• Journal of Computational Design and Engineering
• /
• 제3권3호
• /
• pp.179-190
• /
• 2016

A kind of modified epoxy resin sheet molding compounds of the impeller has been designed. Through the test, the non-metal impeller has a better environmental aging performance, but must do the waterproof processing design. In order to improve the stability of the impeller vibration design, the influence of uncertainty factors is considered, and a multi-objective robust optimization method is proposed to reduce the weight of the impeller. Firstly, based on the fluid-structure interaction, the analysis model of the impeller vibration is constructed. Secondly, the optimal approximate model of the impeller is constructed by using the Latin hypercube and radial basis function, and the fitting and optimization accuracy of the approximate model is improved by increasing the sample points. Finally, the micro multi-objective genetic algorithm is applied to the robust optimization of approximate model, and the Monte Carlo simulation and Sobol sampling techniques are used for reliability analysis. By comparing the results of the deterministic, different sigma levels and different materials, the multi-objective optimization of the SMC molding impeller can meet the requirements of engineering stability and lightweight. And the effectiveness of the proposed multi-objective robust optimization method is verified by the error analysis. After the SMC molding and the robust optimization of the impeller, the optimized rate reached 42.5%, which greatly improved the economic benefit, and greatly reduce the vibration of the ventilation system.

• 디지털융복합연구
• /
• 제14권10호
• /
• pp.271-276
• /
• 2016

본 논문에서는 조명 및 Pose 등의 다양한 환경변화에 강인한 얼굴 및 눈 검출 알고리즘을 제안한다. 일반적으로 눈 검출은 얼굴검출과 동시에 수행되며 조명 및 Pose의 변화에 따라 검출 성능에 영향을 준다. 본 논문에서는 Modified Census Transform 알고리즘 사용하여 환경변화에 강인한 얼굴검출을 수행한다. 눈은 얼굴영역의 중요한 특징으로 주변의 조명 변화 및 안경 등의 다양한 요인으로 검출 성능의 저하 요인이 된다. 이러한 문제점의 해결을 위하여 Gabor transformation과 Feature from Accelerated Segment Test 알고리즘 기반의 눈 검출 알고리즘을 제안한다. 제안된 얼굴검출 알고리즘은 27.4ms의 검출속도와 98.4%의 검출율을 보이며, 눈 검출 알고리즘의 경우 36.3ms의 검출속도와 96.4%의 검출율을 보이는 것을 확인하였다.

• 한국지능시스템학회논문지
• /
• 제19권6호
• /
• pp.773-778
• /
• 2009

본 논문에서는 인간의 감정 변화에 강인한 음성 인식 시스템을 구현하기 위하여 감정 변화의 영향을 최소화 하는 방법에 관한 연구를 수행하였다. 이를 위하여 우선 다양한 감정이 포함된 음성 데이터베이스를 사용하여 감정 변화에 따른 음성 신호의 변화를 관찰하였다. 감정이 포함되지 않은 평상의 음성으로 학습된 음성 인식 시스템에 감정이 포함된 인식 데이터가 입력되는 경우 감정에 따른 음성의 차이가 인식 시스템의 성능을 저하시킨다. 본 연구에서는 감정의 변화에 따라 화자의 성도 길이가 변화한다는 것과 이러한 변화는 음성 인식 시스템의 성능을 저하시키는 원인 중의 하나임을 관찰하였다. 본 연구에서는 이러한 음성의 변화를 감소시키는 방법으로 성도 길이 정규화 방법을 사용한 감정 변화에 강인한 음성 인식 시스템을 개발하였다. HMM을 사용한 단독음 인식 실험에서 제안된 학습 방법을 사용하면 감정 데이터의 오차가 기존방법보다 41.9% 감소되었다.

• 한국지능시스템학회논문지
• /
• 제20권4호
• /
• pp.528-533
• /
• 2010

본 논문에서는 인간의 감정 변화의 영향을 적게 받는 음성 인식 시스템의 학습 방법에 관한 연구를 수행하였다. 이를 위하여 우선 다양한 감정이 포함된 음성 데이터베이스를 사용하여 감정 변화가 음성 신호와 음성 인식 시스템의 성능에 미치는 영향에 관한 연구를 수행하였다. 감정이 포함되지 않은 평상의 음성으로 학습된 음성 인식 시스템에 감정이 포함된 인식 데이터가 입력되는 경우 감정에 따른 음성의 차이가 인식 시스템의 성능을 저하시킨다. 본 연구에서는 감정의 변화에 따라 화자의 성도 길이가 변화한다는 것과 이러한 변화는 음성 인식 시스템의 성능을 저하시키는 원인 중의 하나임을 관찰하였다. 본 연구에서는 이러한 음성의 변화를 포함하는 학습 방법을 제안하여 감정 변화에 강인한 음성 인식 시스템을 개발하였다. HMM을 사용한 단독음 인식 실험에서 제안된 학습 방법을 사용하면 감정 데이터의 오차가 기존 방법보다 28.4% 감소되었다.

• Communications for Statistical Applications and Methods
• /
• 제2권2호
• /
• pp.213-220
• /
• 1995

We consider a class of the weighted logrank tests and 4 types of weights in this class. We propese a test based on the maximum of 4 weighted logrank statistics and suggest a simulation techniqur to obtain the p-value of proposed test. It is shown through the simulation studies that the proposed test is robust and has reasonably good powers comparing with the well known efficient tests.

