• 전자공학회논문지B
• /
• 제29B권11호
• /
• pp.120-129
• /
• 1992

In this paper we propose a fuzzy-neural network(FNN) which includes both advantages of the fuzzy logic and the neural network. The basic idea of the FNN is to realize the fuzzy rule-base and the process of reasoning by neural network and to make the corresponding parameters be expressed by the connection weights of neural network. After constructing the FNN, a novel controller consisting of a conventional P-controller and a FNN is explained. In this control scheme, the rule-base of a FNN are automatically generated by error back-propagation algorithm. Also the parallel connection of the P-controller and the FNN can guarantee the stability of a plant at initial stage before the rules are completely created. Finally the effectiveness of the proposed strategy will be verified by computer simulations using a 2 degree of freedom robot manipulator.

• 대한전기학회논문지:전기물성ㆍ응용부문C
• /
• 제53권10호
• /
• pp.515-522
• /
• 2004

In this study, we tried to investigate that the advantages of SOM(Self Organizing Map) algorithm such as data accumulation ability and the degradation trend trace ability would be adaptable to the analysis of partial discharge pattern recognition. For the purpose, we analyzed partial discharge data obtained from the typical artificial defects in GIS and XLPE power cable system through SOM algorithm. As a result, partial discharge pattern recognition could be well carried out with an acceptable error by use of Kohonen map in SOM algorithm. Also, it was clarified that the additional data could be accumulated during the operation of the algorithm. Especially, we found out that the data accumulation ability of Kohonen map could make it possible to suggest new patterns, which is impossible through the conventional BP(Back Propagation) algorithm. In addition, it is confirmed that the degradation trend could be easily traced in accordance with the degradation process. Therefore, it is expected to improve on-site applicability and to trace real-time degradation trends using SOM algorithm in the partial discharge pattern recognition

• 전자공학회논문지S
• /
• 제35S권2호
• /
• pp.70-78
• /
• 1998

Spike detection in long-term EEG monitoring forepilepsy by wavelet transform(WT), artificial neural network(ANN) and the expert system is presented. First, a small set of wavelet coefficients is used to represent the characteristics of a singlechannel epileptic spikes and normal activities. In this stage, two parameters are also extracted from the relation between EEG activities before the spike event and EEG activities with the spike. then, three-layer feed-forward network employing the error back propagation algorithm is trained and tested using parameters obtained from the first stage. Spikes are identified in individual EEG channels by 16 identical neural networks. Finally, 16-channel expert system based on the context information of adjacent channels is introducedto yield more reliable results and reject artifacts. In this study, epileptic spikes and normal activities are selected from 32 patient's EEG in consensus among experts. The result showed that the WT reduced data input size and the preprocessed ANN had more accuracy than that of ANN with the same input size of raw data. Ina clinical test, our expert rule system was capable of rejecting artifacts commonly found in EEG recodings.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2003년도 추계 학술대회 학술발표 논문집
• /
• pp.95-98
• /
• 2003

In this paper, an adaptive fuzzy neural control of unknown nonlinear systems based on the rapid learning algorithm is proposed for optimal parameterization. We combine the advantages of fuzzy control and neural network techniques to develop an adaptive fuzzy control system for updating nonlinear parameters of controller. The Fuzzy Neural Network(FNN), which is constructed by an equivalent four-layer connectionist network, is able to learn to control a process by updating the membership functions. The free parameters of the AFN controller are adjusted on-line according to the control law and adaptive law for the purpose of controlling the plant track a given trajectory and it's initial values are off-line preprocessing, In order to improve the convergence of the learning process, we propose a rapid learning algorithm which combines the error back-propagation algorithm with Aitken's $\delta$$\^$2/ algorithm. The heart of this approach ls to reduce the computational burden during the FNN learning process and to improve convergence speed. The simulation results for nonlinear plant demonstrate the control effectiveness of the proposed system for optimal parameterization.

• 한국전자통신학회논문지
• /
• 제10권4호
• /
• pp.521-526
• /
• 2015

개인의 신원을 확인하기 위해 인간의 생물학적 특성을 사용하는 방법에 대한 연구가 활발히 진행되고 있다. 심전도를 이용한 생체 인식 기술은 피험자에 피부자극을 일으키지 않고 위조가 어렵다. 기존의 생체 인식 시스템인 지문, 얼굴 등의 인식시스템과 쉽게 접목이 가능하여 다중 생체 인식 시스템으로 응용할 수 있다. 본 논문에서는 이산 웨이블릿 변환 계수를 사용한 심전도의 파형 특성분석법으로 개인을 식별하는 방법을 제안하였다. 심전도 신호의 특징추출은 총 9개의 이산 웨이블릿 변환 계수를 대상으로 상관 계수 분석으로 수행하였다. 식별은 각 클래스의 특징벡터를 입력으로 오류 역전파 신경망을 적용하여 수행하였다. MIT-BIH QT 데이터베이스내 24명의 심전도에 대해 98.88%의 식별율을 나타냈다.

