• 한국음향학회지
• /
• 제25권7호
• /
• pp.319-324
• /
• 2006

본 논문에서는 화자독립 가변어휘 핵심어 검출기의 성능을 개선하기 위하여 두 가지의 새로운 비핵심어 모델링 방법을 제안한다. 첫째는 K-means 알고리즘 기반 monophone 군집화 방법을 개선하기 위해 monophone을 state단위로 결정트리를 기반으로 군집화하여 비핵심어를 모델링하는 방법이다. 둘째는 single state multiple mixture 방법을 개선하기 위해 음절단위 multi-state multiple mixture 방법으로 모델링하는 방법이다. 실험에서 ETRI 표준 한국어 공통음성 단어 DB를 이용하여 트라이폰 모델을 훈련하였고, 훈련에 사용하지 않은 음성데이터를 이용하여 핵심어 검출closed 테스트를 수행하였다. 그리고 사무실 환경에서 4명의 화자가 각각 100문장씩 발성한 400문장의 음성데이터를 이용하여 100단어 핵심어 검출 open 테스트를 수행하였다. 실험 결과 결정트리기반 상태 군집화 방법이 기존의 K-means 알고리듬 기반 monophone clustering 방법보다 핵심어 검출 성능이 28%/29%(closed/open test) 향상되었다 그리고 음절단위 multi-state multiple mixture 방법이 비핵심어 전체를 single state 모델로 구성하는 방법보다 핵심어 검출 성능이 22%/2%(closed/open test) 향상됨으로써 본 논문에서 제안한 두 가지 알고리듬이 우수한 결과를 나타내었다

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제6권2호
• /
• pp.138-143
• /
• 2006

In this paper, we have presented a Sequential Agglomerative Hierarchical Nested (SAHN) algorithm-based data clustering method in fuzzy inference system to achieve optimal performance of fuzzy model. SAHN-based algorithm is used to give possible range of number of clusters with cluster centers for the system identification. The axes of membership functions of this fuzzy model are optimized by using cluster centers obtained from clustering method and the consequence parameters of the fuzzy model are identified by standard least square method. Finally, in this paper, we have observed our model's output performance using the Box and Jenkins's gas furnace data and Sugeno's non-linear process data.

• Wind and Structures
• /
• 제5권2_3_4호
• /
• pp.317-328
• /
• 2002

A numerical investigation on the turbulent flows over a three-dimensional steep hill is presented. The numerical model developed for the present work is based on the finite volume method and the SIMPLE algorithm with a non-staggered grid system. Standard $k-{\varepsilon}$ model and Shih's non-linear model are tested for the validation of the prediction accuracy in the 3D separated flow. Comparisons of the mean velocity and turbulence profiles between the numerical predictions and the measurements show good agreement. The Shih's non-linear model is found to predict mean flow and turbulence better than the Standard $k-{\varepsilon}$. Flow patterns have also been examined to explain the difference in the cavity zone between 2D and 3D hills.

• 약학회지
• /
• 제46권5호
• /
• pp.331-336
• /
• 2002

The purpose of this study is to develop a non-invasive blood glucose measurement method by a portable near infrared (NIR) system which was newly integrated by our lab. The portable NIR system includes a tungsten halogen lamp, a specialized reflectance fiber optic probe and a photo diode array type InGaAs detector; which was developed by a microchip technology based on the lithography. Reflectance NIR spectra of different parts of human body (finger tip, earlobe, and inner lip) were recorded by using a fiber optic probe. The spectra were collected over the spectral range 1100 ∼ 1740 nm. Partial least squares regression (PLSR) was applied for the calibration and validation for the determination of blood glucose. The calibration model from earlobe spectra presented better results, showing good correlation with a glucose oxidase method which is a mostly used standard method. This model predicted the glucose concentration for validation set with a SEP of 33 mg/dL. This study indicated the feasibility for non-invasive monitoring of blood glucose by a portable near infrared system.

• 한국철도학회논문집
• /
• 제9권1호
• /
• pp.63-68
• /
• 2006

This paper proposes a fast Fourier transform(FFT)-based spectral analysis method(SAM) for the dynamic responses of the linear discrete dynamic models with non-proportional damping. The SAM was developed by using discrete Fourier transform(DFT)-theory. To verify the proposed SAM, a three-DOF system with non-proportional viscous damping is considered as an illustrative example. The present SAM is evaluated by comparing the dynamic responses obtained by SAM with those obtained by Runge-Kutta method.

