• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.12a
• /
• pp.197-200
• /
• 2000

최대 수용전력 시계열 데이터를 수집하여 카오스적 성질을 분석하고 퍼지 제어기로부터 추론되어진 제어 값으로 특정 플랜트의 단기예측을 수행하는 카오스 퍼지 제어기를 구성하고 시뮬레이션을 통하여 실제 데이터와의 오차 검토를 통하여 카오스 퍼지 제어기의 강인성을 검증하고 이 시스템을 통하여 얻어진 결과와 실제 데이터를 비교함으로써 제어기의 성능을 평가한다.

• Korean Journal of Materials Research
• /
• v.10 no.5
• /
• pp.364-368
• /
• 2000

SiC fiber reinforced $Si_3N_4$ matrix composites combined with electrical conductive phases of carbon fiber and WC powder fabricated by hot pressing at 1773K. The ability to predict fracture in the ceramic matrix composites was evaluated by measuring simultaneous load-deflection and electrical resistanc difference-deflection curves in four point bending tests. The changes in electrical resistance differences closely corresponded to the fracture behavior of the composites. Different electrical conductive phases are suited to predicting different stages and rates of fracture. These obsevations how that it is possible to perform "in situ" fracture detection in ceramic composites.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2001.06a
• /
• pp.213-216
• /
• 2001

The H.263 Annex I method for the intraframe coding is based on the prediction in DCT domain, unlike JPEG, MPEG-1, and MPEG-2 where the intraframe coding uses block DCT, independent of the neighboring blocks. In this paper, we show the ineffectiveness of H.263 Annex I prediction method by mathematically deriving the spatial domain meaning of H.263 Annex I prediction method. Based on the derivation, we propose a prediction method which is based on the spatial correlation property of image signals. From the experiment and derivation, we verified the proposed method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.2C
• /
• pp.160-169
• /
• 2002

An ATM-PON (Passive Optical Network) system consists of an OLT (Optical Line Termination), multiple ONUs (Optical Network Units) and the optical fiber which has a PON (Passive Optical Network)configuration with a passive optical splitter. To avoid cell collisions on the upstream transmission, an elaborate procedure called as ranging is needed when a new ONU is installed. The ONU can send upstream cells according to the grant provided by the OLT after the procedure. To prevent collisions being generated by the variation of several factors, OLT must performs continuously the cell phase monitoring. It means that the OLT predicts the expected arrival time, monitors the actual arrival time for all upstream cells and calculates the error between the times. Accordingly, TC (Transmission Convergence) chip in the OLT needs a predictor which predicts the time that the cell will arrive for the current grant. In this paper, we implement the predictor by using shift registers of which the length is equivalent to the equalized round trip delay. As each register consists of 8 bit, OLT can identify which ONU sends what type of cell (ranging cell, user cell, idle cell, and mini-slot). Also, TC chip is designed to calculate the effective bandwidth for all ONUs by using the function of predictor. With the time simulation and the measurement of an implemented optical board, we verify the operation of the predictor.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2010.05a
• /
• pp.66-70
• /
• 2010

The performance of 75 ton liquid rocket engine combustion chamber for a space launch vehicle was predicted through firing tests at low pressure. In low pressure tests of 75 ton LRE combustor chamber, the combustion characteristic velocity of 1750 m/sec and the specific impulse of 240 sec were obtained which are higher than the low pressure performance of 30ton combustion chamber. The combustion characteristic velocity of 1770 m/sec and the specific impulse of 278 sec at design point for 75 ton LRE combustion chamber were predicted by using the low/high pressure performance correlation of 30ton LRE combustion chamber.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.3
• /
• pp.133-140
• /
• 2022

We investigate the performance of deep learning-based Korean language models on a task of predicting the score range of Korean essays written by foreign students. We construct a data set containing a total of 304 essays, which include essays discussing the criteria for choosing a job ('job'), conditions of a happy life ('happ'), relationship between money and happiness ('econ'), and definition of success ('succ'). These essays were labeled according to four letter grades (A, B, C, and D), and a total of eleven essay score range prediction experiments were conducted (i.e., five for predicting the score range of 'job' essays, five for predicting the score range of 'happiness' essays, and one for predicting the score range of mixed topic essays). Three deep learning-based Korean language models, KoBERT, KcBERT, and KR-BERT, were fine-tuned using various training data. Moreover, two traditional probabilistic machine learning classifiers, naive Bayes and logistic regression, were also evaluated. Experiment results show that deep learning-based Korean language models performed better than the two traditional classifiers, with KR-BERT performing the best with 55.83% overall average prediction accuracy. A close second was KcBERT (55.77%) followed by KoBERT (54.91%). The performances of naive Bayes and logistic regression classifiers were 52.52% and 50.28% respectively. Due to the scarcity of training data and the imbalance in class distribution, the overall prediction performance was not high for all classifiers. Moreover, the classifiers' vocabulary did not explicitly capture the error features that were helpful in correctly grading the Korean essay. By overcoming these two limitations, we expect the score range prediction performance to improve.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.319-319
• /
• 2019

