• 말소리와 음성과학
• /
• 제3권1호
• /
• pp.87-94
• /
• 2011

The decision directed (DD) approach is widely used to determine a priori SNR from noisy speech signals. In conventional speech enhancement systems with a DD approach, a priori SNR is estimated by using only the magnitude components and consequently follows a posteriori SNR with one frame delay. We propose a phase-dependent two-step a priori SNR estimator based on the minimum mean square error (MMSE) in the log-mel spectral domain so that we can consider both magnitude and phase information, and it can overcome the performance degradation caused by one frame delay. From the experimental results, the proposed estimator is shown to improve the output SNR of enhanced speech signals by 2.3 dB compared to the conventional DD approach-based system.

• Spatial Information Research
• /
• 제19권2호
• /
• pp.29-36
• /
• 2011

본 연구는 표면 파라미터 추정시 고려하는 주요 인자별로 각 조정모델들을 분류하고 그들의 추정정확도를 사전분석함으로써 이들 모델링 인자들이 각 대상파라미터의 추정에 주는 영향을 정량적으로 분석하였다. 현재 지표면형상에 대한 정보를 취득하기 위하여 라이다영상, 항공영상, SAR영상 등 다양한 자료가 활용되고 있고, 이들로부터 지표면 형상을 정량적으로 분석하기 위해서는 임의지점 주위의 관측값들을 이용하여 해당 지점의 형상을 구체적으로 파악하게 된다. 이러한 형상정보는 관측값 범위지정, 가중치방식, 그리고 수학적모델링 등 여러 인자들을 선정하여 산정할 수 있지만, 각 선정인자에 따라 표면의 형상정보는 다르게 산정되고 또한 그 정확도도 상이하게 된다. 따라서, 본 연구에서는 표면의 형상정보추출시 조정계산 인자들 따른 이러한 정확도를 비교함으로써 인자별 추정 정확도 변화 경향에 대한 진단을 실시하였다. 본 연구에서는 표면형상정보로 표고, 경사, 곡면의 2차계수를 대상으로 하고, 수학적함수, 커널크기, 가중유형별로 조정계산모델들을 구성하여 사전통계량을 계산하였고, 이에 따라 전통계량 변화를 비교 분석함으로써, 각 조정모델의 추정성능을 조정계산인자에 따라 정량적으로 비교분석하였다.

• ETRI Journal
• /
• 제36권5호
• /
• pp.721-729
• /
• 2014

We propose a novel phase-based method for single-channel speech enhancement to extract and enhance the desired signals in noisy environments by utilizing the phase information. In the method, a phase-dependent a priori signal-to-noise ratio (SNR) is estimated in the log-mel spectral domain to utilize both the magnitude and phase information of input speech signals. The phase-dependent estimator is incorporated into the conventional magnitude-based decision-directed approach that recursively computes the a priori SNR from noisy speech. Additionally, we reduce the performance degradation owing to the one-frame delay of the estimated phase-dependent a priori SNR by using a minimum mean square error (MMSE)-based and maximum a posteriori (MAP)-based estimator. In our speech enhancement experiments, the proposed phase-dependent a priori SNR estimator is shown to improve the output SNR by 2.6 dB for both the MMSE-based and MAP-based estimator cases as compared to a conventional magnitude-based estimator.

• International Journal of Precision Engineering and Manufacturing
• /
• 제3권4호
• /
• pp.94-100
• /
• 2002

The conventional sliding mode control (SMC) technique requires a priori knowledge of the upperbounds of disturbances and/or modeling uncertainties to assure robustness. This, however, may not be easy to obtain in practical situation. This paper presents a new methodology, a sliding mode control with disturbance estimator (SMCDE), which offers a robust control performance without a priori knowledge about the disturbance. The proposed technique is featured by an average value of the imposed disturbance over a certain period. A nonlinear spring-mass-damper system and a two-link robot system are adopted as illustrative application examples. Control performances such as estimation error and tracking error are compared between the proposed methodology and conventional scheme.

• 한국정밀공학회지
• /
• 제19권1호
• /
• pp.165-171
• /
• 2002

The conventional sliding mode control(SMC) technique requires a priori knowledge of the upperbounds of disturbances or/and modeling uncertainties to assure robustness. This, however, may not to be easy to obtain in practical situation. This paper presents a new methodology, sliding mode control with disturbance estimator(SMCDE), which offers a robust control performance without a priori knowledge about the disturbance. The proposed technique is featured by an average value of the imposed disturbance over a certain period. A nonlinear spring-mass-damper system is adopted as an illustrative example, and a comparative work between the conventional technique and the present one is undertaken.

