• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.20 no.3
• /
• pp.6-11
• /
• 2006

This paper describes the novel control techniques design of VSC(Voltage Sag Corrector) for the purpose of power line quality enhancement. A fast detecting technique of voltage sag is implemented through the detection of instantaneous value on synchronous rotating dq-reference frame. The first order digital filter is added in the detection algorithm to protect the insensitive characteristics against the noise. The relationship between the total detection time and cut-off frequency of the filter is described. The size of the capacitor bank used as the energy storage element is designed from the point of view of input/output energy with circuit analysis. Finally, the validity of the proposed scheme is proven through the simulated results.

• Journal of Korea Multimedia Society
• /
• v.20 no.1
• /
• pp.74-81
• /
• 2017

In learning stage of the preliminary Cut-off indoor positioning scheme, RSSI and UUID data received from beacons at each reference point(RP) are stored in fingerprint map. The fingerprint map and real-time beacon information are compared to identify the nearest K reference points through which the user position is estimated. If the number of K is zero, this scheme cannot estimate user position. We have improved the preliminary Cut-off scheme to get the estimated user position even in the case. The improved scheme excludes the beacon of the weakest signal received by user mobile device and identifies neighborhood reference points using the other beacon information. This procedure are performed repetitively until K > 0. The simulation results confirm that the proposed scheme outperforms K-Nearest-Neighbor (KNN), Cluster KNN and the conventional Cut-off scheme in terms of accuracy while the constraints are guaranteed to be satisfied.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.491-496
• /
• 2023

사전학습 언어모델을 개선하여 고품질의 문장 표현(sentence representation)을 도출하기 위한 다양한 대조 학습 방법에 대한 연구가 진행되고 있다. 그러나, 대부분의 대조학습 방법들은 문장 쌍의 관계만을 고려하며, 문장 간의 유사 정도를 파악하는데는 한계가 있어서 근본적인 대조 학습 목표를 저해하였다. 이에 최근 삼중항 손실 (triplet loss) 함수를 도입하여 문장의 상대적 유사성을 파악하여 대조학습의 성능을 개선한 연구들이 제안되었다. 그러나 많은 연구들이 영어를 기반으로한 사전학습 언어모델을 대상으로 하였으며, 한국어 기반의 비지도 대조학습에 대한 삼중항 손실 함수의 실효성 검증 및 분석은 여전히 부족한 실정이다. 본 논문에서는 이러한 방법론이 한국어 비지도 대조학습에서도 유효한지 면밀히 검증하였으며, 다양한 평가 지표를 통해 해당 방법론의 타당성을 확인하였다. 본 논문의 결과가 향후 한국어 문장 표현 연구 발전에 기여하기를 기대한다.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.3
• /
• pp.145-154
• /
• 2024

In order to operate predictive models utilizing administrative big data, it is crucial to consider policy changes and the characteristics of highly volatile data. Considering this scenario, this study proposes the Cut-off Voting Classifier (CVC) algorithm. This proposed algorithm prevents a sharp decline in accuracy by utilizing multiple weak classifiers. The study validates the proposed algorithm's performance through experiments. The performance evaluation demonstrates the ability to maintain stable prediction rates even in situations with a sharp decline in predictive model accuracy.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.31 no.1
• /
• pp.47-52
• /
• 2018

In this study, a collimator composed of multi-dielectric structures is designed using the phase field design method, a kind of topology optimization methods. It is also purposed to improve the mechanical-structural performance of a collimator by replacing previously used air regions with another dielectric material. Polypropylene and paraffin are selected as the dielectric materials for the design process taking manufacturability into account. The design objective is formulated by integrating the intensity of the electromagnetic field in the pre-determined target area to realize the collimating performance. The model for accurate numerical analysis was derived from the final result obtained from the design process through the simple cut-off method and it shows the improved performance of 105% compared with the free space wave propagation. For the designed model, the possibility of reverse transformation, the mechanical durability evaluation under the compression load, and the electromagnetic performance in the X-band range were also evaluated.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.25 no.4
• /
• pp.287-293
• /
• 2005

