• 한국측량학회지
• /
• 제39권5호
• /
• pp.253-263
• /
• 2021

Traffic data collected using advanced equipment are highly valuable for traffic planning and efficient road operation. However, there is a problem regarding the reliability of the analysis results due to equipment defects, errors in the data aggregation process, and missing data. Unlike other detectors installed for each vehicle lane, radar detectors can yield different error types because they detect all traffic volume in multilane two-way roads via a single installation external to the roadway. For the traffic data of a radar detector to be representative of reliable data, the error factors of the radar detector must be analyzed. This study presents a field survey of variables that may cause errors in traffic volume collection by targeting the points where radar detectors are installed. Video traffic data are used to determine the errors in traffic measured by a radar detector. This study establishes three types of radar detector traffic errors, i.e., artificial, mechanical, and complex errors. Among these types, it is difficult to determine the cause of the errors due to several complex factors. To solve this problem, this study developed a radar detector traffic volume error analysis model using a multiple linear regression model. The results indicate that the characteristics of the detector, road facilities, geometry, and other traffic environment factors affect errors in traffic volume detection.

• International journal of advanced smart convergence
• /
• 제12권2호
• /
• pp.56-66
• /
• 2023

The goal of deep learning is to extract complex features from multidimensional data use the features to create models that connect input and output. Deep learning is a process of learning nonlinear features and functions from complex data, and the user data that is employed to train deep learning models has become the focus of privacy concerns. Companies that collect user's sensitive personal information, such as users' images and voices, own this data for indefinite period of times. Users cannot delete their personal information, and they cannot limit the purposes for which the data is used. The study has designed a deep learning method that employs privacy protection technology that uses distributed collaborative learning so that multiple participants can use neural network models collaboratively without sharing the input datasets. To prevent direct leaks of personal information, participants are not shown the training datasets during the model training process, unlike traditional deep learning so that the personal information in the data can be protected. The study used a method that can selectively share subsets via an optimization algorithm that is based on modified distributed stochastic gradient descent, and the result showed that it was possible to learn with improved learning accuracy while protecting personal information.

• Computers and Concrete
• /
• 제32권6호
• /
• pp.625-637
• /
• 2023

The application of ML approaches in determining the resisting capacity of fire damaged RC columns is introduced in this paper, on the basis of analysis data driven ML modeling. Considering the characteristics of the structural behavior of fire damaged RC columns, the representative five approaches of Kernel SVM, ANN, RF, XGB and LGBM are adopted and applied. Additional partial monotonic constraints are adopted in modelling, to ensure the monotone decrease of resisting capacity in RC column with fire exposure time. Furthermore, additional suggestions are also added to mitigate the heterogeneous composition of the training data. Since the use of ML approaches will significantly reduce the computation time in determining the resisting capacity of fire damaged RC columns, which requires many complex solution procedures from the heat transfer analysis to the rigorous nonlinear analyses and their repetition with time, the introduced ML approach can more effectively be used in large complex structures with many RC members. Because of the very small amount of experimental data, the training data are analytically determined from a heat transfer analysis and a subsequent nonlinear finite element (FE) analysis, and their accuracy was previously verified through a correlation study between the numerical results and experimental data. The results obtained from the application of ML approaches show that the resisting capacity of fire damaged RC columns can effectively be predicted by ML approaches.

• Archives of Plastic Surgery
• /
• 제35권4호
• /
• pp.461-464
• /
• 2008

Purpose: Although the demand for the mammoplasty including reduction or reconstruction is remarkably increasing, the anthropometric measurement for the breast, especially about the nipple areola complex(NAC) of Korean women has not been reported recently. Therefore, the anthropometric measurement about the NAC was performed to suggest the standard size of NAC for Korean women. Methods: Two hundred and twenty five female volunteers in 20's through 50's were included for the study. Questionnaires including the diameter of NAC, the diameter, height of nipple, age, marital status, delivery and lactation history were distributed to the volunteers and collected. Results: The mean values of our study are as follows: the areola diameter is $30.93{\pm}10.07mm$, the nipple diameter is $10.21{\pm}4.14mm$ and the height of nipple is $6.54{\pm}3.74mm$. The diameter of nipple areola complex(NAC) is bigger in old ages. If the volunteers have the history of marriage, delivery and lactation, it is bigger, as well. The height of nipple closely related to individual characters except the correlation between height of nipple and age. Conclusion: It is important to have standard data for the nipple areola complex in order to have good aesthetic results of mammoplasty. Despite the importance, there are a few measurement data for the nipple areola complex(NAC) of Korean women. The result of our study is not the absolute parameter for breast surgery, however it can be used as the standard size for NAC in the Korean female during breast surgery.