• The Journal of the Acoustical Society of Korea
• /
• 제28권3E호
• /
• pp.109-117
• /
• 2009

Identity recognition in real environment with a reliable mode is a key issue in human computer interaction (HCI). In this paper, we present a robust person identification system considering score-based optimal reliability measure of audio-visual modalities. We propose an extension of the modified reliability function by introducing optimizing parameters for both of audio and visual modalities. For degradation of visual signals, we have applied JPEG compression to test images. In addition, for creating mismatch in between enrollment and test session, acoustic Babble noises and artificial illumination have been added to test audio and visual signals, respectively. Local PCA has been used on both modalities to reduce the dimension of feature vector. We have applied a swarm intelligence algorithm, i.e., particle swarm optimization for optimizing the modified convection function's optimizing parameters. The overall person identification experiments are performed using VidTimit DB. Experimental results show that our proposed optimal reliability measures have effectively enhanced the identification accuracy of 7.73% and 8.18% at different illumination direction to visual signal and consequent Babble noises to audio signal, respectively, in comparison with the best classifier system in the fusion system and maintained the modality reliability statistics in terms of its performance; it thus verified the consistency of the proposed extension.

• 한국자동차공학회논문집
• /
• 제16권1호
• /
• pp.134-143
• /
• 2008

The hardware scatters as well as the engine parameters calibration have strong influences on exhaust emissions in recent diesel engines. In this research DoE(Design of Experiments) optimizations were done to study the possibility of minimizing the emission deviations caused by flow rate scatters of the injectors. It has been shown that the optimization of engine calibration, which minimizes the emission deviations, is feasible by establishing a target function representing the emission deviations for test results of maximum, mean and minimum flow rate injectors. It has also been shown that optimization of both emission deviations and emission level is possible by sequential DoE optimizations of the target functions representing the emission level and the emission deviations respectively with the appropriate boundary limits.

• Nuclear Engineering and Technology
• /
• 제55권7호
• /
• pp.2642-2649
• /
• 2023

A real-time unmeasured dynamic response prediction process for the nuclear power plant pressure pipeline is proposed and its performance is tested in the test-loop system (KAERI). The aim of the process is to predict unmeasurable or unreachable dynamic responses such as acceleration, velocity, and displacement by using a limited amount of directly measured physical responses. It is achieved by combining a well-constructed finite element model and robust inverse force identification algorithm. The pressure pipeline system is described by using the displacement-pressure vibro-acoustic formulation to consider fully filled liquid effect inside the pipeline structure. A robust multiphysics modal projection technique is employed for the real-time sensor synchronized prediction. The inverse force identification method is also derived and employed by using Bathe's time integration method to identify the full-field responses of the target system from the modal domain computation. To validate the performance of the proposed process, an experimental test is extensively performed on the nuclear power plant pressure pipeline test-loop under operation conditions. The results show that the proposed identification process could well estimate the unmeasured acceleration in both frequency and time domain faster than 32,768 samples per sec.

• 대한수학회논문집
• /
• 제20권1호
• /
• pp.145-159
• /
• 2005

In this paper we consider the problem of testing statistical hypotheses for unknown parameters in nonlinear regression models and propose three asymptotically equivalent tests based on regression quantiles estimators, which are Wald test, Lagrange Multiplier test and Likelihood Ratio test. We also derive the asymptotic distributions of the three test statistics both under the null hypotheses and under a sequence of local alternatives and verify that the asymptotic relative efficiency of the proposed test statistics with classical test based on least squares depends on the error distributions of the regression models. We give some examples to illustrate that the test based on the regression quantiles estimators performs better than the test based on the least squares estimators of the least absolute deviation estimators when the disturbance has asymmetric and heavy-tailed distribution.

• 대한기계학회논문집A
• /
• 제26권9호
• /
• pp.1931-1942
• /
• 2002

In this paper a formulation of robust optimization is presented and illustrated by a design example of vibratory micro gyroscopes in order to reduce the effect of variations due to uncertainties in MEMS fabrication processes. For the vibratory micro gyroscope considered it is important to match the resonance frequencies of the vertical (sensing) and lateral (driving) modes as close as possible to attain a high sensing sensitivity. A deterministic optimization in which the difference of both the sensing and driving natural frequencies is minimized as an objective function results in highly enhanced performance but apt to be very sensitive to fabrication errors. The formulation proposed is to attain robustness of the performance by including the sensitivity of the response with respect to uncertain variables as a term of objective function to be minimized. This formulation is simple and practically applicable since no detail statistical information on fabrication errors is required. The geometric variables, beam width, length and thickness of vibratory micro gyroscopes are adopted as design variables and at the same time considered as uncertain variables because here occur the fabrication errors. A robustness test in terms of a percentage yield by using the Monte Carlo simulation has shown that the robust optimum produces twice more acceptable designs than the deterministic optimum. Improvement of robustness becomes bigger as the amount of fabrication errors is assumed larger. Considering that the magnitude of fabrication errors and uncertainties in a MEMS structure are comparatively large, the present method is illustrated to be a viable approach for a robust MEMS design.