• 한국음향학회지
• /
• 제18권8호
• /
• pp.32-41
• /
• 1999

본 논문은 chirp신호와 두 개의 근거리 청음기를 이용한 해저퇴적층의 음향학적 특성치 역산기법을 제시한다. 역산문제를 확률론적 모델로 정식화하고, 역산의 해를 역산인자의 a priori분포와 유사도함수의 곱으로 표현되는a posteriori 확률분포로 정의하였다. 퇴적층의 음속과 층두께의 a priori정보를 파형 매칭 기법으로 추정한 후 다수의 퇴적층이 존재하는 환경모델을 부분퇴적층모델로 치환하고, 계측신호와 모의신호의 L₂노음을 이용하여 정의된 목적함수에 대해 반복적인 유전자알고리즘 탐색을 수행하여 탐색공간의 축소로 인한 탐색효율과 결과의 향상을 얻었다. A posteriori 확률분포의 다중적분의 형태로 정의되는 인자의 주변확률분포와 평균의 추정은 유전자알고리즘의 탐색과정에서 선택된 탐색점들을 이용하여 수행되었다. 제시된 역산기법의 검증을 위해 두 가지 퇴적층 환경모델을 설정하고 잡음을 첨가한 합성신호에 대해 역산기법을 적용하여 역산해를 추정하였고 역산결과로부터 본 역산기법의 유용성을 확인하였다.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2003년도 추계학술대회
• /
• pp.401-411
• /
• 2003

Tube expanding process is for combining a heat-sink plate with hair-pins (Cu-tube) through plastic deformation. The two parts, i. e. , heat-sink plate and hair-pins are they components of a heat-exchanger for an air conditioner. This paper presents the analysis and dynamic modeling of a high-speed tube expander which integrates transfer of parts, fixing of parts, and tube expanding into one process. The 3-dimensional modeling of all the parts for the tube-expander was constructed using CATIA$\circledR$. then the CATIA$\circledR$ models are transferred into visuaINastran$\circledR$ to execute the 3-dimensional animation for checking prescribed cycle-time. The technique presented in this paper has been shown to be effective as a priori tool for verifying the design of a high-speed tube expander.

• 전력전자학회논문지
• /
• 제21권3호
• /
• pp.207-214
• /
• 2016

Magnetic circuit is a physical modeling method that is useful in designing and analyzing power transformers, especially for a priori evaluation of leakage and magnetizing inductance before actual production. In this study, a novel modeling approach that uses PSIM magnetic elements adopting gyrator and permeance-capacitances is investigated. A formula to determine the permeance-capacitors in the core and leakage path are established, and a simulation jig is devised to link the physical model and the electrical terminal characteristics with an automated parameter determination process. The derived formula is verified by measurement results of the prototype transformer samples. Given its accuracy and simplicity, this approach is suitable for analyzing and designing LLC resonant transformers whose leakage and magnetizing inductance are very critical to circuit operation.

• The Journal of Korean Physical Therapy
• /
• 제22권1호
• /
• pp.75-81
• /
• 2010

Purpose: The purpose of this study was to apply association rule mining to explore the labyrinthine network of cerebral infarction comorbidity and basic data supply to develop cutting-edge physical therapy protocols for cerebral infarction with comorbidity Methods: From clinic records of enrollees of A Hospital in D city, patients over 18 years of age with cerebral infarction and cerebral infarction comorbidity were recruited as a case group. All diagnoses of that hospital were categorized according to the "International Classification of Disease (ICD)" diagnosis system. We extracted code I63 from the "Korea Classification of Disease (KCD)-4". Associated rule mining was done with a priori modeling and Web nodes to examine the strengths of associations among those diagnoses. The support and confidence values of associated rule mining results were examined. Results: The subjects of this study were 2,267 cerebral infarction patients. E11 (Non-insulin-dependent diabetes mellitus), E78 (Disorders of lipoprotein metabolism and other lipidaemias), G81 (Hemiplegia), I10 (Essential hypertension), and K29 (Gastritis and duodenitis) were high frequency diagnoses, being found in 10% or more of total diagnoses of cerebral infarction from frequency analysis results. The highest frequency diagnosis was 1,042 (46.0%) for I10. The second most frequent diagnosis was for E11(21.5%) while the third most frequent diagnosis was E78 (20.2%). Results from a priori modeling and Web nodes indicated that cerebral infarction has a strong association withessential hypertension, non-insulin-dependent diabetes mellitus, disorders of lipoprotein metabolism and other lipidaemias. Conclusion: Cerebral infarction is associated with hypertension, diabetes mellitus, and disorders of lipoprotein metabolism and other lipidaemias. The result of this study will be helpful to clinicians treating patients with cerebral infarction.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.159.1-159
• /
• 2001

A new time-domain FIR parameter estimation called the receding horizon least square estimation (RHLSE) is suggested for stochastic systems by combining the well known least square estimation with the receding horizon strategy. It can be always obtained without the requirement of any \textit{a priori} information about the horizon initial parameter. A fast algorithm for the suggested estimation is also presented which is remarkable in the view of computational advantage and simple implementation. It is shown that the proposed estimation is robust against temporary modeling uncertainties due to their FIR structure through simulation studies.