in South Korea. Both weather history data including temperature and relative humidity and building information such as floor number and the gross area of apartment units are selected as Building-Information Factors (BIF) to analyze their effects on energy use in apartment buildings. In this paper, we cover annual electricity energy and water consumption among major household utilities that is manipulated by BIF. We conduct Multiple Regression Analysis (MRA) and use neural network-based model such as Multiple-Layer Perceptron (MLP) not only to reveal the relation among the elements such as weather information and BIF toward energy and water consumption model but also to estimate electricity energy and water use forecast. By carrying out this research, the relationship among BIF and energy consumption data sets is turned out to be essential information for establishing a Korean Style Apartment Management Information System as well as establishing a real estate policy on housing prices according to the diversification in floor numbers and gross areas of apartment units.
OpenDKP: Designing an Open Data Knowledge Platform for Disaster Risk Reduction and Management
Dongjun Suh(서동준) 한국디지털콘텐츠학회 2019 한국디지털콘텐츠학회논문지 Vol.20 No.9
With respect to the phases of the entire disaster management cycle, such as prevention, mitigation, preparedness, response, and recovery, data interconnectivity, availability, and sharing with multidisciplinary and cross-domain approaches play a crucial role in disaster risk reduction and management. In this paper, an open data knowledge platform, which is a knowledge sharing platform for disaster data collection, management, distribution, and analysis, designed for use in interdisciplinary studies in the field of disaster risk reduction and management, is presented. The proposed system can play an important role as a knowledge hub for supporting the development of disaster response technology as well as for risk mitigation and disaster response, disaster recovery, and risk assessment studies. In this study, the architectural design of an open disaster data platform that consists of a set of modular platforms, namely an open data knowledge portal, an API gateway, a disaster knowledge-based platform, and a big data analytics platform, is proposed. The proposed platform will enable the researchers and decision makers involved in the analysis of disaster risk to integrate and collaborate on the collected disaster data and the corresponding analyses, which will allow for the highest levels of disaster risk reduction and management.
비디오 스케일러빌리티를 고려한 영상 품질 메트릭 모델링
서동준(Dongjun Suh),김천석(Cheon Seog Kim),배태면(Tae Meon Bae),노용만(Yong Man Ro) 한국방송·미디어공학회 2006 한국방송공학회 학술발표대회 논문집 Vol.2006 No.-
다양한 멀티미디어 사용 환경 조건에 적합한 컨텐츠를 제공하기 위해서는 다양한 소비 환경을 반영할 수 있는 스케일러블 비디오의 제공이 필요하다. 이러한 스케일러블 비디오의 경우, 다양한 조합의 프레임 율, SNR, 공간해상도를 가지는 비디오가 가능하므로 사용자에게 최적의 품질을 제공하는 조합을 결정해야 할 필요가 있다. 따라서 본 논문에서는 프레임율, SNR, 공간해상도 그리고 영상의 모션 속도에 따라 변동할 수 있는 영상 품질을 나타낼 수 있는 새로운 영상 품질 메트릭을 주관적 평가를 통하여 제안한다. 제안한 품질 메트릭은 주관적 품질평가 선호도와의 상관계수(correlation coefficient)가 PSNR과 주관적 품질평가 점수와의 상관계수 평균값에 비해 높은 값을 보였다. 본 논문에서 제안한 품질 메트릭은 다양한 조합의 부호화 조건에 따른 품질측정이 가능하여 제한적인 멀티미디어 소비환경에서 최적의 부호화 조건을 결정할 수 있다.
확장 가능한 영상 부호화에서 전 확장성을 고려한 품질 메트릭
서동준(Dongjun Suh),김천석(Cheon Seog Kim),배태면(Tae Meon Bae),노용만(Yong Man Ro) 한국방송·미디어공학회 2007 방송공학회논문지 Vol.12 No.2
Currently, the efficient adaptive video streaming services and coding techniques are required to ensure quality of service (QoS) in diverse multimedia environments. To maintain optimal video quality in these environments, it is required to support scalable video coding (SVC) technique and proper QoS metric methods. In this paper, we propose a new quality metric that takes in to the temporal, the SNR, the spatial scalability on SVC. Experimental results shows that proposed method has a higher correlation with the subjective quality. We expect that the proposed metric plays an important role in deciding the coding parameter to optimize the quality of the SVC based video.
비 인출식 엔진 회전수 검출을 위한 Signal Conditioning 필터의 설계
안동준(Dongjun Ahn),이동원(Tongwon Lee),김근식(Keunsik Kim),서원수(Wonsoo Suh) 한국자동차공학회 2007 한국자동차공학회 춘 추계 학술대회 논문집 Vol.- No.-
In this paper, we proposed to design strategy of signal conditioning filter systems for the detection of engine RPM data which acquired to cigar jacks for useful facility(plug & play) of fuel efficiency monitoring systems. Signal conditioning filter strategy has high and low pass filters design procedures for the extract of periodical engine RPM signals with noise-like vehicle DC power lines. Also proposed systems using FFT methods for more precious detections of engine RPM datas which decided from regularly intervals of each frequency peaks elements from FFT results. Hardware design procedures includes SMPS design techniques for the stability of systems.
Youngsang Suh,Jiseock Kang,Dongjun Lee 제어로봇시스템학회 2020 제어로봇시스템학회 국제학술대회 논문집 Vol.2020 No.10
This paper presents a new planning method in cluttered environment using a newly defined concept called a maximally occupying convex space (MOCS). Motion planning methods that use safe flight corridor (SFC) generally construct SFC in every new trial based on each initial path. We define a MOCS in a three-dimensional rectangular world to be prebuilt as SFC, prior to the initial path computation that leads to optimization process without running SFC construction phase. In the rectangular world, creating every possible MOCS covers all the free space overlapping with others if there exists a feasible path among those spaces. Thus, expressing MOCSs as nodes and their overlappings as edges, we can formulate a simplified graph representing the free space. Dijkstra algorithm is applied to a simplified graph in this work. This approach enables to omit constructing SFC every new trial, and to use path planning algorithm with small amount of nodes thereby, achieve efficient computation. We provide a completeness of a planning algorithm in a rectangular space, and show computation efficiency compared with a state-of-the-art algorithm based on SFC.