• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.12
• /
• pp.2418-2425
• /
• 1994

Although the error backpropagation(EBP) algorithm based on the gradient descent method is a widely-used learning algorithm of neural networks, learning sometimes takes a long time to acquire accuracy. This paper develops a novel learning method to alleviate the problems of EBP algorithm such as local minima, slow speed, and size of structure and thus to improve performance by adopting other new networks. Gaussian Potential Function networks(GPFN), in parallel with multilayer neural networks. Empirical simulations show the efficacy of the proposed algorithm in function approximation, which enables us to train networks faster with the better generalization capabilities.

• Journal of the Institute of Convergence Signal Processing
• /
• v.4 no.3
• /
• pp.41-48
• /
• 2003

This paper presents an indirect adaptive neuro controller using modified chaotic neural networks(MCNN) for nonlinear dynamic system. A modified chaotic neural networks model is presented for simplifying the traditional chaotic neural networks and enforcing dynamic characteristics. A new Dynamic Backpropagation learning method is also developed. The proposed MCNN paradigm is applied to the system identification of a MIMO system and the indirect adaptive neuro controller. The simulation results show good performances, since the MCNN has robust adaptability to nonlinear dynamic system.

• 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.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.7 no.5
• /
• pp.830-836
• /
• 2006

In this paper described an auto tuning neuro-fuzzy PID controller based neural network. The PID type controller has been widely used in industrial application due to its simply control structure, easy of design, and inexpensive cost. However, control performance of the PID type controller suffers greatly from high uncertainty and nonlinearity of the system, large disturbances and so on. In this paper will design to take advantage of neural network fuzzy theory and pid controller auto toning technique. The value of initial scaling factors of the proposed controller were determined on the basis of the conventional PID controller parameters tuning methods and then they were adjusted by using neural network control techniques. This controller simple structure and computational complexity are less, and also application is easy and performance is excellent in system that is strong and has nonlinearity to system dynamic behaviour change or disturbance. Finally, the proposed auto tuning neuro-fuzzy controller is applied to magnetic levitation. Simulation results demonstrated that the control performance of the proposed controller is better than that of the conventional controller.

• Proceedings of the Korean Society for Emotion and Sensibility Conference
• /
• 1999.11a
• /
• pp.388-395
• /
• 1999

본 논문의 목적은 퍼지 엔트로피를 이용하여 비선형신호를 예측하는 것이다. 이 방법은 분할된 여러 부 공간(subspace)에 대해 입력 데이터로부터 퍼지 엔트로피를 이용하여 각각의 규칙에 등급을 정하여 불필요한 제어규칙을 제거하여 바람직한 규칙베이스를 구성하도록 한 것이다. 적용되는 퍼지 신경망의 기본적인 구조는 퍼지 제어기의 규칙베이스와 추론의 과정을 신경회로망을 이용하여 구현하며 퍼지 제어규칙의 매개변수들은 역전파 알고리즘에 의해 적응되어진다. 또한 매개변수의 수를 줄이기 위하여 제어규칙의 결론부의 출력값은 신경망의 가중치로 구성하였다. 결국 퍼지 신경망의 복잡도를 줄일 수 있다. Mackey-Glass 시계열의 예측에 대한 컴퓨터 시뮬레이션을 통하여 본 논문에서 제안한 방법의 효율성을 입증하고, 제안된 방법을 EEG 생리신호 분석에 이용될 수 있다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.29 no.2 s.233
• /
• pp.303-310
• /
• 2005

The systems with probe and SIL(Solid Immersion Lens) mechanisms have been researched as the technology to perform NFR(Near Field Recording). Most of them use the flying head mechanism to accomplish high recording density and fast data transfer rate. In this paper, ABS shape of flying head was optimized with the object of securing the maximum compliance ability of OFH. We suggest low different optimization processes to predict the static flying characteristics for the OFH. Two different approximation methods, regression analysis and back propagation neural network were used. And we compared the result of directly connected(between CAE and optimizer) method and two approximated optimization results. Design Optimization Tool(DOT) and ${\mu}GA$ were used as the optimizers.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.11
• /
• pp.2264-2275
• /
• 1994

In this paper, image compression based on neural networks is presented with block classification and coding. Multilayer neural networks with error back-propagation learning algorithm are used to transform the normalized image date into the compressed hidden values by reducing spatial redundancies. Image compression can basically be achieved with smaller number of hidden neurons than the numbers of input and output neurons. Additionally, the image blocks can be grouped for adaptive compression rates depending on the characteristics of the complexity of the blocks in accordance with the sensitivity of the human visual system(HVS). The quantized output of the hidden neuron can also be entropy coded for an efficient transmission. In computer simulation, this approach lie in the good performances even with images outside the training set and about 25:1 compression rate was achieved using the entropy coding without much degradation of the reconstructed images.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.54 no.3
• /
• pp.144-151
• /
• 2005

This paper presents a forecasting technique of the short-term marginal price (SMP) using an Artificial Neural Network (ANN). The SW forecasting is a very important element in an electricity market for the optimal biddings of market participants as well as for market stabilization of regulatory bodies. Input data are organized in two different approaches, time-axis and day-axis approaches, and the resulting patterns are used to train the ANN. Performances of the two approaches are compared and the better estimate is selected by a composition rule to forecast the SMP. By combining the two approaches, the proposed composition technique reflects the characteristics of hourly, daily and seasonal variations, as well as the condition of sudden changes in the spot market, and thus improves the accuracy of forecasting. The proposed method is applied to the historical real-world data from the Korea Power Exchange (KPX) to verify the effectiveness of the technique.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.13 no.2
• /
• pp.63-70
• /
• 1999

This paper is proposed a new method to deal with the optimized auto-tuning for the Pill controller which is used to the process-control in various fields. First of all, in this method, 1st order delay system with dead time which is modelled from the unit step response of the system is Pade-approximated, then initial values are determined by the Ziegler-Nichols method. So we can find the parameters of Pill controller so as to minimize the fuzzy criterion function which includes the maximum overshoot, damping ratio, rising time and settling time. Finally, after studying the parameters of Pill controller by Backpropagation of Neural-Network, when we give new K, L, T values to Neural-Network, the optimized parameter of Pill controller is found by Neural-Network Program.rogram.

• Journal of Welding and Joining
• /
• v.17 no.5
• /
• pp.98-106
• /
• 1999

In GMAW, the spatters are generated according to the variation of the arc. Of the arc is stable, Few spatters are generated. But if unstable, too many spatters are generated. So, this means the spatters are dependent on the arc state. The aim of this study is to accurately estimate the arc state. To do this, the generated spatters were captured under the some welding conditions, and the waveforms of the arc voltage and welding current were collected. From the collected signals, the waveform factors and their standard deviations were extracted. Using these factors as input parameters of multi-layer artificial neural network, the learning for the weight of the generated spatters is performed and the estimation results to the real spatter are assessed. Obtained results are as follow: the linear correlation coefficient between the estimated result and the real spatters was 0.9986. And although the average convergence error was set 0.002, the estimated error to the real spatter was within 0.1 gr/min at each welding condition. In the estimation for the weight generated spatters, the result with multi-layer neural network was far better than with multiple regression analysis. Especially, even though under the welding condition which the arc state is unstable (the spatter is generated much more), very excellent estimation performance was shown.