• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.151-153
• /
• 2006

we present a method of sequential speech enhancement, where we infer clean speech signal using a Rao-Blackwellized particle filter (RBPF), given a noise-contaminated observed signal. In contrast to Kalman filtering-based methods, we consider a non-Gaussian speech generative model that is based on the generalized auto-regressive (GAR) model. Model parameters are learned by a sequential Newton-Raphson expectation maximization (SNEM), incorporating the RBPF. Empirical comparison to Kalman filter, confirms the high performance of the proposed method.

• Journal of Information Processing Systems
• /
• v.16 no.1
• /
• pp.145-154
• /
• 2020

Maximum likelihood (ML) is the best estimator asymptotically as the number of training samples approaches infinity. This paper deduces an adaptive algorithm for blind signal processing problem based on gradient optimization criterion. A parametric density model is introduced through a parameterized generalized distribution family in ML framework. After specifying a limited number of parameters, the density of specific original signal can be approximated automatically by the constructed density function. Consequently, signal separation can be conducted without any prior information about the probability density of the desired original signal. Simulations on classical biomedical signals confirm the performance of the deduced technique.

• The Korean Journal of Applied Statistics
• /
• v.6 no.2
• /
• pp.277-287
• /
• 1993

This paper is a generalization of single- and two-level skip-lot sampling plans to n-level, which can considerably reduce inspection cost when the level of submitted quality is high. In every skipping inspection of the generalized sampling plan, not only skipping parameters but also inspection fractions can be freely choosed. The general formula of the operating characteristic function for the n-level skip-lot sampling plan is derived. Also the operating characteristic curves of a reference plan, two-level and three-level skip-lot sampling plans are compared.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.19 no.2
• /
• pp.200-207
• /
• 2009

Complex systems are made of many subsystems, those are developed and manufactured by many part companies. Even though the information for a part is necessary to analyze the performance of the other part, it is not so easy to get the information for that part from other companies due to many reasons like security or compatibilities. If the modal parameters of a system between the connecting points are available, we can reconstruct a reduced model for that system in a physical coordinate not in a generalized coordinate. The assemble of the equations of motion for the main system and the reduced equations of motion for the connected system can give a response of the main system considering the effects of connected systems. The results show that the proposed method can give the response of a system accurately. The rule for the selection of modes is to use the fundamental modes whose natural frequencies are low.

• Bulletin of the Korean Mathematical Society
• /
• v.30 no.1
• /
• pp.29-34
• /
• 1993

J. Leray [7] proposed a sufficient condition ofr the solvability of the Cauchy problem on the initial hyperplane x$_{1}$=0 with Cauchy data which are holomorphic with respect to the variables parallel to some analytic subvariety S of the initial hyperplane. Limiting the problem to the case of operators with constant coefficients, A. Kaneko [2] proposed a new sharper sufficient condition. Later we generalized this condition and showed that it is necessary and sufficient for the solvability of the Cauchy problem for the hyperfunction Cauchy data and the distribution Cauchy data which contain variables parallel to S as holomorphic parameters in [5, 6]. In this paper, we extend the results in [6] to the case of operators with variable coefficients and show that it is sufficient for the solvability of the Cauchy problem for the hyperfunction Cauchy data. Our main theorem can be considered as an example of a deep theorem on micro-hyperbolic systems by Kashiwara-Schapira [4] and we give a direct proof based on an elementary sweeping out procedure developed in Kaneko [3].

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06a
• /
• pp.830-835
• /
• 1994

In this paper, a new method of statistically evaluating an output response probability distribution of a memory type non-linear system is practically derived based on a zero-memory type non-linear equivalent system. That is, first, the objective system is approximately and functionally separated into two functional parts, i.e., a zero-memory type non-linear part and a memory type linear part according to the well-known Wiener's idea. A whole mathematical frame of the output probability distribution is evaluated in an approximate but generalized form, based on the equivalent zero-memory type non-linear part. The memory effects between the input and the output of the system are reflected in the statistical parameters and the expansion coefficients.

• Bulletin of the Korean Chemical Society
• /
• v.16 no.10
• /
• pp.957-968
• /
• 1995

Generalized Multichannel Quantum Defect theory (MQDT) was implemented to the vibrational predissociation of triatomic van der Waals molecules in the previous paper [Bull. Korean Chem. Soc, 12, 228 (1991)]. Implementation was limited to the calculation of the scattering matrix. It is now extended to the calculation of the predissociation spectra and the final rotational distribution of the photofragment. The comparison of the results with those obtained by other methods, such as Golden-rule type calculation, infinite order sudden approximation (IOS), and close-coupling method, shows that the implementation is successful despite the fact that transition dipole moments show more energy dependence than other quantum defect parameters. Examination of the short-range channel basis functions shows that they resemble angle-like functions and provide the validity of the IOS approximation. Besides the validity of the latter, only a few angles are found to play the major role in photodissociation. In addition to the implementation of MQDT, more progress in MQDT itself is made and reported here.

• Bulletin of the Korean Chemical Society
• /
• v.18 no.9
• /
• pp.965-972
• /
• 1997

An approximate molecular theory of classical fluids based on the nonrandom lattice statistical-mechanical theory is presented. To obtain configurational Helmholtz free energy and equation of state (EOS), the lattice-hole theory of the Guggenheim combinatorics is approximated by introducing the nonrandom two-fluid theory. The approximate nature in the derivation makes the model possible to unify the classical lattice-hole theory and to describe correctly the configurational properties of real fluids including macromolecules. The theory requires only two molecular parameters for a pure fluid. Results obtained to date have demonstrated that the model correlates quantitatively the first- and second-order thermodynamic properties of real fluids. The basic simplicity of the model can readily be generalized to multicomponent systems. The model is especially relevant to (multi) phase equilibria of systems containing molecularly complex species.

• Structural Engineering and Mechanics
• /
• v.19 no.2
• /
• pp.153-172
• /
• 2005

A new load/displacement parameter method is proposed for the simultaneous control of applied loads and structural displacements at one or more points. The procedure is based on a generalized Riks' method, which utilizes load/displacement parameters as scaling factors to analyze post-buckling phenomena including snap-through or snap-back. The convergence characteristics are improved by employing new relaxation factors through an incremental displacement parameter, particularly in a region that exhibits severe numerical instability. The improved performance is illustrated by means of a numerical example.

• Nuclear Engineering and Technology
• /
• v.7 no.2
• /
• pp.119-126
• /
• 1975

On the assumption that the spectral distribution of prompt fission neutrons released from supercritical accidents can be expressed by the generalized Cranberg form with two spectral parameters, which is then transformed into the single parameter form, a variation of the fission spectrum-averaged cross-sections for some threshold reactions with varying the spectral parameter has teen calculated using an electronic computer. It appears that the average cross-sections are very sensitive to the spectral deformation, especially those for the detectors having the threshold at high neutron energy are high compared to those for the detectors of which the threshold energies are comparatively low.