• Proceedings of the IEEK Conference
• /
• 2001.06c
• /
• pp.25-28
• /
• 2001

RBFN has some problem that because the basis function isn't orthogonal to each others the number of used basis function goes to big. In this reason, the Wavelet Neural Network which uses the orthogonal basis function in the hidden node appears. In this paper, we propose the composition method of the actual function in hidden layer with the scaling function which can represent the region by which the several wavelet can be represented. In this method, we can decrease the size of the network with the pure several wavelet function. In addition to, when we determine the parameters of the scaling function we can process rough approximation and then the network becomes more stable. The other wavelets can be determined by the global solutions which is suitable for the suggested problem using the genetic algorithm and also, we use the back-propagation algorithm in the learning of the weights. In this step, we approximate the target function with fine tuning level. The complex neural network suggested In this paper is a new structure and important simultaneously in the point of handling the determination problem in the wavelet initialization.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.20 no.1
• /
• pp.22-28
• /
• 2010

Approximate analysis for a building installed with a friction damper is performed to get insight of its dynamic behavior. Energy balance equation is used to have a closed analytical form solution of dynamic magnification factor(DMF). It is found out that DMF is dependent on friction force ratio and resonance frequency. Approximation of DMF and equivalent damping ratio of a friction damper is proposed with such assumption that the building with a friction damper shows harmonic steady-state response and narrow banded response behavior near resonance frequency. Linear transfer function from input external force to output building displacement is suggested from the simplified DMF equation. Root mean square of a building displacement is derived under earthquake-like random excitation. Finally, design procedure of a friction damper is proposed by finding friction force corresponding to target control ratio. Numerical analysis is carried out to verify the proposed design procedure.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2005.06a
• /
• pp.387-392
• /
• 2005

Double-deck train have studied in the next generation train in KRRI. Double-deck train have more seat capacities compared with single deck vehicles and is a efficient, reliable and comfortable alternative train. Because of heavy weight, weight minimization of double-deck train carbody is imperative to reduce cost and extend life-time of train. Weight minimization problem of the double-deck train car-body is required to decide 66 design variables of thicknesses for large aluminum extruded panel while satisfying stress constraints. Design variables are too many and one execution of structural analysis of double-deck train carbody is time-consuming. Therefore, we adopt approximation technique to save computational cost of optimization process. Metamodels such as response surface model (RSM) and kriging model are used to approximate model-based optimization is described. RSM is easy to obtain and expressed explicit function, but this is not suitable for highly nonlinear and large scaled problems. Kriging model employs an interpolation scheme and is developed in the fields of spatial statistics and geostatistics. Target of this design is to find optimum thickness of AEP to minimize weight of doulbe-deck train carbody. In this study, meta model techniques are introduced to carry out weight minimization of a double-deck train car-body.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.3
• /
• pp.206-213
• /
• 2004

This paper presents a steering angle error compensation (SAEC) algorithm that is appropriate for rapidly moving sources. The Proposed algorithm utilizes a modal covariance matrix from multiple frequency components instead of the multiple snapshots in a narrowband SAEC, and estimates the steering error by maximizing the wideband WVDR output power using a first-order Taylor series approximation of the modal steering vector in terms of the steering error. As such, the steering error can be compensated with short observation times. Several simulations using artificial and sea trial data are used to demonstrate the Performance of the proposed algorithm.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.38 no.5
• /
• pp.1-6
• /
• 2001

In this paper, we propose the initial optimized structure of the Radial Basis Function Network(RBFN) which is more simple in the part of the structure and converges more faster than Neural Network. For this, we use the analysis method using time frequency localization and we can decide the initial structure of the RBFN suitable for the given problem. When we compose the hidden nodes with the radial basis functions whose localization are similar with the target function in the plane of the time and frequency, we can make a good decision of the initial structure having an ability of approximation.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.17 no.5
• /
• pp.680-687
• /
• 2014

PGA(phase gradient autofocus) has been widely used to remove motion induced phase errors in the ISAR(inverse synthetic aperture radar) imaging. The critical process for the processing time and image quality is windowing stage in PGA. In this paper, the new method to determine window size based on polynomial least square approximation is proposed. Moreover, dominant range bins are selected for efficient phase error estimation, which improve image quality and speed up convergence. The simulation results show that the proposed algorithm provides high quality ISAR images while computational efficiency of inherent PGA is retained.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.303-306
• /
• 2003