• 한국정밀공학회지
• /
• 제19권6호
• /
• pp.51-61
• /
• 2002

A tracking control scheme on the XY ball-screw drive system in the presence of nonlinear dynamic friction is proposed. A nonlinear dynamic friction is regarded as the Lund-Grenoble friction model to compensate effects of friction. The conventional VSC method that often has been used as a non-model-based friction controller has poor tracking performance in high-precision position tracking application since it cannot compensate the friction effect below a certain precision level completely. Thus to improve the precise position tracking performance, we propose the integral type VSC method combined with the friction-model-based observer. Then this control scheme has the high precise tracking performance compared with the non-model-baked VSC method and the PID control method with a similar observer. This fact is shown through the experiment on the XY ball-screw drive system with the nonlinear dynamic friction.

• 음성과학
• /
• 제9권1호
• /
• pp.111-135
• /
• 2002

An annotation model of a non-native speech database has been devised, wherein English is the target language and Korean is the native language. The proposed annotation model features overt transcription of predictable linguistic information in native speech by the dictionary entry and several predefined types of error specification found in native language transfer. The proposed model is, in that sense, different from other previously explored annotation models in the literature, most of which are based on native speech. The validity of the newly proposed model is revealed in its consistent annotation of 1) salient linguistic features of English, 2) contrastive linguistic features of English and Korean, 3) actual errors reported in the literature, and 4) the newly collected data in this study. The annotation method in this model adopts the widely accepted conventions, Speech Assessment Methods Phonetic Alphabet (SAMPA) and the TOnes and Break Indices (ToBI). In the proposed annotation model, SAMPA is exclusively employed for segmental transcription and ToBI for prosodic transcription. The annotation of non-native speech is used to assess speaking ability for English as Foreign Language (EFL) learners.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1996년도 한국자동제어학술회의논문집(국내학술편); 포항공과대학교, 포항; 24-26 Oct. 1996
• /
• pp.1-4
• /
• 1996

It is difficult that a non-translational motion in a block is estimated by the block matching algorithm (BMA). In this paper, a nodal-displacement-based deformation model is used for this reason. This model assumes that a selected number of control nodes move freely in a block and that displacement of any interior point can be interpolated from nodal displacements. As a special case with a single node this model is equivalent to a translational model. And this model can represent more complex deformation using more nodes. We used an iterative gradient based search algorithm to estimate nodal displacement. Each iteration involves the solution of a simple linear equation. This method is called the deformable block matching algorithm (DBMA).

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2007년도 추계학술대회 논문집
• /
• pp.7-10
• /
• 2007

The current model based sensorless method has many benefits that it can be robust control for large load torque. However, this method should determine a coefficient of back electro motive force(back-emf). This coefficient is varied by load torque and speed. Also, the coefficient determining equation is not exist, so it is determined only by experiment. On the other hands, using only back-emf estimatior method can not drive in low speed area and it has weakness in load variation. For these problems, this paper suggests the hybrid sensorless method that mixes the back-emf estimator regarding saliency and the current based sensorless model. This estimator offers not only non-necessary coefficient for current sensorless model, but also wide speed area operating in no specific transition method.

• Nuclear Engineering and Technology
• /
• 제55권12호
• /
• pp.4350-4356
• /
• 2023

Radiation detection systems working at high count rates suffer from the overlapping of their output electric pulses, known as pulse pile-up phenomenon, resulting in spectrum distortion and degradation of the energy resolution. Pulse tail extrapolation is a pile-up correction method which tries to restore the shifted baseline of a piled-up pulse by extrapolating the overlapped part of its preceding pulse. This needs a mathematical model which is almost always nonlinear, fitted usually by a nonlinear least squares (NLS) technique. NLS is an iterative, potentially time-consuming method. The main idea of the present study is to replace the NLS technique by an integration-based non-iterative method (NIM) for pulse tail extrapolation by an exponential model. The idea of linear extrapolation, as another non-iterative method, is also investigated. Analysis of experimental data of a NaI(Tl) radiation detector shows that the proposed non-iterative method is able to provide a corrected spectrum quite similar with the NLS method, with a dramatically reduced computation time and complexity of the algorithm. The linear extrapolation approach suffers from a poor energy resolution and throughput rate in comparison with NIM and NLS techniques, but provides the shortest computation time.