• 대한기계학회논문집
• /
• 제18권8호
• /
• pp.1939-1951
• /
• 1994

This study introduces a new tool breakage detecting technology comprised of an unsupervised neural network combined with adaptive time series autoregressive(AR) model where parameters are estimated recursively at each sampling instant using a parameter adaptation algorithm based on an RLS(Recursive Least Square). Experiment indicates that AR parameters are good features for tool breakage, therefore it can be detected by tracking the evolution of the AR parameters during milling process. an ART 2(Adaptive Resonance Theory 2) neural network is used for clustering of tool states using these parameters and the network is capable of self organizing without supervised learning. This system operates successfully under the wide range of cutting conditions without a priori knowledge of the process, with fast monitoring time.

• International Journal of Control, Automation, and Systems
• /
• 제6권5호
• /
• pp.767-775
• /
• 2008

Methods used for exploring marine resources and spaces include positioning a probe under water and then recalling it after a specified time. Hydro-acoustic Releasers are commonly used for positioning and retrieving of such exploration equipment. The most important factor in this kind of system is the reliability for recalling the instruments. The frequently used ultrasonic signal detection method can detect ultrasonic signals using a fixed comparator, but because of increased rates of errors due to outside interferences, information is repetitively acquired. This study presents an effective ultrasonic signal detection algorithm using the characteristics of a resonance and adaptive comparator Combined with the FSK+ASK modulator. As a result, approximately 8.8% of ultrasonic wave communication errors caused by background noise and transmission losses were reduced for effectively detecting ultrasonic waves. Furthermore, the resonance circuit's quality factor was enhanced (Q = 120 to 160). As such, the bias voltage of the transistor (Vb= 3.3 to 6.8V) was increased thereby enhancing the frequency's selectivity.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2005년도 춘계학술대회
• /
• pp.186-189
• /
• 2005

본 논문에서는 LVQ를 퍼지화한 새로운 퍼지 학습 법칙들을 제안하였다. 퍼지 LVQ 학습법칙 1은 기존의 학습률 대신에 퍼지 학습률을 사용하였는데 이는 조건 확률의 퍼지화에 기반을 두고 있다. 퍼지 LVQ 학습법칙 2는 클래스들 사이에 존재하는 입력벡터가 결정 경계선에 대한 정보를 더 가지고 있는 것을 반영한 것이다. 이 새로운 퍼지 학습 법칙들을 improved IAFC(Integrted Adaptive Fuzzy Clustering)신경회로망에 적용하였다. improved IAFC신경회로망은 ART-1 (Adaptive Resonance Theory)신경회로망과 Kohonen의 Self-Organizing Feature Map의 장점을 취합한 퍼지 신경회로망이다. 제안한 supervised IAFC 신경회로망 1과 supervised IAFC neural 신경회로망 2의 성능을 오류 역전파 신경회로망의 성능과 비교하기 위하여 iris 데이터를 사용하였는데 Supervised IAFC neural network 2가 오류 역전파 신경회로망보다 성능이 우수함을 보여주었다.

• 한국지능시스템학회논문지
• /
• 제15권6호
• /
• pp.673-681
• /
• 2005

원자력 발전소는 안정성 및 신뢰성 확보가 가장 중요하므로 고장의 감지 및 진단 시스템의 개발은 원전 자체가 구축하고 있: 다중의 하드웨어 중첩도(hardware redundancy)에도 불구하고 가장 중요한 문제로 취급되고 있다. 본 논문에서는 원저 PWR 증기발생기에서 발생한 고장을 진단하기 위한 알고리듬의 개발을 위해 시스템에서 발생한 고장을 감지하고 분류할 수 있는 ART2 시경회로망 기반 고장진단방법을 제안한다. 고장진단시스템은 발생한 고장을 감지하기 위한 고장감지부, 변화된 시스템파라미터를 추정하기 위한 파라미터 추정부 및 발생한 고장의 종류를 알아내기 위한 고장분류부로 구성된다. 고장분류부는 여러 경계인수를 갖는 ART2(adaptive resonance theory 2) 신경회로망을 이용한 고장분류기로 구성된다. 제안한 고장진단 알고리듬을 증기발생기의 고장진단문제에 적용하여 성능을 확인하였다.

• 한국정보처리학회논문지
• /
• 제1권2호
• /
• pp.215-224
• /
• 1994

본 논문은 신경회로망 EART(Extended Adaptive Resonance Theory)를 이용한 회전 과 크기 변화에 무관한 지문인식에 관한 연구이다. 지문 농담 화상($515{\times}512$)을 적응 문턱 값을 이용하여 융선와 골을 분리하여 이진화 영상으로 바꾼후 이를 다시 세선화 영상으로 만든다. 이진 세선화 영상으로부터 지문의 특징점 중 식별에 가장 큰영향을 주는 분기점과 끝점을 $3{\times}3$마스크를 사용해서 추출한다. 이렇게 추출된 분기점과 끝 점의 개수, 그리고 분기점으로 이루어진 볼록 다각형의 내각을 회전변화와 크기변화 에 영향을 받지않는 가중코드(weighted code)로 된 40*10 특징점 행렬로 나타낸 후 이를 신경회로망 EART의 입력으로 했다. 신경망을 이용한 본 시스템은 세선화 영상에 대한 어떠한 복원 처리 과정도 없이 영상의 회전과 크기 변화에 대해서도 매우 효과 적이고도 만족할 만한 결과를 얻을 수 있었다.

