• Journal of Korea Water Resources Association
• /
• v.31 no.1
• /
• pp.13-25
• /
• 1998

It is necessary to develop methodologies for the application of artificial neural network into hydrologic rainfall-runoff process, although there is so much applicability by using the functions of associative memory based on recognition for the relationships between causes and effects and the excellent fitting capacity for the nonlinear phenomenon. In this study, some problems are presented in the application procedures of artificial neural networks and the simulation of runoff hydrograph experiences are reviewed with nonlinear functional approximator by artificial neural network for rainfall-runoff relationships in a watershed. which is regarded as hydrdologic black box model. The neural network models are constructed by organizing input and output patterns with the deserved rainfall and runoff data in Pyoungchang river basin under the assumption that the rainfall data is the input pattern and runoff hydrograph is the output patterns. Analyzed with the results. it is possible to simulate the runoff hydrograph with processing element of artificial neural network with any hydrologic concepts and the weight among processing elements are well-adapted as model parameters with the assumed model structure during learning process. Based upon these results. it is expected that neural network theory can be utilized as an efficient approach to simulate runoff hydrograph and identify the relationship between rainfall and runoff as hydrosystems which is necessary to develop and manage water resources.

• Journal of Korean Association for Spatial Structures
• /
• v.20 no.2
• /
• pp.51-58
• /
• 2020

Recently, deep learning that is the most popular and effective class of machine learning algorithms is widely applied to various industrial areas. A number of research on various topics about structural engineering was performed by using artificial neural networks, such as structural design optimization, vibration control and system identification etc. When nonlinear semi-active structural control devices are applied to building structure, a lot of computational effort is required to predict dynamic structural responses of finite element method (FEM) model for development of control algorithm. To solve this problem, an artificial neural network model was developed in this study. Among various deep learning algorithms, a recurrent neural network (RNN) was used to make the time history response prediction model. An RNN can retain state from one iteration to the next by using its own output as input for the next step. An eleven-story building structure with semi-active tuned mass damper (TMD) was used as an example structure. The semi-active TMD was composed of magnetorheological damper. Five historical earthquakes and five artificial ground motions were used as ground excitations for training of an RNN model. Another artificial ground motion that was not used for training was used for verification of the developed RNN model. Parametric studies on various hyper-parameters including number of hidden layers, sequence length, number of LSTM cells, etc. After appropriate training iteration of the RNN model with proper hyper-parameters, the RNN model for prediction of seismic responses of the building structure with semi-active TMD was developed. The developed RNN model can effectively provide very accurate seismic responses compared to the FEM model.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4
• /
• pp.36-42
• /
• 1996

In this paper, we have proposed the robust estimation of HMM parameters which is based on CDDN(codeword dependent distance normalization)and codeword weighting by distance. The proposed method has used a distance normalization based on the characteristics of a codeword dependent distribution and have computed fuzzy contributions of codeword to a input vector with a fuzzy objective function. From experimental results, we have shown the effectiveness of the proposed method in that the correction rate of the proposed method is improved 4.5% over the conventional FVQ based method. Especially, the application of distance weighting to smoothing of output probability is improved the performance of 2.5% compared to distance based codeword weighting.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.16 no.3
• /
• pp.20-28
• /
• 2002

This paper describes the time sharing type high frequency resonant inviter can be used as power of induction heating. This closed inverter can be obtained output frequency three times than switching frequency by composing three unit inviter of conventional Half-Bridge serial resonant inverter in parallel with input power source also, this reduce switching loss because it has ZVS function. The analysis of the proposed circuit is generally described by using the normailized proposed parameters. The principle of basic operating and the its charasteristics are extimated by the parameters such as switching frequency($\mu$), the variation of Phase angle($\phi$) of Phase-shift. Experimental results are presented to verify theoretical discussion. This preposed inverter will be able to be prastically used as a power supply in various fields as induction, heating application, DC-DC converter etc.

• Journal of the Korean Geotechnical Society
• /
• v.18 no.5
• /
• pp.37-42
• /
• 2002

Artificial neural networks are efficient computing techniques that are widely used to solve complex problems in many fields. In this paper a liquefaction potential was estimated by using a back propagation neural network model applicated to cyclic triaxial test data, soil parameters and site investigation data. Training and testing of the network were based on a database of 43 cyclic triaxial test data from 00 sites. The neural networks are trained by modifying the weights of the neurons in response to the errors between the actual output values and the target output value. Training was done iteratively until the average sum squared errors over all the training patterns were minimized. This generally occurred after about 15,000 cycles of training. The accuracy from 72% to 98% was shown for the model equipped with two hidden layers and ten input variables. Important effective input variables have been identified as the NOC,$D_10$ and (N$_1$)$_60$. The study showed that the neural network model predicted a CSR(Cyclic shear stress Ratio) of silty-sand reasonably well. Analyzed results indicate that the neural-network model is more reliable than simplified method using N value of SPT.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.42 no.7 s.337
• /
• pp.27-34
• /
• 2005

Iterative detection (ID) has proven to be a near-optimal solution for concatenated Finite State Machines (FSMs) with interleavers over an additive white Gaussian noise (AWGN) channel. When perfect channel state information (CSI) is not available at the receiver, an adaptive ID (AID) scheme is required to deal with the unknown, and possibly time-varying parameters. The basic building block for ID or AID is the soft-input soft-output (SISO) or adaptive SISO (A-SISO) module. In this paper, Reduced State SISO (RS-SISO) algorithms have been applied for complexity reduction of the A-SISO module. We show that serially concatenated CPM (SCCPM) with AID has turbo-like performance over fading ISI channels and also RS-A-SISO systems have large iteration gains. Various design options for RS-A-SISO algorithms are evaluated. Recently developed density evolution technique is used to analyze RS-A-SISO algorithms. We show that density evolution technique that is usually used for AWGN systems is also a good analysis tool for RS-A-SISO systems over frequency-selective fading channels.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.37 no.1
• /
• pp.8-18
• /
• 2000

