• 전기학회논문지P
• /
• 제59권2호
• /
• pp.163-172
• /
• 2010

A robust positioning control system has been studied using a friction parameter observer and a recurrent fuzzy neural network based on the sliding model. To estimate a nonlinear friction parameters of the LuGre friction model, a dual friction model-based observer is introduced. In addition, an approximating method for a system uncertainty has been developed using a recurrent fuzzy neural network technique to improve positioning performance. Experimental results have been presented to validate the performance of a proposed intelligent compensation scheme.

• 산업기술연구
• /
• 제3권
• /
• pp.95-101
• /
• 1983

This study aims at the determination of the coefficienties of runoff and infiltration affecting runoff. The rating curve is more available than the peak flood runoff to determine flood control plan of flood control reservoir and the volume of hydroelectric power plant, or to make multipurpose dam. In hydrologic analysis and design, it is necessary to develop relations between precipitation and runoff, possible using some of the factors affecting runoff as parameters. In order to calculate the runoff discharge, the runoff process constituting elements are divided to the surface runoff, the subsurface runoff and the groundwater runoff. By comparing the computed hydrograph with the measured hydrograph, determinned the watershed TANK Model constant Varying the tank model constant for approximating the computed hydrograph to the measured hydrograph.

• 응용통계연구
• /
• 제21권4호
• /
• pp.659-664
• /
• 2008

Missing continuous covariates are pervasive in the use of generalized linear models for medical data. Multiple imputation is the most common and easy-to-do method of dealing with missing covariate data. However, there are always serious warnings in using this method. There should be concern to make imputed values more proper. In this paper, proper imputation from posterior predictive distribution is developed for implementing with arbitrary priors. We use empirical distribution of the posterior for approximating the posterior predictive distribution, to sample from it. This method is preferable in comparison with a presented imputation method of us which uses a full model to impute missing values using available software. The proposed methods are implemented on glucocorticoid data.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 추계학술대회
• /
• pp.1846-1851
• /
• 2003

Approximate optimization has become popular in engineering field such as MDO and Crash analysis which is time consuming. To accomplish efficient approximate optimization, accuracy of approximate model is very important. As surrogate model, Kriging have been widely used approximating highly nonlinear system . Because Kriging employs interpolation method, it is adequate for deterministic computer simulation. Because there are no random errors and measurement errors in deterministic computer simulation, instead of classical DOE ,space filling experiment design which fills uniformly design space should be applied. In this work, various space filling designs such as maximin distance design, maximum entropy design are reviewed. And new design improving maximum entropy design is suggested and compared.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제5권2호
• /
• pp.145-150
• /
• 2005

We proposed the indirect adaptive fuzzy model based sliding mode controller to control nonaffine nonlinear systems. Takagi-Sugano fuzzy system is used to represent the nonaffine nonlinear system and then inverted to design the controller at each sampling time. Also sliding mode component is employed to eliminate the effects of disturbances, while a fuzzy model component equipped with an adaptation mechanism reduces modeling uncertainties by approximating model uncertainties. The proposed controller and adaptive laws guarantee that the closed-loop system is stable in the sense of Lyapunov and the output tracks a desired trajectory asymptotically.

• 대한전자공학회논문지
• /
• 제24권2호
• /
• pp.280-291
• /
• 1987

In this paper, the linear autoregressive model of EMG signal for four basic arm functions was presented and parameters for each function were estimated. The signal identification was carried out using function discrimination algorithm. It was validated that EMG signal was a widesense stationary process and the linear autoregressive model of EMG signal was constructed through approximating it to Gaussian process. It was confined that Levinson-Durbin algoridthm is a more appropriate one than the recursive least square method for parameter estimation of the linear model. Optimal function discrimination was acquired when sampling frequency was 500Hz and two electrodes were attached to bicep and tricep muscle, respectively. Parameter values were independent of variance and the number of minimum data for function discrimination was 200. Bayesian discrimination method turned out to be a better one than parallel filtering method for functional discrimination recognition.