기후변화로 인한 극한 기후 상황의 증가로 홍수기 홍수피해와 갈수기 가뭄피해가 심화되고 있으며, 수자원 관리에 대한 어려움이 발생하고 있다. 효율적인 수자원 관리를 위해 국내에는 약 1,8000여개의 댐을 운영하고 있으며, 댐의 유입량과 저수량을 감안하여 물을 적절하게 방류하는 것을 목적으로 한다. 그러기 위해서는 유입량이 우선적으로 확보되어야 하며, 더 나아가 유입량을 미리 예측할 수 있다면 더욱 효율적인 댐 운영이 가능할 것이다. 기존에는 수위나 유량을 예측하기 위해서는 주로 물리적 모형이 사용되어 왔으며, 물리적 모형은 매개변수 결정을 위한 많은 자료를 필요로 하고 그 과정에서 많은 불확실성을 포함하고 있기 때문에 계산 과정을 거치는 동안 다양한 오차가 반복 누적되는 단점이 있다. 반면에 시계열 데이터 예측을 위한 알고리즘 LSTM(Long Short-Term Memory)은 입력된 데이터와 출력된 데이터를 동시에 이용하여 보다 정확한 예측 값을 얻을 수 있다. 따라서 본 연구는 다목적댐의 유입유량 예측을 위해 구글에서 제공하는 딥러닝 오픈소스 라이브러리를 활용하여 LSTM모형을 구축하고 댐 유입유량을 예측하였다. 분석 자료로는 wamis에서 제공하는 용담댐의 2006년부터 2018년까지의 시간당 유입량 자료를 사용하였으며, 입력 데이터로 모형을 학습한 후 2018년의 유입량을 예측하였다. 예측 값의 정확도를 판단하기 위해 2018년의 실제 유입량 자료와 비교하였다.

• The Journal of the Acoustical Society of Korea
• /
• v.17 no.2
• /
• pp.50-55
• /
• 1998

트리코딩과 시영역 하모닉 스케일링을 결합하여 6.4 및 4.8 kbits/s급 음성부호화기 를 제안하였다. 부호화기는 완전 후방 적응적이고 또 하모닉 스케일링 때문에 저지연은 아 니다. 부호화기의 에러 성능을 향상시키기 위하여 트리코더에 새로운 적응 피치 예측기, 적 응 이득 함수, 단구간 적응 예측 알고리듬 등을 제안하였다. 새로운 코드 트리와 적응 이득 함수, 새로운 후방 적응 피치 예측기, 잡음에 강인한 단구간 적응 예측 알고리듬 등을 이상 적인 채널과 잡음의 영향을 받는 채널에 대하여 각각 그 성능을 평가하였다. 두 문장씩 쌍 으로 비교한 청취실험 결과, 6.4kbits/s coder (2-to-1 TDHS/2 bits/sample tree coding)의 음질은 6400samples/s로 표본화된 6-bit logPCM의 음질과 대등하였다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.3C
• /
• pp.102-111
• /
• 2005

In this paper, we propose an advanced hardware architecture for the fast inter prediction engine of the video coding standard MPEG-4 AVC. We describe the algorithm and derive the hardware architecture emphasizing and real time operation of the quarter_pel based motion estimation. The fast inter prediction engine is composed of block segmentation, motion estimation, motion compensation, and the fast quarter_pel calculator. The proposed architecture has been verified by ARM-interfaced emulation board using Excalibur & Virtex2 FPGA, and also by synthesis on Samsung 0.18 um CMOS technology. The synthesis result shows that the proposed hardware can operate at 62.5MHz. In this case, it can process about 88 QCIF video frames per second. The hardware is being used as a core module when implementing a complete MPEG-4 AVC video encoder chip for real-time multimedia application.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.10 no.5
• /
• pp.435-441
• /
• 2017

Detecting an inter-harmonic accurately is not easy work, because it has small magnitude, and its frequency which can be observed is not an integer multiple of fundamental frequency. In this paper, a new method using filter bank system and adaptive predictor is proposed. Filter bank system decomposes input signal to sub bands. In adaptive predictor, inter-harmonic is detected with decomposed sub band signal as input, and error signal as output. In this scheme, input-output characteristic of adaptive predictor is notch filter, as predicted harmonic is canceled in error signal, so detecting an inter-harmonic can be possible. Magnitude and frequency of detected inter-harmonic is estimated by recursive algorithm. The performances of proposed method are evaluated to sinusoidal signal model synthesized with harmonics and inter-harmonics. And validity of the method is proved as comparing the inter-harmonic detection results to MUSIC and ESPRIT.