We presented a BGA(Ball Grid Array) package for RF circuit modules and extracted its electrical parameters. As the frequency of RF system devices increases, the effect of its electrical parasitics in the wireless communication system requires new structure of RF circuit modules because of its needs to be considered of electrical performance for minimization and module mobility. RF circuit modules with BGA packages can provide some advantages such as minimization, shorter circuit routing, and noise improvement by reducing electrical noise affected to analog and digital mixed circuits, etc. We constructed a BGA package of ITS(Intelligent Transportation System) RF module and measured electrical parameters with a TDR(Time Domain Reflectometry) equipment and compared its electrical parasitic parameters with PCB RF circuits. With a BGA substrate of 3${\times}$3 input and output terminals, we have found that self capacitance of BGA solder ball is 68.6fF, and self inductance 146pH, whose values were reduced to 34% and 47% of the value of QFP package structure. S11 parameter measurement with a HP4396B Network Analyzer showed the resonance frequency of 1.55GHz and the loss of 0.26dB. Routing length of the substrate was reduced to 39.8mm. Thus, we may improve electrical performance when we use BGA package structures in the design of RF circuit modules.

• Proceedings of the KSRS Conference
• /
• 2008.10a
• /
• pp.220-223
• /
• 2008

A camera system for the satellite application performs the mission of observation by measuring radiated light energy from the target on the earth. As a development stage of the system, the signal level analysis by estimating the number of electron collected in a pixel of an applied CCD is a basic tool for the performance analysis like SNR as well as the data path design of focal plane electronic. In this paper, two methods are presented for the calculation of the number of electrons for signal level analysis. One method is a quantitative assessment based on the CCD characteristics and design parameters of optical module of the system itself in which optical module works for concentrating the light energy onto the focal plane where CCD is located to convert light energy into electrical signal. The other method compares the design\ parameters of the system such as quantum efficiency, focal length and the aperture size of the optics in comparison with existing camera system in orbit. By this way, relative count of electrons to the existing camera system is estimated. The number of electrons, as signal level of the camera system, calculated by described methods is used to design input circuits of AD converter for interfacing the image signal coming from the CCD module in the focal plane electronics. This number is also used for the analysis of the signal level of the CCD output which is critical parameter to design data path between CCD and A/D converter. The FPE(Focal Plane Electronics) designer should decide whether the dividing-circuit is necessary or not between them from the analysis. If it is necessary, the optimized dividing factor of the level should be implemented. This paper describes the analysis of the electron count of a camera system for a satellite application and then of the signal level for the interface design between CCD and A/D converter using two methods. One is a quantitative assessment based on the design parameters of the camera system, the other method compares the design parameters in comparison with those of the existing camera system in orbit for relative counting of the electrons and the signal level estimation. Chapter 2 describes the radiometry of the camera system of a satellite application to show equations for electron counting, Chapter 3 describes a camera system briefly to explain the data flow of imagery information from CCD and Chapter 4 explains the two methods for the analysis of the number of electrons and the signal level. Then conclusion is made in chapter 5.

• Korean Journal of Optics and Photonics
• /
• v.19 no.3
• /
• pp.229-236
• /
• 2008

The optimum conditions of second harmonic generation (SHG) can be successfully achieved experimentally using single pass and double pass methods of a pumping beam. The beam has a power of several Watts radiated by a DOFA (Diode Laser Oscillator & Fiber Amplifier) system, which is a high power CW wavelength tunable infrared laser system, in a PPLN (Periodically Poled MgO doped Lithium Niobate) nonlinear crystal. In the case of a single pass method, the parameters are the wavelength of 535 nm for SHG and the output power of 245 mW generated from the pumping input beam with wavelength of 1070 nm and the power of 2.45 W at phase matching temperature of $108.9^{\circ}C$. The conversion efficiency of SHG was 10%. In order to enhance the output of SHG, the double pass method of the SHG system of a PPLN using a concave mirror for the retroreflection and a pair of wedged flat windows for phase compensation was also presented. In this double pass system, we obtained the SHG output beam with the wavelength of 535 nm and the maximum power of 383 mW at optimum phase matching temperature of $108.5^{\circ}C$ by using an incident pumping beam with wavelength of 1070 nm and the power of 2.45 W. The maximum conversion efficiency is 15.6%, which is more than that of the single pass method.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.7
• /
• pp.970-976
• /
• 2007

In this study, Polynomial Network Pattern Classifier(PNC) based on Fuzzy Inference Mechanism is designed and its parameters such as learning rate, momentum coefficient and fuzzification coefficient are optimized by means of Particle Swarm Optimization. The proposed PNC employes a partition function created by Fuzzy C-means(FCM) clustering as an activation function in hidden layer and polynomials weights between hidden layer and output layer. Using polynomials weights can help to improve the characteristic of the linear classification of basic neural networks classifier. In the viewpoint of linguistic analysis, the proposed classifier is expressed as a collection of "If-then" fuzzy rules. Namely, architecture of networks is constructed by three functional modules that are condition part, conclusion part and inference part. The condition part relates to the partition function of input space using FCM clustering. In the conclusion part, a polynomial function caries out the presentation of a partitioned local space. Lastly, the output of networks is gotten by fuzzy inference in the inference part. The proposed PNC generates a nonlinear discernment function in the output space and has the better performance of pattern classification as a classifier, because of the characteristic of polynomial based fuzzy inference of PNC.