• Journal of the Korea Society of Computer and Information
• /
• v.16 no.8
• /
• pp.39-47
• /
• 2011

Ergezinger's method, one of the layer-by-layer algorithms used for multilyer perceptrons, consists of an output node and can make premature saturations in the output's weight because of using linear least squared method in the output layer. These saturations are obstacles to learning time and covergence. Therefore, this paper expands Ergezinger's method to be able to use an output vector instead of an output node and introduces a learning rate to improve learning time and convergence. The learning rate is a variable rate that reflects the correlation coefficient between new weight and previous weight while updating hidden's weight. To compare the proposed method with Ergezinger's method, we tested iris recognition and nonlinear approximation. It was found that the proposed method showed better results than Ergezinger's method in learning convergence. In the CPU time considering correlation coefficient computation, the proposed method saved about 35% time than the previous method.

• Geophysics and Geophysical Exploration
• /
• v.12 no.1
• /
• pp.99-104
• /
• 2009

We propose a waveform inversion method for SH-wave data obtained in a shallow seismic refraction survey, to determine a 2D inhomogeneous S-wave profile of shallow soils. In this method, a 2.5D equation is used to simulate SH-wave propagation in 2D media. The equation is solved with the staggered grid finite-difference approximation to the 4th-order in space and 2nd-order in time, to compute a synthetic wave. The misfit, defined using differences between calculated and observed waveforms, is minimised with a hybrid heuristic search method. We parameterise a 2D subsurface structural model with blocks with different depth boundaries, and S-wave velocities in each block. Numerical experiments were conducted using synthetic SH-wave data with white noise for a model having a blind layer and irregular interfaces. We could reconstruct a structure including a blind layer with reasonable computation time from surface seismic refraction data.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.12 no.1
• /
• pp.106-115
• /
• 1988

A numerical procedure for determining the airfoil shape that maximizes the lift is presented. The structure of the flow field is calculated by iteratively coupling potential flow and boundary analysis using the viscous-inviscid interaction method. The potential flow field is obtained by the vortex panel method and boundary layer flow is analyzed by means of integral approximation method which is capable of handling the laminar, transition and turbulent flow regimes. As the result of this study, it is found that the calculated flow regimes have good agreement with the existing experimented data. Davidon-Fletcher-Powell method and Augmented Lagrange Multiplier method are used for the optimal techniques. NACA 23012, NACA 65-3-21, NACA 64-2-415, NACA 64-2-A215 airfoils are used for determining the optimal airfoil shapes as a basic and compensate airfoils. Optimal design showed that the lift coefficients are increased by 17.4% at M$_{0}$=0.2 and 29% at M$_{0}$=0.3, compared with those of basic airfoil.oil.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2002.11a
• /
• pp.59-62
• /
• 2002

Wavelength selection and prediction algorithm for determining hematocrit are investigated. A model based on the difference in optical density induced by the pulsation of heart beat is developed by taking approximation of Twersky's theory on the assumption that the variation of blood vessel size is small during arterial pulsing[1]. A device is constructed with a five-wavelength LED array as light source. The selected wavelengths are two isobestic points and three in compensation for tissue scattering. Data are collected from 549 out-patients who are randomly grouped as calibration and prediction sets. The range of percent hematocrit was 19.3∼51.8. The ratio of the variations of optical density between systole and diastole at two different wavelengths is used as a variable. We selected several such variables that show high reproducibility among all variables. Multiple linear regression analysis is made. The relative percent error is 8% and the standard deviation is 3.67 for the calibration set. The relative % error and standard deviation of the prediction set are 8.2% and 3.69 respectively. We successfully demonstrate the possibility of non-invasive hematocrit measurement, particularly, using the wavelengths below 1000nm.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.6
• /
• pp.842-848
• /
• 2008

This paper deal with uncertainty problem by using Type-2 fuzzy logic set for nonlinear system modeling. We design Type-2 fuzzy logic system in which the antecedent and the consequent part of rules are given as Type-2 fuzzy set and also analyze the performance of the ensuing nonlinear model with uncertainty. Here, the apexes of the antecedent membership functions of rules are decided by C-means clustering algorithm and the apexes of the consequent membership functions of rules are learned by using back-propagation based on gradient decent method. Also, the parameters related to the fuzzy model are optimized by means of particle swarm optimization. The proposed model is demonstrated with the aid of two representative numerical examples, such as mathematical synthetic data set and Mackey-Glass time series data set and also we discuss the approximation as well as generalization abilities for the model.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.58 no.2
• /
• pp.399-406
• /
• 2009