• 센서학회지
• /
• 제8권1호
• /
• pp.45-52
• /
• 1999

이 논문에서는 오차 역전파 학습 알고리듬의 학습 속도를 향상시키기 위한 새로운 학습 방법을 제안한다. 제안하고자 하는 방법은 시그모이드 형태를 갖는 신경회로망의 활성화 함수(activation function) 자체에 고차항(higher order)을 적절히 이용하여 초기 학습 단계에서 발생할 수 있는 조기 포화(premature saturation) 현상을 계산량의 큰 증가 없이 효과적으로 대처할 수 있다. 고차항을 이용함으로써 은닉층 활성화 함수의 도합수가 작은 값으로 감소함에 따라 신경망의 연결 강도를 학습시키는 학습율은 적응적으로 큰 값을 갖게 된다. 또한, 은닉층에 고차항을 이용하는 제안한 방법에 모멘텀(momentum) 학습 알고리듬을 결합하는 새로운 hybrid 학습 방법을 제안한다. 컴퓨터 모의 실험을 통해 제안하고자 하는 학습 방법과 기존의 방법들과의 학습 속도 성능을 비교한다.

• 정보처리학회논문지B
• /
• 제11B권4호
• /
• pp.449-456
• /
• 2004

본 논문에서는 모호성을 포함하고 있는 시계열 패턴인식을 위한 새로운 인식모델인 순환퍼지기억장치를 제안한다. 순환퍼지기억장치는 기존의 퍼지기억장치에 순차적인 입력패턴를 처리하고 시간적 관련성을 표현할 수 있는 순환층을 추가함으로써 확장된 모델이다. 본 논문에서 제안하는 순환퍼지기억장치는 입력과 출력사이의 관련정도를 설정하기 위해 헤비안 방식의 학습알고리즘을 사용한다. 그리고 순환퍼지기억장치의 순환층에 필요한 가중치를 학습하기 위해서 오류역전파 알고리즘을 이용한다. 본 논문에서는 제안하는 모델을 음성신호의 경계를 추출하는 문제에 적용하여 성능을 평가한다.

• 한국안전학회지
• /
• 제17권1호
• /
• pp.11-17
• /
• 2002

The purpose of this study was to present the method to choose the optimization machining condition for the wire electric machine. This was completed by examining the ever-changing quality of the material and by improving the function of the wire electric discharge machine. Precision metal mold products and the unmanned wire electric discharge machining system were used and then applied in industrial fields. This experiment uses the wire electric discharge machine with brass wire electrode of 0.25mm. To measure the precision of the machining surface, average values are obtained from 3 samples of measures of center-line average roughness by using a third dimension gauge and a stylus surface roughness gauge. In this experiment, we changed no-node voltage to 7 and 9, pulse-on-time to $6{\mu}s,\;8{\mu}s$ and $10{\mu}s$, pulse-off-time to $8{\mu}s,\;10{\mu}s$ and $13{\mu}s$, and experimented on wire tension at room temperature by 1000gf, 1200gf, and 1400gf, respectively.

• 한국정밀공학회지
• /
• 제18권4호
• /
• pp.209-218
• /
• 2001

Ultrasonic motors are newly developed motors which are expected to be useful as actuators in many practical systems such as robot arms or manipulators because of several advantages against the electromagnetic motors. However, the precise control of the ultrasonic motor is generally difficult due to the absence of appropriate and rigorous mathematical model. Furthermore, owing to heavy nonlinearity, the position control of a pendulum system driven by the ultrasonic motor has a problem that control method using multiple neural network controllers based on a fuzzy inference system that can determine the initial position of the pendulum in the beginning of control operation. In addition, and appropriate neural network controller that has been learned to operate well at the corresponding initial position is adopted by switching schemes. The effectiveness of the proposed method was verified and evaluated from real experiments.

• 한국의류학회지
• /
• 제26권1호
• /
• pp.152-159
• /
• 2002

Neural networks are used to predict the sense of touch of polyurethane coated fabrics. In this study, we used the multi layer perceptron (MLP) neural networks in Neural Connection. The learning algorithm for neural networks is back-propagation algorithm. We used 29 polyurethane coated fabrics to train the neural networks and 4 samples to test the neural networks. Input variables are 17 mechanical properties measured with KES-FB system, and output variable is the sense of touch of polyurethane coated fabrics. The influence of MLF function, the number of hidden layers, and the number of hidden nodes on the prediction accuracy is investigated. The results were as follows: MLP function, the number of hidden layer and the number of hidden nodes have some influence on the prediction accuracy. In this work, tangent function, the architecture of the double hidden layers and the 24-12-hidden nodes has the best prediction accuracy with the lowest RMS error. Using the neural networks to predict the sense of touch of polyurethane coated fabrics has hotter prediction accuracy than regression approach used in our previous study.