• 전자공학회논문지
• /
• 제51권9호
• /
• pp.165-170
• /
• 2014

음성 인식을 위한 전처리기로 주변 잡음을 제거해 주는 음성향상 기법이 강조되고 있다. 다양한 음성향상 기법들 중 코드북 기반 음성향상 기법은 nonstationary 잡음 환경에서도 효율적으로 동작한다. 하지만, 기존 코드북 기반 음성향상 기법에서는 입력 신호와 음성 및 잡음 코드벡터 간에 미스매치가 발생하여 부정확한 게인이 추정되는 문제가 있다. 본 논문에서는 부정확한 게인을 보상하기 위해 long-term 잡음 추정 알고리즘을 사용하여 매 프레임 별로 신호 대 잡음비기반의 Normalized Weighting Factor (NWF)를 구하고, 이것을 기존 게인에 보상하는 방식을 제안한다. 제안된 코드북 기반 음성향상 기법은 기존 코드북 기반 음성향상 기법에 비해 향상된 성능을 보였다.

• 말소리와 음성과학
• /
• 제3권2호
• /
• pp.65-70
• /
• 2011

Recently, audio-visual speech recognition (AVSR) has been studied to cope with noise problems in speech recognition. In this paper we propose a novel method of deciding weighting factors for audio-visual information fusion. We adopt the particle swarm optimization (PSO) to weighting factor determination. The AVSR experiments show that PSO-based normalized confidence measures (NCM) improve the rejection performance of mis-recognized words by 33%.

• 전자공학회논문지B
• /
• 제33B권5호
• /
• pp.17-24
• /
• 1996

In this paper, we present a robust performance improvement method for the NLCF(normalized left coprime factor) uncertain structure using loop shaping and the structure singular value. For this, we select weighting functions for a loop shaping considering condition numer, and transform the NLCF uncertain structure into the 4-block structure. However, we can't get a good performance on account of the restriction of weighting functions. To cope with this, we motivate the use of structured singular vlaue in the robust performance improvement procedure. After all, the robust performance improvement can be obtained by a factor W$_{a}$ and a scaling factor of D-K iteration.

• 한국기계가공학회지
• /
• 제15권2호
• /
• pp.131-138
• /
• 2016

In order to improve the performance of an extended Kalman filter, a simplified indirect inference method (SIIM) fuzzy logic system (FLS) is proposed. The proposed FLS is composed of two fuzzy input variables, four fuzzy rules and one fuzzy output. Two normalized fuzzy input variables are the variance between the trace of a prior and a posterior covariance matrix, and the residual error of a Kalman algorithm. One fuzzy output variable is the weighting factor to adjust for the Kalman gain. There is no need to decide the number and the membership function of input variables, because we employ the normalized monotone increasing/decreasing function. The single parameter to be determined is the magnitude of a universe of discourse in the output variable. The structure of the proposed FLS is simple and easy to apply to various nonlinear state estimation problems. The simulation results show that the proposed FLS has strong adaptability to estimate the states of the incoming/outgoing moving objects, and outperforms the conventional extended Kalman filter algorithm by providing solutions that are more accurate.

• 한국정밀공학회지
• /
• 제21권2호
• /
• pp.153-159
• /
• 2004

A control-structure combined optimal design problem is discussed taking a 3-D truss structure as a design object. We use descriptor forms for a controlled object and a generalized plant because the structural parameters appear naturally in these forms. We consider not only minimum weight design problem for structure system, but also suppression problem of the effect of disturbances for control system as the purpose of the design. By minimizing the linear sum of the normalized structural objective function and control objective function, it is possible to make optimal design by which the balance of the structural weight and the control performance is taken.

• 대한음성학회지:말소리
• /
• 제47호
• /
• pp.131-140
• /
• 2003

We propose a new method of generating client models for HMM based text-dependent speaker verification system with only a small amount of training data. To make a client model, statistical methods such as segmental K-means algorithm are widely used, but they do not guarantee the quality or reliability of a model when only limited data are avaliable. In this paper, we propose a blind speech segmentation based on level building DTW algorithm as an alternative method to make a client model with limited data. In addition, considering the fact that voiced sounds have much more speaker-specific information than unvoiced sounds and energy of the former is higher than that of the latter, we also propose a new score evaluation method using the observation probability raised to the power of weighting factor estimated from the normalized log energy. Our experiment shows that the proposed methods are superior to conventional HMM based speaker verification system.

• 한국통신학회논문지
• /
• 제10권2호
• /
• pp.82-87
• /
• 1985

一定한 펄스反復周波數로 動作하는 MTI信號處理器에 대하여 MTI필터나 積分器의 傳達函數가 주어졌을 때 地上클러터에 대한 性能 比較 方法을 考察하였다. MTI信號處理器는 橫斷필터(transversal filter)로, 클러터의 電力密度스펙트럼은 Gaussian形態로 모델화하였고, MTI信號處理器에 대한 MTI改善因子(Improvement Factor)를 算出하여 比較하였다. MTI필터 및 Hanning加重函數를 使用한 積分器와 이들의 從屬接續인 경우에 대하여 正規化된 스펙트럼 너비에 대한 MTI改善因子의 變化를 例示하였다.

• 한국대기환경학회지
• /
• 제20권4호
• /
• pp.539-553
• /
• 2004

Up to the present time, many methods to estimate emissions from a particular diesel engines have wholly depended on the quantity of diesel fuel consumed. Then, the recommended emission factors were normalized by fuel consumption, and further total activity was estimated by the total fuel consumed. One of main purposes in the study is newly to develop emission factors for the railroad diesel rolling stock (RDRS) and to estimate a total amount of major gaseous pollutants from the RDRS in Korea. Prior to develop a Korean mode emission factor. the emission factor from the USEPA was simply applied for comparative studies. When applying the USEPA emission factors, total exhaust emissions from the RDRS in Korea were estimated by 28,117tons of NOx, 2,832.3tons of CO, and 1,237.5tons of HC, etc in 2001. In this study, a emission factor for the RDRS, so called the KoRail mode (the Korean Railroad mode) has been developed on the basis of analyzing the driving pattern of the Gyeongbu-Line especially for the line-haul mode. Explicitly to make the site specific emission factors, many uncertainty problems concerning weighting factors for each power mode, limited emission test, incomplete data for RDRS, and other important input parameters were extensively examined. Total exhaust emissions by KoRail mode in Korea were estimated by 10,960tons of NOx, and 4,622tons of CO, and so on in the year of 2001. The emissions estimated by the USEPA mode were 2.6 times higher for NOx, and 1.6 times lower for CO than those by the KoRail mode. As a conclusion, based on the emission calculated from both the USEPA mode and the KoRail mode, the RDRS is considered as one of the significant mobile sources for major gaseous pollutants and thus management plans an(1 control strategies for the RDRS must be established to improve air quality near future in Korea.