• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.86-89
• /
• 2003

To model a numerical problem space under the limitation of available data, we need to extract sparse but key points from the space and to efficiently approximate the space with them. This study proposes a sampling method based on the search process of genetic algorithm and a space modeling method based on least-squares approximation using the summation of Gaussian functions. We conducted simulations to evaluate them for several kinds of problem spaces: DeJong's, Schaffer's, and our original one. We then compared the performance between our sampling method and sampling at regular intervals and that between our modeling method and modeling using a polynomial. The results showed that the error between a problem space and its model was the smallest for the combination of our sampling and modeling methods for many problem spaces when the number of samples was considerably small.

• Proceedings of the KIEE Conference
• /
• 1999.11c
• /
• pp.824-826
• /
• 1999

In this paper, we design an Fuzzy-Neural Networks(FNN) by means of divisions of fuzzy input space with multi-input variables. Fuzzy input space of Yamakawa's FNN is divided by each separated input variable, but that of the proposed FNN is divided by mutually combined input variables. The membership functions of the proposed FNN use both triangular and gaussian membership types. The parameters such as apexes of membership functions, learning rates, momentum coefficients, weighting value, and slope are adjusted using genetic algorithms. Also, an aggregate objective function(performance index) with weighting value is utilized to achieve a sound balance between approximation and generalization abilities of the model. To evaluate the performance of the proposed model, we use the data of sewage treatment process.

• Publications of The Korean Astronomical Society
• /
• v.7 no.1
• /
• pp.51-61
• /
• 1992

The nonlinear stochastic behavior of chaotic inflation is characterized by the 'scaling' effect. Using a simple criterion for the appearance of scaling behavior in the ${\lambda}{\phi}^4$ inflation model, we show explicitly that in this limit the onset of the scaling regime does not require any special initial conditions and that it is independent of the self-coupling constant ${\lambda}$. Non-Gaussian statistics in adiabatic fluctuations are important only for super-horizon scales and the scaling regime does not lead to any significant statistical properties on currently observable scales. However, the scaling effect gives some cosmological consequences very different from what we expect in the naive diffusion approximation for quantum fluctuations. The classical (deterministic) treatment of the inflation field (essentially a quantum mechanical object.) becomes valid towards the end of inflation.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.26 no.3
• /
• pp.198-205
• /
• 2006

The necessity of nondestructively inspecting austenitic steels, fiber-reinforced composites, and other inherently anisotropic materials has stimulated considerable interest in developing beam models for anisotropic media. The properties of slowness surface playa key role in the beam models based on the paraxial approximation. In this paper, we apply a modular multi-Gaussian beam (MMGB) model to study the effects of material anisotropy on ultrasonic beam profile. It is shown that the anisotropic effects of beam skew and excess beam divergence enter into the MMGB model through parameters defining the slope and curvature of the slowness surface. The overall beam profile is found when the quasilongitudinal(qL) beam propagates in the symmetry plane of transversely isotropic austenitic steels. Simulation results are presented to illustrate the effects of these parameters on ultrasonic beam diffraction and beam skew. The MMGB calculations are also checked by comparing the anisotropy factor and beam skew angle with other analytical solutions.

• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.368-371
• /
• 1999

In this paper, we propose a neural network processor architecture with on-chip learning and with reconfigurability according to the data dependencies of the algorithm applied. For the neural network model applied, the proposed architecture can be configured into either SIMD or SRA(Systolic Ring Array) without my changing of on-chip configuration so as to obtain a high throughput. However, changing of system configuration can be controlled by user program. To process activation function, which needs amount of cycles to get its value, we design it by using PWL(Piece-Wise Linear) function approximation method. This unit has only single latency and the processing ability of non-linear function such as sigmoid gaussian function etc. And we verified the processing mechanism with EBP(Error Back-Propagation) model.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.34S no.6
• /
• pp.58-66
• /
• 1997

This paper proposes a multiresolution surface parameter estimation method for range images. Based on robust estimation of surface parameters, it approximates a patch to a planar surface in the locally adaptive window. Selection of resolution is made pixelwise by comparing a locally computed homogeneity measure with th eglobal threshold determined by te distribution of the approximation error. The proposed multiresolution surface parameter estimation method is applied to range image reconstruction. Computer simulation results with noisy rnag eimages contaminated by additive gaussian noise and impulse noise show that the proposed multiresolution reconstruction method well preserves step and roof edges compared with the conventional methods. Also the segmentation method based on the estimated surface parameters is shown to be robust to noise.

• Journal of Communications and Networks
• /
• v.16 no.6
• /
• pp.600-612
• /
• 2014

In this paper, we consider the problem of spectrum sensing for advanced television systems committee (ATSC) digital television (DTV) signal detection. To exploit the cyclostationarity of the ATSC DTV signals, we employ spectral correlation density (SCD) as the decision statistic and propose an optimal detection algorithm. The major difficulty is in obtaining the probability distribution functions of the SCD. To overcome the difficulty, we probabilistically model the pilot frequency location and employ Gaussian approximation for the SCD distribution. Then, we obtain a practically implementable detection algorithm that outperforms the industry leading systems by 2-3 dB. We also propose various techniques that greatly reduce the system complexity with performance degradation by only a few tenths of decibels. Finally, we show how robust the system is to the estimation errors of the noise power spectral density level and the probability distribution of the pilot frequency location.

• Journal of Communications and Networks
• /
• v.9 no.4
• /
• pp.466-472
• /
• 2007

Single carrier block transmission with frequency domain equalization(SC-FDE) has been shown to be a promising candidate in ultra-wideband(UWB) communications. In this paper, we analyze the performance of SC-FDE over UWB communications with channel estimation error. The probability density functions of the frequency domain minimum mean-squared error(MMSE) equalizer taps are derived in closed form. The error probabilities of single carrier block transmission with frequency domain MMSE equalization under imperfect channel estimation are presented and evaluated numerically. Compared with the simulation results, our semi-analytical analysis yields fairly accurate bit error rate performance, thus validating the use of the Gaussian approximation method in the performance analysis of the SC-FDE system with channel estimation error.

• Journal of Welding and Joining
• /
• v.31 no.4
• /
• pp.23-27
• /
• 2013

Laser heat source was used for automotive interior and exterior parts to reduce weights. Typically, 900's nm wavelength of laser has been widely used for polymer joining, however, the transmittance of the laser beam thorough clear polymers such as PMMA or PC has been an issue to overcome. To solve this issue, 1,940nm laser was applied on the clear polymer for the better absorption and 900nm laser beam was used for main laser for the joining. Conventional Gaussian or Elliptical heat source approximation has limitation in polymer which had deeper skin depth where major laser beam absorbs. To accurately simulate the physical laser beam absorption and joining optical properties were experimentally measured for the computer FEM simulation. The simulation results showed close correlation between theoretical and experimental results. The developed laser process is expected to increase productivity and gap closing which can cause failure of joining in laser material processing.

• Journal of Astronomy and Space Sciences
• /
• v.14 no.1
• /
• pp.67-79
• /
• 1997

We have designed the quasi optics of the dual channel receiver to observe the radio sources with 100GHz-band and 150GHz-band simultaneously. We introduced the general quasi optics and the relation between the Gaussian beam and thin lens approximation. We determined the parameters of the quasi optic components to match the beam waist at cassegrain focus with that of feed horn well.