• MALSORI
• /
• no.52
• /
• pp.145-160
• /
• 2004

In this paper, a new noise compensation method based on the eigenvoice framework in feature space is proposed to reduce the mismatch between training and testing environments. The difference between clean and noisy environments is represented by the linear combination of K eigenvectors that represent the variation among environments. In the proposed method, the performance improvement of speech recognition systems is largely affected by how to construct the noisy models and the bias vector set. In this paper, two methods, the one based on MAP adaptation method and the other using stereo DB, are proposed to construct the noisy models. In experiments using Aurora 2 DB, we obtained 44.86% relative improvement with eigen-environment method in comparison with baseline system. Especially, in clean condition training mode, our proposed method yielded 66.74% relative improvement, which is better performance than several methods previously proposed in Aurora project.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.13 no.2
• /
• pp.153-158
• /
• 2020

In this paper, we studied an algorithm for estimating a desired target by removing noise and interference in a wireless communication environment. When an information signal with a mixture noise and interference receive a receiver, noise and interference signals must be removed to accurately estimate a desired target. In order to divide the received signal region into two spatial, a power spectrum is obtained by analyzing a correlation matrix, covariance, eigen vector, and eigen value. The proposed spectrum is an algorithm that can remove noise and interference, and analyzes the existing algorithm and target estimation performance through simulation. As a result of simulation, the target estimation resolution of existing algorithm is more than 10°, but the resolution of the proposed algorithm is less than 10°. The proposed algorithm has improved the resolution of about 5° than the exiting algorithm. The proposed algorithm proved that the target estimation accuracy and resolution are superior to the existing algorithm.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2016.02a
• /
• pp.167.1-167.1
• /
• 2016

The interaction between adsorbed water and hydrogen on metallic surfaces is important for fundamental understanding of heterogeneous catalysis and electrode surface reactions in acidic environment. Here, we explore a long-standing question of whether hydronium ion can exist or not on a Pt surface coadsorbed with atomic hydrogen and water. Studies based on mass spectrometry and infrared spectroscopy show clear evidence that hydrogen atoms are converted into hydrated protons on a Pt(111) surface. The preferential structures of hydrated protons are identified as multiply hydrated $H_5O_2{^+}$ and $H_7O_3{^+}$ species rather than as hydronium ions. The multiply hydrated protons may be regarded as two dimensional zundel ($H_5O_2{^+}$) and Eigen cation ($H_7O_3{^+}$) in water-metal interface. These surface-bound hydrated protons may be key surface intermediates of the electrochemical interconversion between adsorbed hydrogen atoms and solvated protons.

• Journal of Institute of Control, Robotics and Systems
• /
• v.1 no.2
• /
• pp.105-111
• /
• 1995

This paper proposes a segmentation algorithm using gray-level discontinuity and surface reflectance ratio of input images obtained under different illumination conditions. Each image is divided by a certain number of subregions based on the thresholds. The thresholds are determined using the histogram of fusion image which is obtained by ANDing the multiple input images. The subregions of images are projected on the eigenspace where their bases are the major eigenvectors of image matrix. Points in the eigenspace are classified into two clusters. Images associated with the bigger cluster are fused by revised ANDing to form a combined edge image. Missing edges are detected using surface reflectance ration and chain code. The proposed algorithm obtains more accurate edge information and allows to more efficiently recognize the environment under various illumination conditions.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.12
• /
• pp.1522-1531
• /
• 1988

The purpose of this paper is to explain eigenstructure based spatial spectrum estimation algorithms computing better estimates than the other approaches. Also, as an approach to overcome performance degradations in a multipath environments, the notion of forward and backwark spatial smoothing is discussed. Intensive simulation results,which include the comparisons of the eigenbased spatial spectral estimation algorithms in the situations of faulty estimation of the number of signals, are presented. The simulation results have shown that overestimation of the number of signals is more desirable than underestimation in using EV (Eigen Vector) and MUSIC (Multiple Signal Classification) algorithms and that underestimation of the number of signals is better strategy than overestimation in using eigenstructure based LP(Linear Prediction) algorithms.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.34 no.11
• /
• pp.1613-1622
• /
• 2010

In this paper, the ultraprecision positioning control of an ultrasonic linear motor in a high-vacuum environment is presented. The bolt-clamped Langivin type transducer (BLT) with the 3rd longitudinal; and 6th lateral vibration modes was developed, which was excited by using the Eigen resonance frequency for two vibration modes in order to generate stable and high power. In practical applications, however, even if a geometrical design has an Eigen frequency, discordance between both mode frequencies can be generated by the contact mechanism and because of manufacturing errors as well as environmental factors. Both mode frequencies were precisely matched by adjusting the impedence. By using this method, the BLT can be driven under any environmental conditions. The nominal characteristic trajectory following(NCTF) control method was adopted to control the positioning of the system in vacuum. The developed linear motor stage show high positioning accuracy with 5 nm.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.21 no.1
• /
• pp.20-27
• /
• 2010

In this paper, we presented a signal model of STAP and actual environment of clutter. The novel estimation method of clutter covariance matrix using median value is proposed to overcome serious performance degradation after NHD in nonhomogeneous clutter. Eigen value characteristic is improved through diagonal loading. Target detection ability and SINR loss of the proposed method though MSMI statistic is also compared with conventional method using average value. The simulation results, confirm the proposed method has better performance than others.

• Proceedings of the Korea Electromagnetic Engineering Society Conference
• /
• 2003.11a
• /
• pp.43-46
• /
• 2003

This paper presents capacity characteristics of the indoor LOS(Line-Of-Sight) propagation channel for MIMO system at 5GHz. The distance between antenna elements, their moving path, and number of transmitting and receiving antennas can be determined by wanted eigen-vlaue, and channel capacity of the MIMO communication channel using only reliable simulation without measurements. The simulation uses 3D Ray tracing and patch scattering model to which electromagnetic material constants are applied. As distance between antenna elements increases, distribution of the eigen-value show a tendency to decrease, but channel capacity increases in LOS environment. However, despite of short distance between antenna elements, large value of channel capacity is obtained in positions which have high AS. When the position of receiver antennas are shifted, channel capacity hardly changed, and as number of antenna elements increases, channel capacity also increases regularly.

• Proceedings of the IEEK Conference
• /
• 2003.07c
• /
• pp.2737-2740
• /
• 2003

This paper has been studied an algorithm for EMG signal amplitude estimation in noisy environment. The proposed method has the first stage decomposing the row vector from the delayed EMG signal and the second stage computing the eigenvalues by the eigen decomposition from the covariance matrix of the EMG signal matrix. The last stage is the estimation of RMS values from the eigenvalues. The proposed method was effective when the amplitude of the EMG signal is small, which means the signal to noise ratio is low.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.8C
• /
• pp.769-774
• /
• 2005

If the angle of arrival (AOA) of a source is perturbed due to some reasons in a statistical way as in the environment of wireless mobile communications, a new model appropriate for such environment should be used instead of the point source model. In this paper, an angel-perturbed source model is proposed and an estimation method based on the eigen-decomposition tecklique is investigated under the model. The asymptotic distribution of the estimation errors is obtained to observe the statistical properties.