• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09b
• /
• pp.71-75
• /
• 2003

오류 역전파 학습 알고리즘을 이용하여 영상 인식에 적용 할 경우에는 은닉층의 노드 수를 경험적으로 설정하므로, 학습시간과 지역최소화 및 정체현상이 발생한다. 그리고 ARTI 알고리즘은 입력 패턴과 저장 패턴간의 측정 방법인 유사성 검증 방법과 경계 변수의 설정에 따라 인식률이 좌우된다. 경계 변수의 값이 크면 입력 패턴과 저장 패턴사이에 약간의 차이만 있어도 새로운 카테고리(Category)로 분류하고, 반대로 경계 변수의 값이 적으면 입력 패턴과 저장 패턴 사이에 많은 차이가 있더라도 유사성이 인정되어 입력 패턴들을 대략적으로 분류한다. 따라서 ART1 알고리즘을 영상 인식에 적용하기 위해서는 경계 변수를 경험적으로 설정하므로 인식률에 부정적인 영향을 갖는 문제점이 있다. 따라서 본 논문에서는 개선된 ART1 알고리즘과 지도 학습 방법을 결합하여 신경망의 은닉층 노드를 동적으로 변화시키는 자가 생성지도 학습 알고리즘을 제안한다. 제안된 신경망에서 입력층과 은닉층의 학습 구조에는 ART1 알고리즘을 개선하여 적용하고, 은닉층과 출력층의 학습 구조에는 은닉층에서 승자로 선택된 노드와 출력층 노드와 연결된 가중치만을 조정하고 Delta-Bar-Delta 알고리즘을 적용한다. 제안된 방법의 학습 성능을 분석하기 위하여 학생증 영상에서 추출한 학번 패턴 분류에 적용한 결과, 기존의 신경망 학습 알고리즘보다 학습 성능이 개선됨을 확인하였다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.34 no.1
• /
• pp.61-65
• /
• 2010

A commercially available soft PZT wafer that is poled in thickness direction is subjected to longitudinal tensile stress loading in both short and open-circuit conditions. Variations of electric displacement in thickness direction and in-plane strains are measured over time during the loading. Different material responses in the two electrical boundary conditions are explained by the effects of piezoelectrically produced internal electric field on linear material moduli and domain switching mechanisms. Finally, a free energy model of normal distribution is introduced to explain the observed creep behavior, and its predictions are compared with experimental observations.

• Journal of Sensor Science and Technology
• /
• v.8 no.1
• /
• pp.45-52
• /
• 1999

In this raper, an enhanced learning method is proposed for improving the learning speed of the error back propagation learning algorithm. In order to cope with the premature saturation phenomenon at the initial learning stage, a variation scheme of active functions is introduced by using higher order functions, which does not need much increase of computation load. It naturally changes the learning rate of inter-connection weights to a large value as the derivative of sigmoid function abnormally decrease to a small value during the learning epoch. Also, we suggest the hybrid learning method incorporated the proposed method with the momentum training algorithm. Computer simulation results show that the proposed learning algorithm outperforms the conventional methods such as momentum and delta-bar-delta algorithms.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1997.10a
• /
• pp.454-457
• /
• 1997

The purpose of this study is to develop an optimal model, using existing models, that is able to estimate the amount of spatter utilizing artificial neural network in the short circuit transfer mode of gas metal arc (GMA) welding. The amount of spatter generated during welding can become a barometer which represents the process stability of metal transfer in GMA welding, and it depends on some factors which constitute a periodic waveforms of welding current and arc voltage in short circuit GMA welding. So, the 12 factors, which could express the characteristics for the waveforms, and the amount of spatter are used as input and output variables of the neural network, respectively. Two neural network models to estimate the amount of spatter are proposed: A neural network model, where arc extinction is not considered, and a combined neural network model where it is considered. In order to reduce the calculation time it take to produce an output, the input vector and hidden layers for each model are optimized using the correlation coefficients between each factor and the amount of spattcr. The est~mation performance of each optimized model to the amount of spatter IS assessed and compared to the est~mation performance of the model proposed by Kang. Also, through the evaluation for the estimation performance of each optimized model, it is shown that the combined neural network model can almost perfectly predict the amount of spatter.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.41 no.8
• /
• pp.131-137
• /
• 2004

