• Proceedings of the Korea Information Processing Society Conference
• /
• 2004.11a
• /
• pp.505-508
• /
• 2004

본 논문은 동적 3-D 뉴로 시스템(A Dynamic 3-D Neuro System)모델을 이용하여 오프라인 필기체 숫자 인식 실험을 하였다. 3-D 뉴로 시스템 모델을 사용함으로써 기존에 교육된 정보를 유지하면서 새로운 정보를 추가할 수 있는 추가학습이 가능했고, 동일한 범주의 정보에 대해서는 반복교육 횟수에 따라 교육정도가 점점 누적되는 반복교육이 가능했다. 교육과정을 통해 누적된 정보로부터 일반화된 패턴(Generalized Pattern)을 도출해 인식시 사용할 수 있었다. 패턴 인식기는 피드백루틴을 통해 미지의 입력이미지를 원형이미지로 복원한 후, 그 결과 데이터를 사용하여 문자를 인식하도록 동작한다. NIST의 MNIST 데이터베이스를 사용해 실험을 하였고, 결과로 $99.0\%$의 정인식률을 얻었다.

• Proceedings of the KIEE Conference
• /
• 2001.07a
• /
• pp.318-322
• /
• 2001

본 논문은 뉴로-퍼지 모델의 구조 학습을 이용한 단기 전력 수요 예측시스템에서 예측치별로 신뢰도를 계산하는 체계적인 방법을 제안한다. 예측시스템의 신뢰도를 추정하는 작업은 특히 신경회로망과 같은 경험적 모델을 실제 활용하기 위해서 필수적인 연구로 인식되고 있다. 본 논문에서 제안하는 출력별 신뢰 구간 계산 방법은 지역 표현하는 뉴로-퍼지 모델의 특성을 활용하여 학습된 퍼지 규칙 각각에 대해 신뢰도를 추정하는 Local reliability measure 기법을 사용한다. 제안된 신뢰도 계산이 가능한 단기 전력 수요 예측시스템은 먼저 결정 트리를 이용하여 초기 구조를 생성하고, 이를 초기 구조 뱅크에 저장한다. 저장된 초기 구조 뱅크를 이용하여 뉴로-퍼지 모델을 학습하고, 학습된 퍼지 규칙의 신뢰도를 추정한다. 제안된 시스템의 실효성을 검증하기 위해서 한국 전력에서 수집한 1996년과 1997년의 실제 전력 수요 데이터를 이용하여 한 시간 앞의 수요를 예측하는 모의 실험을 수행하고 실험 결과를 비교 분석한다.

• Journal of the Korean Society for Precision Engineering
• /
• v.15 no.3
• /
• pp.107-118
• /
• 1998

Two-phase neuro-system identification method is presented. The 1$^{st}$-phase identification uses conventional neural network mapping for modeling an input-output system. The 2$^{nd}$ -phase modeling is also performed sequentially using the 1$^{st}$-phase modeling errors. In the 2$^{nd}$ a phase modeling, newly generated input signals, which are obtained by summing the 1st-phase modeling error and artificially generated uniform series, are utilized as system's I-O mapping elements. The 1$^{st}$-phase identification is interpreted as a “Real Model” system identification because it uses system's real data(i.e., observations and control inputs) while the 2$^{nd}$ -phase identification as a “Artificial Model” identification because of using artificial data. Experimental results are given to verify that the two-phase neuro-system identification could reduce the overall modeling errors.rrors.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.9 no.5
• /
• pp.472-479
• /
• 1999

In this paper, we propose a land cover pattern classifier for remote sensing image by using neuro-fuzzy algorithm. The proposed pattem classifier has a 3-layer feed-forward architecture that is derived from generic fuzzy perceptrons, and the weights are con~posed of h u y sets. We also implement a neuro-fuzzy pattern classification system in the Visual C++ environment. To measure the performance of this, we compare it with the conventional neural networks with back-propagation learning and the Maximum-likelihood algorithms. We classified the remote sensing image into the eight classes covered the majority of land cover feature, selected the same training sites. Experimental results show that the proposed classifier performs well especially in the mixed composition area having many classes rather than the conventional systems.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.03a
• /
• pp.95-98
• /
• 1998

