• 말소리와 음성과학
• /
• 제1권4호
• /
• pp.127-132
• /
• 2009

The feature vectors which are used in conventional speaker recognition (SR) systems may have many correlations between their neighbors. To improve the performance of the SR, many researchers adopted linear transformation method like principal component analysis (PCA). In general, the linear transformation of the feature vectors is based on concatenated form of the static features and their dynamic features. However, the linear transformation which based on both the static features and their dynamic features is more complex than that based on the static features alone due to the high order of the features. To overcome these problems, we propose an efficient method that applies linear transformation and temporal information of the features to reduce complexity and improve the performance in speaker verification (SV). The proposed method first performs a linear transformation by PCA coefficients. The delta parameters for temporal information are then obtained from the transformed features. The proposed method only requires 1/4 in the size of the covariance matrix compared with adding the static and their dynamic features for PCA coefficients. Also, the delta parameters are extracted from the linearly transformed features after the reduction of dimension in the static features. Compared with the PCA and conventional methods in terms of equal error rate (EER) in SV, the proposed method shows better performance while requiring less storage space and complexity.

• 한국정밀공학회지
• /
• 제26권3호
• /
• pp.88-97
• /
• 2009

The vacuum die casting techniques can diminish the porosity of products and provide better surface appearance by the ordinary high pressure die casting process. The vacuum system can also reduce the cold laps in the die casting process and minimize the overflow pockets of the die. The vacuum system does not need high pressures to die cast compared to the ordinary die casting process, and so enables die casting of large parts for a given machine size. Parts made by the vacuum system have higher strength and more elongation than parts made by the ordinary die casting systems. In this paper, we designed and produced the Magnesium seat frames using the vacuum die casting processes. The new Magnesium seat frame was designed to satisfy safety regulations. Some safety test procedures of the seat frame were simulated by the finite element method. We obtained 10% weight reduction by design modification of seat frames compared to the current model. Flow simulations were carried out to minimize the trial and error in producing the parts. The die casted parts using vacuum systems resulted in better mechanical characteristics and no defects compared to those without vacuum systems.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1991년도 하계학술대회 논문집
• /
• pp.657-665
• /
• 1991

Although orthogonal function is introduced in control theory in early 1970's, it is not perfect. Since the concept of integral operator by Chen and Hsiao in mid 1970's, orthogonal function (for example Walsh, Block-pulse, Haar, Laguerre, Legendre, Chebychev etc) has been widely applied In system's analysis and identification, model reduction, state estimation, optimal control, signal processing, image processing, EEG, and ECG etc. The reason why Walsh Functions introduces in control theory is that as integral of Walsh function is also developed in Walsh orthogonal function, if we transfer give system into integral equation and introduce Walsh function. We can know that system's characteristic by algebraical expression. This approach is based on least square error and that result is expressed as computer calculation and partly continuous constant value which is easy to apply. Such a Walsh function has been actively studied in USA, TAIWAN, INDO, CHINA, EUROPE etc and in domestic, author has studied it for 10 years since it was is introduced in 1982. This paper is consider the that author has studied for 10 years and Walsh function's efficiency.

• Nuclear Engineering and Technology
• /
• 제52권6호
• /
• pp.1327-1333
• /
• 2020

The global warming phenomenon emerges from the issue of climate change, which attracts the attention of intellectuals towards clean energy sources from dirty energy sources. Among clean sources, nuclear energy is getting immense attention among policymakers. However, the role of nuclear energy in pollution emissions reduction has remained inconclusive and demand for further investigation. Therefore, the current study contributes to extend knowledge by investigating the nexus between nuclear energy, economic growth, and CO₂ emissions in a developing country context such as Pakistan for the period between 1973 and 2017. The auto-regressive distributive lag model summarizes the nuclear energy has negative effect on environmental pollution as it releases carbon emission in the environment. Moreover, vector error correction Granger causality provides evidence for bidirectional causality between nuclear energy and carbon emissions. These interesting findings provide new insight, and policy guidelines provided based on these results.

• 한국정밀공학회지
• /
• 제21권10호
• /
• pp.127-136
• /
• 2004

This paper is the first of two companion papers concerning the analysis and design of a pneumatic vibration isolation system. The design optimization of the pneumatic vibration isolation system is required for the reduction of cost, endeavor and time, and it needs modeling and calculation algorithm. The nonlinear models are devised from the fluid mechanical expression for components of the system and the calculation algorithm is derived from the mathematical relationship between the models. It is shown that the orifice makes the nonlinear property of the transmissibility curve that the resonant frequency changes by the amplitude of excited vibration. Linearization of the nonlinear models is tried to reduce elapsed time and truncation error accumulation and to enable the transmissibility calculation of the system with multi damping chambers. The equivalent mechanical models generated by linearization clarify the function of each component of the system and lead to the linearized transfer function that can give forth to the transmissibility exactly close to that of nonlinear models. The modified successive under-relaxation method is developed to calculate the linearized transfer function.

