• 대한음성학회지:말소리
• /
• 제55권
• /
• pp.131-142
• /
• 2005

A mismatch between the training and the test conditions often causes a drastic decrease in the performance of the speech recognition systems. In this paper, non-linear transformation techniques based on histogram equalization in the acoustic feature space are studied for reducing the mismatched condition. The purpose of histogram equalization(HEQ) is to convert the probability distribution of test speech into the probability distribution of training speech. While conventional histogram equalization methods consider only the probability distribution of a test speech, for noise-corrupted test speech, its probability distribution is also distorted. The transformation function obtained by this distorted probability distribution maybe bring about miss-transformation of feature vectors, and this causes the performance of histogram equalization to decrease. Therefore, this paper proposes a new method of calculating noise-removed probability distribution by using assumption that the CDF of noisy speech feature vectors consists of component of speech feature vectors and component of noise feature vectors, and this compensated probability distribution is used in HEQ process. In the AURORA-2 framework, the proposed method reduced the error rate by over $44\%$ in clean training condition compared to the baseline system. For multi training condition, the proposed methods are also better than the baseline system.

• Structural Engineering and Mechanics
• /
• 제68권4호
• /
• pp.423-442
• /
• 2018

In this paper, masonry infilled reinforced concrete (RC) frames are analyzed through a probabilistic approach. A macro-modeling technique, based on an equivalent diagonal pin-jointed strut, has been resorted to for modelling the stiffening contribution of the masonry panels. Since it is quite difficult to decide which mechanical characteristics to assume for the diagonal struts in such simplified model, the strut width is here considered as a random variable, whose stochastic characterization stems from a wide set of empirical expressions proposed in the literature. The stochastic analysis of the masonry infilled RC frame is conducted via the Probabilistic Transformation Method by employing a set of space transformation laws of random vectors to determine the probability density function (PDF) of the system response in a direct manner. The knowledge of the PDF of a set of response indicators, including displacements, bending moments, shear forces, interstory drifts, opens an interesting discussion about the influence of the uncertainty of the masonry infills and the resulting implications in a design process.

• Structural Engineering and Mechanics
• /
• 제67권5호
• /
• pp.439-455
• /
• 2018

Within a probabilistic framework, this paper addresses the determination of the static structural response of beams and frames with partially restrained (semi-rigid) connections. The flexibility of the nodal connections is incorporated via an idealized linear-elastic behavior of the beam constraints through the use of rotational springs, which are here considered uncertain for taking into account the largely scattered results observed in experimental findings. The analysis is conducted via the Probabilistic Transformation Method, by modelling the spring stiffness terms (or equivalently, the fixity factors of the beam) as uniformly distributed random variables. The limit values of the Eurocode 3 fixity factors for steel semi-rigid connections are assumed. The exact probability density function of a few indicators of the structural response is derived and discussed in order to identify to what extent the uncertainty of the beam constraints affects the resulting beam response. Some design considerations arise which point out the paramount importance of probability-based approaches whenever a comprehensive experimental background regarding the stiffness of the beam connection is lacking, for example in steel frames with semi-rigid connections or in precast reinforced concrete framed structures. Indeed, it is demonstrated that resorting to deterministic approaches may lead to misleading (and in some cases non-conservative) outcomes from a design viewpoint.

• 한국해안해양공학회지
• /
• 제2권2호
• /
• pp.83-95
• /
• 1990

현재 해안공학에 있어서 불규칙파랑을 표현하는 방법으로서 스펙트럼분석법과 파별해석법이 있다. 그 중,스펙트럼해석법은 수심이 깊고 입사파랑의 분산성이 큰 영역에, 파랑분석법은 수심이 얕고 비선형성이 강한 영역 및 석파와 같은 불연속현상을 취급할 경우에 각각적용 되어진다. 그러므로 현지 불규칙파랑장을 계통적으로 취급하기 위해서는 어느 수심영역에서 스펙트럼해석으로부터 파별분석으로 전의해야 할 것인가에 대한 연구가 필요하게 된다. 본 연구는 현지 불규칙파랑의 전파에 의한 변형을 계통적으로 취급할 수 있는 구체적인 방법을 확립하기 위해서 행한 일연의 연구로서 입사파랑의 방향분석이 없는 2차원 불규칙파에 대한 천수을 행하여 스펙트럼해석법의 적용한계, 스펙트럼에서 추정되어지는 파고, 주기의 결합확률분포, 불규칙파의 천수, 석파변형에 대한 파별분석법의 적용한계 등에 대해서 검토한 것이다.