In radar tracking, since the sensor measures range, azimuth and elevation angle of a target, the measurement equation is nonlinear and the extended Kalman filter (EKF) is applied to nonlinear estimation. The conventional EKF has been widely used as a nonlinear filter for radar tracking, but the considerably large measurement error due to the linearization of nonlinear function in highly nonlinear situations may deteriorate the performance of the EKF To solve this problem, a fuzzy-model-based Kalman filter (FMBKF) is proposed for radar tracking. The FMBKF uses a local model approximation based on a TS fuzzy model instead of a Jacobian matrix to linearize nonlinear measurement equation. The hybrid GA and RLS method is used to identify the premise and the consequent parameters and the rule numbers of this TS fuzzy model. In two-dimensional radar tracking problem, the proposed method is compared with the conventional EKF.

• ETRI Journal
• /
• v.27 no.6
• /
• pp.747-758
• /
• 2005

The problem of selection among competing models has been a fundamental issue in statistical data analysis. Good fits to data can be misleading since they can result from properties of the model that have nothing to do with it being a close approximation to the source distribution of interest (for example, overfitting). In this study we focus on the preference among models from a family of polynomial regressors. Three decades of research has spawned a number of plausible techniques for the selection of models, namely, Akaike's Finite Prediction Error (FPE) and Information Criterion (AIC), Schwartz's criterion (SCH), Generalized Cross Validation (GCV), Wallace's Minimum Message Length (MML), Minimum Description Length (MDL), and Vapnik's Structural Risk Minimization (SRM). The fundamental similarity between all these principles is their attempt to define an appropriate balance between the complexity of models and their ability to explain the data. This paper presents an empirical study of the above principles in the context of model selection, where the models under consideration are univariate polynomials. The paper includes a detailed empirical evaluation of the model selection methods on six target functions, with varying sample sizes and added Gaussian noise. The results from the study appear to provide strong evidence in support of the MML- and SRM- based methods over the other standard approaches (FPE, AIC, SCH and GCV).

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.05a
• /
• pp.65-68
• /
• 2001

RBFN has some problem that because the basis function isnt orthogonal to each others the number of used basis function goes to big. In this reason, the Wavelet Neural Network which uses the orthogonal basis function in the hidden node appears. In this paper, we propose the composition method of the actual function in hidden layer with the scaling function which can represent the region by which the several wavelet can be represented. In this method, we can decrease the size of the network with the pure several wavelet function. In addition to, when we determine the parameters of the scaling function we can process rough approximation and then the network becomes more stable. The other wavelets can be determined by the global solutions which is suitable for the suggested problem using the genetic algorithm and also, we use the back-propagation algorithm in the learning of the weights. In this step, we approximate the target function with fine tuning level. The complex neural network suggested in this paper is a new structure and important simultaneously in the point of handling the determination problem in the wavelet initialization.

• Wind and Structures
• /
• v.13 no.3
• /
• pp.235-256
• /
• 2010

The aeroelastic stability of bridge decks equipped with multiple tuned mass dampers is studied. The problem is attacked in the time domain, by representing self-excited loads with the aid of aerodynamic indicial functions approximated by truncated series of exponential filters. This approach allows to reduce the aeroelastic stability analysis in the form of a direct eigenvalue problem, by introducing an additional state variable for each exponential term adopted in the approximation of indicial functions. A general probabilistic framework for the optimal robust design of multiple tuned mass dampers is proposed, in which all possible sources of uncertainties can be accounted for. For the purposes of this study, the method is also simplified in a form which requires a lower computational effort and it is then applied to a general case study in order to analyze the control effectiveness of regular and irregular multiple tuned mass dampers. A special care is devoted to mistuning effects caused by random variations of the target frequency. Regular multiple tuned mass dampers are seen to improve both control effectiveness and robustness with respect to single tuned mass dampers. However, those devices exhibit an asymmetric behavior with respect to frequency mistuning, which may weaken their feasibility for technical applications. In order to overcome this drawback, an irregular multiple tuned mass damper is conceived which is based on unequal mass distribution. The optimal design of this device is finally pursued via a full domain search, which evidences a remarkable robustness against frequency mistuning, in the sense of the simplified design approach.