• Smart Structures and Systems
• /
• 제29권5호
• /
• pp.729-746
• /
• 2022

In this paper, a nonlinear structural model updating methodology based on the Deep Belief Network (DBN) is proposed. Firstly, the instantaneous parameters of the vibration responses are obtained by the discrete analytical mode decomposition (DAMD) method and the Hilbert transform (HT). The instantaneous parameters are regarded as the independent variables, and the nonlinear model parameters are considered as the dependent variables. Then the DBN is utilized for approximating the nonlinear mapping relationship between them. At last, the instantaneous parameters of the measured vibration responses are fed into the well-trained DBN. Owing to the strong learning and generalization abilities of the DBN, the updated nonlinear model parameters can be directly estimated. Two nonlinear shear-type structure models under two types of excitation and various noise levels are adopted as numerical simulations to validate the effectiveness of the proposed approach. The nonlinear properties of the structure model are simulated via the hysteretic parameters of a Bouc-Wen model and a Giuffré-Menegotto-Pinto model, respectively. Besides, the proposed approach is verified by a three-story shear-type frame with a piezoelectric friction damper (PFD). Simulated and experimental results suggest that the nonlinear model updating approach has high computational efficiency and precision.

• 에너지공학
• /
• 제23권3호
• /
• pp.36-41
• /
• 2014

The world's overall need for natural gas (NG) has been growing up fast, especially in the residential sector. The better the estimation of residential NG consumption (RNGC) distribution, the better decision-making for a residential NG policy such as pricing, demand estimation, management options and so on. Approximating the distribution of RNGC is complicated by zero observations in the sample. To deal with the zero observations by allowing a point mass at zero, a mixture model of RNGC distributions is proposed and applied. The RNGC distribution is specified as a mixture of two distributions, one with a point mass at zero and the other with full support on the positive half of the real line. The model is empirically verified for household RNGC survey data collected in Korea. The mixture model can easily capture the common bimodality feature of the RNGC distribution. In addition, when covariates were added to the model, it was found that the probability that a household has non-expenditure significantly varies with some variables. Finally, the goodness-of-fit test suggests that the data are well represented by the mixture model.

• 자원ㆍ환경경제연구
• /
• 제12권3호
• /
• pp.453-467
• /
• 2003

CVM 연구에서 설문자료로부터 WTP 분포를 근사화하는 작업은 표본내에 포함된 0의 응답 때문에 문제가 발생할 수 있다. 이 때 WTP의 이중적 분포를 무시하면, 즉 연속적인 분포를 가정하면 후생치 도출에 있어 왜곡이 있을 수 있다. 본 연구는 이런 문제를 해결하기 위해 혼합모형을 제안하고 적용한다. 혼합모형은 영에서 점질량으로 이루어진 분포와 양의 연속적 분포를 혼합하는 특징을 갖고 있다. 이 경우 지불의사가 전혀 없는 확률은 모수 ${\rho}$로 표현되고, 별도로 식별가능하며, 최우추정법을 이용하면 일치추정량을 얻을 수 있다. 본 연구에서는 한려해상국립공원을 대상으로 혼합모형의 적합성에 관한 실증연구를 수행하였는데 그 결과 혼합모형의 적용이 바람직한 것으로 나타났다.

• 대한임베디드공학회논문지
• /
• 제16권4호
• /
• pp.145-151
• /
• 2021

In this paper, we propose a novel visualization technique to explain the predictions of deep neural networks. We use knowledge distillation (KD) to identify the interior of a black-box model for which we know only inputs and outputs. The information of the black box model will be transferred to a white box model that we aim to create through the KD. The white box model will learn the representation of the black-box model. Second, the white-box model generates attention maps for each of its layers using Grad-CAM. Then we combine the attention maps of different layers using the pixel-wise summation to generate a final saliency map that contains information from all layers of the model. The experiments show that the proposed technique found important layers and explained which part of the input is important. Saliency maps generated by the proposed technique performed better than those of Grad-CAM in deletion game.