The aim of this paper is to find the weather conditions which make the abnormal propagation of a radar bear In order to analyze the weather conditions which cause superrefraction or ducting, the meteorological data of the west sea and the south sea of Korea are classified which are observed during three years from 2000. Atmospheric indexes of refraction with increasing altitude are calculated and the rate of variation of temperature and hmidity at the altitude where the index is very low are observed. It is found that unwanted radar echoes by anomalous propagation are showed up only when the atmospheric indexes of refraction at a altitude is less than -150/km and the reverse layer of temperature appears with a sudden drop of humidity at the altitude.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.9 no.4
• /
• pp.435-445
• /
• 1997

It was studied the phenomena of transient freezing of an inclined water-saturated enclosure. One side of the test section was cooled and the other sides were insulated. The effects of the initial temperature, the inclination angle on the temperature field and the shape of the ice-water interface were observed. In the beginning of freezing, with increasing the inclination angle the freezing rate was increased and in the stable density layer centered $4^{\circ}C$, the freezing was fast as the convective fluid flow became small. When the initial temperature was above the $4^{\circ}C$, the frozen thickness in the upper part of inclined surface was thinner than that in the lower part, but with time the frozen thickness of upper part was thicker than that of lower part, below the $4^{\circ}C$, the frozen thickness in the upper part was thicker than that of lower part from the begining, and above the $8^{\circ}C$ in the beginning upper part was thinner with concave, but with time thicker the upper part, vanishing concave.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.7 no.1
• /
• pp.36-39
• /
• 2014

MFC(microbial fuel cell) is the device to produce the electricity by using the microbes which are living in the waste water. In this paper, the electric characteristics of the MFC were investigated according to each different structure and electrode materials. The voltage being reversed phenomenon was observed in the MFC which uses the cupper plate as the cathode material. This result comes from the oxidation reaction of the cupper plate electrode in this MFC. And this MFC has lower output voltage than one that has a platinum plate electrode. The smaller gap distance of the cupper plate electrode of the MFC showed the higher output voltage. The larger electrode area of the cupper plate electrode showed that the reaching time of the output voltage to the maximum value was delayed.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.12 no.4
• /
• pp.619-627
• /
• 1999

다분야 통합 시스템의 설계문제는 다량의 설계변수와 구속조건으로 구성되며 다수의 공학적 현상으로 연관되어 있다. 다분야 통합 최적설계 문제를 효과적으로 다루기 위해서는 다양한 해석분야의 공학적 설계원리를 동시에 고려하여 균형 있고 유기적인 방법으로 최적의 설계를 결정하는 체계적인 설계자동화기술이 요구된다. 다분야 통합 설계문제를 위한 효율적인 설계방법론으로 분리기반 최적화 기법이 적용되는데 이 방법은 한 단위의 대규모 설계문제를 여러 개의 하부시스템으로 분리하여 독립적으로 최적화를 수행하고 각 하부 시스템으로부터의 설계해 사이의 중재 및 통합화를 거쳐 최종적으로 수렴된 최적설계를 찾는 방법이다. 본 논문에서는 분리기반 최적화기법을 다분야 통합최적 설계문제에 적용하는데 필요한 시스템분리기법을 유전알고리즘 및 다층 역전 파 신경회로망을 이용하여 정립하였다. 시스템분리기법을 검증하기 위해 최근 미국 Boeing사에서 개발중인 고속 민간항공기인 HSCT의 시뮬레이션기반 설계문제를 이용하였다. 대규모 설계시스템의 분리결과는 전체 설계문제의 특성을 파악하기 위한 자료로 활용되며 향후, 분리기반 최적화과정에서 최종적으로 통합된 최적설계를 탐색하는데 필요한 기반구조를 제공한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.2C
• /
• pp.31-40
• /
• 2005

we propose an enhanced self-generation supervised algorithm that by combining an ART algorithm and the delta-bar-delta method. Form the input layer to the hidden layer, ART-1 and ART-2 are used to produce nodes, respectively. A winner-take-all method is adopted to the connection weight adaption so that a stored pattern for some pattern is updated. we test the recognition of student identification, a certificate of residence, and an identifier from container that require nodes of hidden layers in neural network. In simulation results, the proposed self-generation supervised learning algorithm reduces the possibility of local minima and improves learning speed and paralysis than conventional neural networks.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.4 no.6
• /
• pp.554-560
• /
• 1999

In this paper, a vibration suppression method using an energy stored in winding inductance and an induced v voltage of the Linear Stepping Motor(LSM) is shown, and it is applied to a new one-phase excitation method A And a magnetic equivalent circuit is based on the structure of the LSM, and then the electric equivalent circuit of the LSM is derived by solving equations for the magnetic equivalent circuit. Several dynamic characteristics of the LSM are analyzed by the ACSL with the voltage equations, the force equations and the kinetic equation, a and are measured by experimental system.