• 대한전기학회논문지:전력기술부문A
• /
• 제48권9호
• /
• pp.1055-1063
• /
• 1999

This paper present a conceptual possibilistic approach using fuzzy set theory to manage the uncertainties in the given reliability input date of the practical power system. In this paper, an algorithm is introduced to calculate the possibilstic reliability indices according to the degree of uncertainty in the given data. The probability distribution function can be transformed into an appropriate possibilstic representation using the probability-Possibility Consistency principle(PPCP) algorithm. In this the algorithm, the transformation is performation by making a compromise between the transformation consistency and the human updating experience. Fuzzy classifcation theory is applied to reduced the number of load data. The fuzzy classification method determines the closeness of load data points by assigning them to various clusters and then determening the distance between the clusters. The IEEE-RTS with 32-generating units is used to demonstrate the capability of the proposed algorithm.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.408-413
• /
• 2005

A new methodology for discrete non-Gaussian probability density estimation is investigated in this paper based on a dynamic Bayesian network (DBN) and kernel functions. The estimator consists of a DBN in which the transition distribution is represented with kernel functions. The estimator parameters are determined through a recursive learning algorithm according to the maximum likelihood (ML) scheme. A discrete-type Poisson distribution is generated in a simulation experiment to evaluate the proposed method. In addition, an unknown probability density generated by nonlinear transformation of a Poisson random variable is simulated. Computer simulations numerically demonstrate that the method successfully estimates the unknown probability distribution function (PDF).

• 한국농공학회논문집
• /
• 제55권5호
• /
• pp.49-58
• /
• 2013

Appropriate random variables and probability density functions based on statistical analysis should be defined to execute reliability analysis. Most studies have focused on only normal distributions or assumed that the variables showing non-normal characteristics follow the normal distributions. In this study, the reliability problem with non-normal probability distribution was dealt with using the convolution method in the case that the integration domains of variables are limited to a finite range. The results were compared with the traditional method (linear transformation of normal distribution) and Monte Carlo simulation method to verify that the application was in good agreement with the characteristics of probability density functions with peak shapes. However it was observed that the reproducibility was slightly reduced down in the tail parts of density function.

• International Journal of Precision Engineering and Manufacturing
• /
• 제4권6호
• /
• pp.12-19
• /
• 2003

This paper presents the effect of boundary conditions in various failure pressure models published for the estimation of failure pressure. Furthermore, this approach is extended to the failure prediction with the aid of a failure probability model. The first order Taylor series expansion of the limit state function is used in order to estimate the probability of failure associated with each corrosion defect in buried pipelines for long exposure period with unit of years. A failure probability model based on the von-Mises failure criterion is adapted. The log-normal and standard normal probability functions for varying random variables are adapted. The effects of random variables such as defect depth, pipe diameter, defect length, fluid pressure, corrosion rate, material yield stress, material ultimate tensile strength and pipe thickness on the failure probability of the buried pipelines are systematically investigated for the corrosion pipeline by using an adapted failure probability model and varying failure pressure model.

• 로봇학회논문지
• /
• 제6권3호
• /
• pp.247-254
• /
• 2011

This paper presents a method of sonar grid map matching for topological place recognition. The proposed method provides an effective rotation invariant grid map matching method. A template grid map is firstly extracted for reliable grid map matching by filtering noisy data in local grid map. Using the template grid map, the rotation invariant grid map matching is performed by Ring Projection Transformation. The rotation invariant grid map matching selects candidate locations which are regarded as representative point for each node. Then, the topological place recognition is achieved by calculating matching probability based on the candidate location. The matching probability is acquired by using both rotation invariant grid map matching and the matching of distance and angle vectors. The proposed method can provide a successful matching even under rotation changes between grid maps. Moreover, the matching probability gives a reliable result for topological place recognition. The performance of the proposed method is verified by experimental results in a real home environment.

• 한국물환경학회지
• /
• 제27권1호
• /
• pp.9-18
• /
• 2011

In this study the optimal volume for non-point source control retention is estimated considering spatio-temporal variation of land surface. The 3-parameter mixed exponential probability density function is used to represent the statistical properties of rainfall events, and NRCS-CN method is applied as rainfall-runoff transformation. The catchment drainage area is divided into individual $30m{\times}30m$ cells, and runoff curve number is estimated at each cell. Using the derived probability density function theory, the stormwater probability density function at each cell is derived from the rainfall probability density function and NRCS-CN rainfall-runoff transformation. Considering the antecedent soil moisture condition at each cell and the spatial variation of CN value at the whole catchment drainage area, the ensemble stormwater capture curve is established to estimate the optimal volume for an non-point source control retention. The comparison between spatio-temporally varied land surface and constant land surface is presented as a case study for a urban drainage area.