초음파센서는 저렴성, 단순한 구조, 기계적 강인성, 사용상의 적은 제약 등의 이점 때문에 다양한 응용분야에 적용된다. 물체의 인식에 초음파센서를 사용하기에는 낮은 분해능을 초래하는 불량한 방향성과 측정오류를 유발하는 반사성의 어려움을 내재하고 있다. 이런 문제를 개선하기 위해서 다양한 센사의 배열형태에서 많은 수의 센서를 사용하거나, 일정 수의 센서를 사용할 경우에는 센서의 배열을 기계적으로 이동시킨다. 본 논문에서는 물체의 패턴인식에 있어서 가장 기본적인 거리, 물체크기, 물체각도 값을 얻기 위해 간단하게 구성된 전자회로를 부가하여 초음파센서의 송출전압을 여러 단계로 변경시켜 얻어낸 데이터에 뉴로퍼지 기반의 지능적 계산 알고리즘을 적용하여 개선된 결과를 얻는다.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.16 no.10
• /
• pp.1833-1840
• /
• 1992

In the diagnostic process for automobile, Neuro-Fuzzy technique was compared with the conventional diagnostic method for the verification of performance, and proto-type system was developed. For the utilities of the system, 1/3 octave filter(band-pass filter) and A/D converter were used for data acquisition and then data were analyzed using octave band processing and pattern recognition using hamming network algorithm. In order to raise the reliability of the diagnostic results by considering many operating variables and condition of automobile to be diagnosed, fuzzy inference technique was applied in combining several information. The validation of this diagnostic system was examined through computer simulation and experiment, and it showed an acceptable performance for diagnostic process.

• Journal of KIISE:Software and Applications
• /
• v.34 no.7
• /
• pp.585-594
• /
• 2007

The back propagation algorithm took a long time to learn the input patterns and was difficult to train the additional or repeated learning patterns. So Aleksander proposed the binary neural network which could overcome the disadvantages of BP Network. But it had the limitation of repeated learning and was impossible to extract a generalized pattern. In this paper, we proposed a dynamic 3 dimensional Neuro System which was consisted of a learning network which was based on weightless neural network and a feedback module which could accumulate the characteristic. The proposed system was enable to train additional and repeated patterns. Also it could be produced a generalized pattern by putting a proper threshold into each learning-net's discriminator which was resulted from learning procedures. And then we reused the generalized pattern to elevate the recognition rate. In the last processing step to decide right category, we used maximum response detector. We experimented using the MNIST database of NIST and got 99.3% of right recognition rate for training data.

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 1992.06a
• /
• pp.538-591
• /
• 1992

• Journal of Digital Convergence
• /
• v.17 no.1
• /
• pp.239-247
• /
• 2019

An accurate prediction of emotion is a very important issue for the sake of patient-centered medical device development and emotion-related psychology fields. Although there have been many studies on emotion prediction, no studies have applied the heart rate variability and neuro-fuzzy approach to emotion prediction. We propose ANFEP(Adaptive Neuro Fuzzy System for Emotion Prediction) HRV. The ANFEP bases its core functions on an ANFIS(Adaptive Neuro-Fuzzy Inference System) which integrates neural networks with fuzzy systems as a vehicle for training predictive models. To prove the proposed model, 50 participants were invited to join the experiment and Heart rate variability was obtained and used to input the ANFEP model. The ANFEP model with STDRR and RMSSD as inputs and two membership functions per input variable showed the best results. The result out of applying the ANFEP to the HRV metrics proved to be significantly robust when compared with benchmarking methods like linear regression, support vector regression, neural network, and random forest. The results show that reliable prediction of emotion is possible with less input and it is necessary to develop a more accurate and reliable emotion recognition system.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.37 no.2
• /
• pp.41-53
• /
• 2000

In this paper, the Maekjin and the Yangdorak Diagnosis algorithm by using a neuro-fuzzy method is proposed and it is implemented on the DSP-based system. Maekjin is measured by 3-channels of the Maekjin board through Maekjin probe which is attached on Chon, Kwan and Chuk of patient's wrist. First, we experiment Chon, Kwan and Chuk, 3-parts simultaneously and second perform one part of Chon, Kwan and Chuk respectively, The experimental results show that the Maekjin signal is measured precisely with any Maekjin probe. In Yangdorak diagnosis, the pulse generated by electric stimulator stimulates a portion of body and the response signal is measured through electrodes which is attached on representative points of 12 kyungmaks. The experimental methods are (1) 1 channel-measure, (2) 2 channels-measure, (3) 6 channels-measure and (4) 24 channels-measure. A fuzzy diagnosis is performed and neural networks is learned using fuzzy values as inputs, and we show that neuro-fuzzy diagnosis method is performed well.