In this study, Polynomial Radial Basis Function Neural Network(pRBFNN) based on Fuzzy Inference System is designed and its parameters such as learning rate, momentum coefficient, and distributed weight (width of RBF) are optimized by means of Particle Swarm Optimization. The proposed model can be expressed as three functional module that consists of condition part, conclusion part, and inference part in the viewpoint of fuzzy rule formed in 'If-then'. In the condition part of pRBFNN as a fuzzy rule, input space is partitioned by defining kernel functions (RBFs). Here, the structure of kernel functions, namely, RBF is generated from HCM clustering algorithm. We use Gaussian type and Inverse multiquadratic type as a RBF. Besides these types of RBF, Conic RBF is also proposed and used as a kernel function. Also, in order to reflect the characteristic of dataset when partitioning input space, we consider the width of RBF defined by standard deviation of dataset. In the conclusion part, the connection weights of pRBFNN are represented as a polynomial which is the extended structure of the general RBF neural network with constant as a connection weights. Finally, the output of model is decided by the fuzzy inference of the inference part of pRBFNN. In order to evaluate the proposed model, nonlinear function with 2 inputs, waster water dataset and gas furnace time series dataset are used and the results of pRBFNN are compared with some previous models. Approximation as well as generalization abilities are discussed with these results.

• Proceedings of the KIEE Conference
• /
• 2000.07d
• /
• pp.2843-2845
• /
• 2000

A lane detection based on a road model or feature all need correct acquirement of information on the lane in a image, It is inefficient to implement a lane detection algorithm through the full range of a image when being applied to a real road in real time because of the calculating time. This paper defines two searching range of detecting lane in a road, First is searching mode that is searching the lane without any prior information of a road, Second is recognition mode, which is able to reduce the size and change the position of a searching range by predicting the position of a lane through the acquired information in a previous frame. It is allow to extract accurately and efficiently the edge candidates points of a lane as not conducting an unnecessary searching. By means of removing the perspective effect of the edge candidate points which are acquired by using the inverse perspective transformation, we transform the edge candidate information in the Image Coordinate System(ICS) into the plane-view image in the World Coordinate System(WCS). We define linear approximation filter and remove the fault edge candidate points by using it This paper aims to approximate more correctly the lane of an actual road by applying the least-mean square method with the fault-removed edge information for curve fitting.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.42 no.12
• /
• pp.103-110
• /
• 2005

The performance of maximum likelihood(ML) detection for the given channel is analyzed in spatially multiplexed MIMO system. In order to obtain the vector symbol error rate, we define error vectors which represent the geometrical relation between lattice points. The properties of error vectors are analyzed to show that all lattice points in infinite lattice almost surely have four nearest neighbors after random channel transformation. Using this information and minimum distance obtained by the modified sphere decoding algorithm, we formulate the analytical performance of vector symbol error over the given channel. To verify the result, we simulate ML performance over various random channel which are classified into three categories: unitary channel, dense channel, and sparse channel. From the simulation results, it is verified that the derived analytical result gives a good approximation about the performance of ML detector over the all random MIMO channels.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.38 no.6
• /
• pp.26-37
• /
• 2001

In order to improve the several problems of the general backpropagation, we propose a method using a fuzzy logic system for automatic tuning of the activation function gain in the backpropagation. First, we researched that the changing of the gain of sigmoid function is equivalent to changing the learning rate, the weights, and the biases. The inputs of the fuzzy logic system were the sensitivity of error respect to the last layer and the mean sensitivity of error respect to the hidden layer, and the output was the gain of the sigmoid function. In order to verify the effectiveness of the proposed method, we performed simulations on the parity problem, function approximation, and pattern recognition. The results show that the proposed method has considerably improved the performance compared to the general backpropagation.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.26 no.5
• /
• pp.393-399
• /
• 2013

In this paper, a novel approach to estimate cohesive laws for the mode I fracture of the graphene is presented by combining molecular dynamic simulations and an inverse algorithm based on field projection method and finite element method. The determination of crack-tip cohesive laws of the graphene based on continuum mechanics is a non-trivial inverse problem of finding unknown tractions and separations from atomic simulations. The displacements of molecular dynamic simulations in a region far away from the crack tip are transferred to finite element nodes by using moving least square approximation. Inverse analyses for extracting unknown cohesive tractions and separation behind the crack tip can be carried out by using conservation nature of the interaction J- and M-integrals with numerical auxiliary fields which are generated by systematically imposing uniform surface tractions element-by-element along the crack surfaces in finite element models. The preset method can be a very successful approach to extract crack-tip cohesive laws from molecular dynamic simulations as a scale bridging method.