• 한국의학물리학회지:의학물리
• /
• 제30권4호
• /
• pp.104-111
• /
• 2019

Purpose: Online magnetic resonance-guided adaptive radiotherapy (MRgART), an emerging technique, is used to address the change in anatomical structures, such as treatment target region, during the treatment period. However, the electron density map used for dose calculation differs from that for daily treatment, owing to the variation in organ location and, notably, air pockets. In this study, we evaluate the dosimetric effect of electron density override on air pockets during online ART for pancreatic cancer cases. Methods: Five pancreatic cancer patients, who were treated with MRgART at the Seoul National University Hospital, were enrolled in the study. Intensity modulated radiation therapy plans were generated for each patient with 60Co beams on a ViewrayTM system, with a 45 Gy prescription dose for stereotactic body radiation therapy. During the treatment, the electron density map was modified based on the daily MR image. We recalculated the dose distribution on the plan, and the dosimetric parameters were obtained from the dose volume histograms of the planning target volume (PTV) and organs at risk. Results: The average dose difference in the PTV was 0.86Gy, and the observed difference at the maximum dose was up to 2.07 Gy. The variation in air pockets during treatment resulted in an under- or overdose in the PTV. Conclusions: We recommend the re-contouring of the air pockets to deliver an accurate radiation dose to the target in MRgART, even though it is a time-consuming method.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 V
• /
• pp.2775-2778
• /
• 2003

Neural network-based fault diagnosis algorithm to detect and isolate faults in the nonlinear systems is proposed. In the proposed method, the fault is detected when the errors between the system output and the neural network nominal system output cross a predetermined threshold. Once a fault in the system is detected, the system outputs are transferred to the fault classifier by ART2 NN (adaptive resonance theory 2 neural network) for fault isolation. From the computer simulation results, it is verified that the proposed fault diagonal method can be performed successfully to detect and isolate faults in a nonlinear system.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.127.1-127
• /
• 2001

In this paper we propose an input/output model based fault diagnosis method to detect and isolate single faults in the robot arm control system. The proposed algorithm is functionally composed of three main parts-parameter estimation, fault detection, and isolation, When a change in the system occurs, the errors between the system output and the estimated output cross a predetermined threshold, and once a fault in the system is detected, and in this zone the estimated parameters are transferred to the fault classifier by ART2(adaptive resonance theory 2) neural network for fault isolation. Since ART2 neural network is an unsupervised neural network fault classifier does not require the knowledge of all possible faults to isolate the faults occurred in the system. Simulations are carried out to evaluate the performance of the proposed ...

• 제어로봇시스템학회논문지
• /
• 제3권3호
• /
• pp.244-251
• /
• 1997

In this paper, we propose a fault diagnosis algorithm to detect and isolate multiple faults in a system. The proposed fault diagnosis algorithm is based on a multiple fault classifier which consists of two ART2 NN(adaptive resonance theory2 neural network) modules and the algorithm is composed of three main parts - parameter estimation, fault detection and isolation. When a change in the system occurs, estimated parameters go through a transition zone in which residuals between the system output and the estimated output cross the threshold, and in this zone, estimated parameters are transferred to the multiple faults classifier for fault isolation. From the computer simulation results, it is verified that when the proposed diagnosis algorithm is performed successfully, it detects and isolates faults in the position control system of a DC motor.

• Smart Structures and Systems
• /
• 제20권6호
• /
• pp.657-668
• /
• 2017

In this research, an active vibration suppression of a smart beam having piezoelectric sensor and actuators is investigated by designing separate controllers comprising a linear quadratic regulator and a neural network. Firstly, design of a smart beam which consists of a cantilever aluminum beam with surface bonded piezoelectric patches and a designed mechanism having a micro servomotor with a mass attached arm for obtaining variations in the frequency response function are presented. Secondly, the frequency response functions of the smart beam are investigated experimentally by using different piezoelectric patch combinations and the analytical models of the smart beam around its first resonance frequency region for various servomotor arm angle configurations are obtained. Then, a linear quadratic regulator controller is designed and used to simulate the suppression of free and forced vibrations which are performed both in time and frequency domain. In parallel to simulations, experiments are conducted to observe the closed loop behavior of the smart beam and the results are compared as well. Finally, active vibration suppression of the smart beam is investigated by using a linear controller with a neural network based adaptive element which is designed for the purpose of overcoming the undesired consequences due to variations in the real system.