• Radiation Oncology Journal
• /
• 제2권1호
• /
• pp.129-137
• /
• 1984

In brachytherapy, it is important to determine the positions of the radiation sources which are inserted into a patient and to estimate the dose resulting from the treatment. Calculation of the dose distribution throughout an implant is so laborious that it is rarely done by manual methods except for model cases. It is possible to calculate isodose distributions and tumor doses for individual patients by the use of a microcomputer. In this program, the dose rate and dose distributions are calculated by numerical integration of point source and the localization of radiation sources are obtained from two radiographs at right angles taken by a simulator developed for the treatment planning. By using microcomputer for brachytherapy, we obtained the result as following 1. Dose calculation and irradiation time for tumor could be calculated under one or five seconds after input data. 2. It was same value under$\pm2\%$ error between dose calculation by computer program and measurement dose. 3. It took about five minutes to reconstruct completely dose distribution for intracavitary irradiation. 4. Calculating by computer made remarkly reduction of dose errors compared with Quimby's calculation in interstitial radiation implantation. 5. It could calculate the biological isoffect dose for high and low dose rate activities.

• 한국유체기계학회 논문집
• /
• 제17권2호
• /
• pp.48-53
• /
• 2014

A low-noise regenerative blower is developed for fuel cell application by combining the FANDAS-Regen code and design optimization algorithm under several performance constraints for flow capacity, static pressure, efficiency and power consumption. The optimized blower design model is manufactured with some impeller modification based on low noise design concept and tested by using aerodynamic performance chamber facility and narrow-band noise measurement apparatus. The measured results of the optimized blower satisfy the performance requirements and are also compared favorably with the FANDAS-Regen prediction results within a few percent relative error. Furthermore, the present study shows the remarkable noise reduction by 26 dBA can be achieved through design optimization and low noise design concept.

• 로봇학회논문지
• /
• 제11권3호
• /
• pp.156-162
• /
• 2016

For the underwater localization, acoustic sensor systems are widely used due to greater penetration properties of acoustic signals in underwater environments. On the other hand, the good penetration property causes multipath and interference effects in structured environment too. To overcome this demerit, a localization method using the attenuation of electro-magnetic(EM) waves was proposed in several literatures, in which distance estimation and 2D-localization experiments show remarkable results. However, in 3D-localization application, the estimation difficulties increase due to the nonuniform (doughnut like) radiation pattern of an omni-directional antenna related to the depth direction. For solving this problem, we added a depth sensor for improving underwater 3D-localization with the EM wave method. A micro scale pressure sensor is located in the mobile node antenna, and the depth data from the pressure sensor is calibrated by the curve fitting algorithm. We adapted the depth(z) data to 3D EM wave pattern model for the error reduction of the localization. Finally, some experiments were executed for 3D localization with the fast calculation and less errors.

• 응용통계연구
• /
• 제18권1호
• /
• pp.137-146
• /
• 2005

신경망은 전통적인 시계열 기법들에 비해 대체적으로 예측성능의 우수함이 입증되었으나 계절성과 추세성을 갖는 시계열자료에 대해 예측력이 떨어지는 단점을 가지고 있다. 최근에는 Ensemble 기법인 Bagging Algorithm과 신경망의 혼합모형인 Bagging Neural Network이 개밭되었다. 이 기법은 분산과 편향을 많이 줄여줌으로써 더 좋은 예측을 할 수 있는 것으로 나타났다. 그러나 Ensemble 기법을 이용한 예측모형은 시계열자료를 적합 시키는데 있어 초기부여확률 및 예측자 선정시의 문제점을 가지고 있다. 이에 본 연구에서는 이러한 문제점을 해결하고 더불어 예측력을 향상시키기 위한 방법으로 초기부여확률이 균일분포가 아닌 순차적인 형태의 분포를 제시하고 신경망을 예측자로 활용한 변형된 Ensemble Algorithm을 제안한다. 또한 예측모형의 평가를 위해 실제자료를 가지고 기존 예측모형들과 제안한 방법을 이용하여 예측하고 각 MSE의 비교를 통하여 예측정확도를 알아보고자 한다.

• 유통과학연구
• /
• 제11권7호
• /
• pp.15-22
• /
• 2013

Purpose - The study investigates the influence of tourism and hospitality industry on economic growth and CO2 emissions. Research design, data, and methodology - In the empirical analysis, unit root tests, cointegration test and vector error correction model regression using time series data of South Korea from the first quarter of 1970 to the third quarter of 2010 are performed to examine the long-run equilibrium relationship and short-run dynamics among the tourism and hospitality industry, CO2 emissions, economic growth and other industry sectors. Results - Results indicate that a long-run equilibrium relationship exists among these variables. Furthermore, the tourism and hospitality industry and CO2 emissions have high significant positive effect on economic growth. The tourism and hospitality industry in Korea, in turns, shows a high significant positive impact on economic growth while the industry sector incursa high significant negative impact on CO2 emissions. Conclusions - The tourism and hospitality industry in Korea may havebeen prompted by several factors such as accelerated process of technological innovation or energy and environmental policies. These findings suggest that the effectively managed tourism and hospitality sector in Korea has resulted in both economic growth and a reduction in CO2 emissions.