• 웰빙융합연구
• /
• 제5권4호
• /
• pp.51-56
• /
• 2022

Purpose: By applying an ultrasonic mechanical device to the liquid fertilizer storage in the pig dropping treatment plant, the initial odor of the odor source is reduced, and the air dilution drainage of the complex odor is fundamentally recognized to facilitate odor treatment on the mechanical and chemical biological treatment devices at the rear. Research design, data and methodology: The odor concentration on the site boundary was measured to confirm the state of reduction. In order to prevent the spread of odor from the collection of the pig dropping treatment plant, it was measured by installing an ultrasonic generator inside the installation wall after installing the sealing wall. Results: The average value of the March and April measurement data remained close to neutral at 8.2 after 8.6 treatment before pH treatment, decreased 97.3% from 462 mg/L before SS treatment to 10.5 mg/L after treatment, and the composite odor was reduced by 85% from 20 to 3 before treatment. It was confirmed that ammonia (NH3) was reduced by 99% from 5.8 ppm to 0.09 ppm, and general bacteria were also reduced by 99% from 3,200 CFU/mL to 57 CFU/mL Conclusion: Applying the ultrasonic air ejector hybrid system and zigzag air complex module development product to resource circulation centers or sewage treatment facilities is thought to reduce inconvenience to residents due to odors caused.

• 한국건축시공학회지
• /
• 제24권3호
• /
• pp.355-362
• /
• 2024

본 연구의 주요 목표는 아파트 단지 수선유지 비용을 예측하기 위해 딥러닝 기법을 적용한 예측 모델 구축 프레임워크를 제안하는 것이다. 아파트 건물을 이상적인 상태로 관리하기 위해서는 지속적인 유지 및 시의적절한 수리가 필수적이다. 아파트 단지는 광범위한 면적, 공동 시설, 다수의 주거 동, 서비스 지역 등으로 인해 유지관리가 복잡하다. 또한, 아파트의 안전성 보장, 가치 유지 및 경제적 효율성 때문에 경제적이고 합리적인 유지보수의 중요성이 점점 커지고 있다. 그러나 아파트 단지 수선유지는 다양한 외부 요인의 영향을 받고 데이터 수집이 어려워 연구가 부족한 상황이다. 따라서 본 연구는 실제 아파트 단지 유지보수 비용 데이터를 기반으로 딥러닝 기법을 활용해 유지보수 비용을 예측하는 모델 개발 프레임워크를 제시하고자 한다. 본 연구의 프레임워크 및 결과는 실질적으로 아파트 단지의 유지보수 비용 예측에 활용될 수 있으며, 궁극적으로 아파트 단지의 시설 관리 향상에 기여할 것이다.

• 한국융합학회논문지
• /
• 제9권6호
• /
• pp.25-31
• /
• 2018

항공탑재시험은 복잡해지는 무기체계 개발과정에서 데이터 획득을 위해 수행하는 중요한 시험 중 하나이다. 본 논문은 무기체계 개발 과정중 수행하는 항공탑재시험용 POD 시스템에 대한 설계 내용 및 시험 결과를 소개한다. 항공탑재시험용 POD는 좌 우 2개의 POD가 설계 및 제작되었고 각각의 POD는 항공기 연료탱크와 동일한 외형과 질량특성을 갖도록 하여 감항인증 절차를 생략할 수 있도록 하였다. 또한 무기체계 개발에 필요한 표적 영상 데이터 계측, 항법 데이터 획득, 알고리듬 검증 및 분석에 필요한 기준 데이터를 획득할 수 있도록 구성품들을 적절하게 배치하였다. 항공탑재시험용 POD 시스템은 기계적, 전기적 요소들이 모두 반영된 복합적인 시스템이며 개발된 POD 시스템은 반복적으로 항공탑재시험에 사용되어 무기체계 개발에 필요한 다양한 데이터를 성공적으로 획득하였다.

• 한국태양에너지학회 논문집
• /
• 제34권2호
• /
• pp.82-90
• /
• 2014

The MERRA reanalysis data provided online by NASA was applied to predict the annual energy productions of two largest wind farms in Korea. The two wind farms, Gangwon wind farm and Yeongyang wind farm, are located on complex terrain. For the prediction, a commercial CFD program, WindSim, was used. The annual energy productions of the two wind farms were obtained for three separate years of MERRA data from June 2007 to May 2012, and the results were compared with the measured values listed in the CDM reports of the two wind farms. As the result, the prediction errors of six comparisons were within 9 percent when the availabilities of the wind farms were assumed to be 100 percent. Although further investigations are necessary, the MERRA reanalysis data seem useful tentatively to predict adjacent wind resources when measurement data are not available.

• 한국조사연구학회:학술대회논문집
• /
• 한국조사연구학회 2002년도 추계학술대회 발표논문집
• /
• pp.5-28
• /
• 2002

• Bulletin of the Korean Chemical Society
• /
• 제8권3호
• /
• pp.220-221
• /
• 1987