Ultrasonic guided waves are gaining increasing attention for the inspection of platelike and rodlike structures. At the same time, inspection methods that do not require contact with the test piece are being developed for advanced applications. This paper capitalizes on recent advances in the areas of guided wave ultrasonics and noncontact ultrasonics to demonstrate a superior method for the nondestructive detection of defects thinning simulating hidden corrosion in thin aluminum plates. The proposed approach uses EMAT(electro-magnetic acoustic transducer) for the noncontact generation and detection of guided waves. Interesting features in the dispersive behavior of selected guided modes are used for the detection of plate thinning. It is shown that mode cutoff measurements provide a qualitative detection of defects thinning. Measurement of the mode group velocity can be also used to quantify depth thinning.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.8
• /
• pp.1393-1400
• /
• 2006

This paper optimizes the access probabilities (APs) in a network resource allocation scheme based on access probabilities in order that the waiting time and the blocking probability are minimized under the given constraints, and obtains its performance. In order to optimize APs, an infinite number of balance equations is reduced to a finite number of balance equations by applying Neuts matrix geometric method. And the nonlinear programming problem is converted into a linear programming problem. As a numerical example, the performance measures of waiting time and blocking probability for optimal access probabilities and the maximum utilization under the given constraints are obtained. And it is shown that the scheme with optimal APs gives more performance build-up than the strategy without optimization.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.2
• /
• pp.263-268
• /
• 2003

In this paper, we have investigated double gate (DG) MOSFET structure, which has main gate (NG) and two side gates (SG). We know that optimum side gate voltage for each side gate length is about 3V in the main gate 50nm. Also, we know that optimum side gate length for each for main gate length is about 70nm. DG MOSFET shows a small threshold voltage roll-off. From the I-V characteristics, we obtained IDsat=550$mutextrm{A}$/${\mu}{\textrm}{m}$ at VMG=VDS=1.5V and VSG=3.0V for DG MOSFET with the main gate length of 50nm and the side gate length of 70nm. The subthreshold slope is 86.2㎷/decade, transconductance is 114$mutextrm{A}$/${\mu}{\textrm}{m}$ and DIBL (Drain Induced Barrier Lowering) is 43.37㎷. Then, we have investigated the advantage of this structure for the application to multi-input NAND gate logic. Then, we have obtained very high cut-off frequency of 41.4GHz in the DG MOSFET.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.30 no.1
• /
• pp.17-22
• /
• 2017

In this study, we designed the metallic reception part of a patch antenna using the phase field design method. The design object function is formulated with the S-parameter value which represent the return loss so that it is targeted to maximize radiation efficiency at a target frequency. The initial model of a patch antenna was designed via the ordinary theory based approach and its performance was enhanced by changing the structural configuration of the metallic part using the phase field design method combined with the double well potential functions. The final shape was proposed by removing the gray scale area along the structural boundary by employing a cut-off method. The proposed shape shows that the radiation efficiency at target frequency is significantly improved compared with the initial patch shape. The finite element analysis and optimization precess was performed using the commercial package COMSOL and Matlab programming.

• Journal of the Korean Society of Radiology
• /
• v.9 no.7
• /
• pp.515-521
• /
• 2015

This study assessed the degradation of image quality caused by grid artifacts and $moir{\acute{e}}$ pattern artifacts in a stationary grid, and the degradation of image quality caused by cut off artifacts in a moving grid. X-ray images were acquired in a stationary grid and a moving grid with X-ray exposure conditions of 100 cm, 80 kVp, and 30 mA using a CDRAD phantom and a 24 cm thickness acrylic phantom. Observer's perception of X-ray imaging using CDRAD Analyzer was mean 49.36, standard deviation 3.76, maximum 55.56, and minimum 38.67 in the stationary grid, and 47.04, 12.69, 55.56, and 20.89, respectively, in the moving grid. The stationary grid was superior to the moving grid in terms of the mean and standard deviation of observer's perception.