• Phonetics and Speech Sciences
• /
• 제1권1호
• /
• pp.69-73
• /
• 2009

This paper proposes an efficient global covariance-based principal component analysis (GCPCA) for speaker identification. Principal component analysis (PCA) is a feature extraction method which reduces the dimension of the feature vectors and the correlation among the feature vectors by projecting the original feature space into a small subspace through a transformation. However, it requires a larger amount of training data when performing PCA to find the eigenvalue and eigenvector matrix using the full covariance matrix by each speaker. The proposed method first calculates the global covariance matrix using training data of all speakers. It then finds the eigenvalue matrix and the corresponding eigenvector matrix from the global covariance matrix. Compared to conventional PCA and Gaussian mixture model (GMM) methods, the proposed method shows better performance while requiring less storage space and complexity in speaker identification.

• Journal of the Korean Operations Research and Management Science Society
• /
• 제39권4호
• /
• pp.137-152
• /
• 2014

In this paper, I suggest several techniques to estimate covariance matrix and compare the performance of the global minimum variance portfolio (GMVP) in terms of out of sample mean standard deviation and return. As a result, the return differences among the GMVPs are insignificant. The mean standard deviation of the GMVP using historical covariance is sensitive to the estimation window and the number of assets in the portfolio. Among the model covariance, the GMVP using constant systematic risk ratio model or using short sale restriction shows the best performance. The performance difference between the GMVPs using historical covariance and model covariance becomes insignificant as the historical covariance is estimated with longer estimation window. Lastly, the implied volatilities from ELW prices do not lead to superior performance to the historical variance.

• International Journal of Aeronautical and Space Sciences
• /
• 제3권1호
• /
• pp.19-29
• /
• 2002

Decentralized filtering for a formation flight instrumentation system by INS/GPS integration is considered in this paper. An elaborate tuning method of the measurement noise covariance is suggested to compensate modeling errors caused by decentralizing the extended Kalman filter. It does not require large data transfer between formation vehicles. Covariance analysis exhibits the superior performance of the proposed approach when compared with the existent decentralized filter and the global filter, which has the target-filter performance.

• Proceedings of the Korean Environmental Sciences Society Conference
• /
• 한국환경과학회 2003년도 International Symposium on Clean Environment
• /
• pp.47-50
• /
• 2003

$CO_2$ concentration profile was measured to investigate whether $CO_2$ concentration at one level (i.e., eddy covariance measurement level) can be used to estimate storage term without significant uncertainty at broadleaf deciduous forest at Kwangneung experiment forest in Korea. Based on t-test with significance level of 5％, there was no statistical difference between storage term from one-level $CO_2$ concentration and one from $CO_2$ profile measurement. Storage term constitutes on average 5％ of half hourly net ecosystem exchange (NEE) even at unstable stability (i.e., well mixed condition), indicating that storage term should be considered even at daytime, which is sometimes neglected.

• Korean Journal of Agricultural and Forest Meteorology
• /
• 제1권1호
• /
• pp.60-71
• /
• 1999

The micrometeorological tower flux network is the cornerstone of the global terrestrial vegetation monitoring. The eddy covariance technique used for tower fluxes is derived from the conservation of mass and is most applicable for steady-state conditions over flat, extended, and uniform vegetation. This technique allows us to obtain surface fluxes of energy budget components, greenhouse and trace gases, and other pollutants. The quality-controlled flux data are invaluable to validate various models with temporal scales ranging from minutes to years and spatial scales ranging from a few meters to hundreds of kilometers. In this paper, we review the theoretical background of this important eddy covariance technique, examine the measurement criteria and corrections, and finally suggest some measurement strategies that may facilitate coordinated flux measurements among different disciplines and provide a strong infrastructure for the global flux network.

• Journal of information and communication convergence engineering
• /
• 제9권2호
• /
• pp.207-211
• /
• 2011

This paper presents a method for facial feature representation and recognition from the Covariance Matrices of the Gabor-filtered images. Gabor filters are a very powerful tool for processing images that respond to different local orientations and wave numbers around points of interest, especially on the local features on the face. This is a very unique attribute needed to extract special features around the facial components like eyebrows, eyes, mouth and nose. The Covariance matrices computed on Gabor filtered faces are adopted as the feature representation for face recognition. Geodesic distance measure is used as a matching measure and is preferred for its global consistency over other methods. Geodesic measure takes into consideration the position of the data points in addition to the geometric structure of given face images. The proposed method is invariant and robust under rotation, pose, or boundary distortion. Tests run on random images and also on publicly available JAFFE and FRAV3D face recognition databases provide impressively high percentage of recognition.

• Korean Journal of Agricultural and Forest Meteorology
• /
• 제9권4호
• /
• pp.260-267
• /
• 2007

This research note introduces the procedure of the quality control and quality assurance applied to the eddy covariance data collected at the two KoFlux sites (i.e., Gwangneung forest and Haenam farmland). The quality control was conducted through several steps based on micrometeorological theories and statistical tests. The data quality was determined at each step of the quality control procedure and was denoted by five different quality flags. The programs, which were used to perform the quality control, and the quality assessed data are available at KoFlux website (http://www.koflux.org/).

• Ocean and Polar Research
• /
• 제26권3호
• /
• pp.453-463
• /
• 2004

The Antarctic Peninsula is important in terms of global warming research due to pronounced increase of air temperature over the last century. The first eddy covariance system was established at King Sejong Station located in the northern region of the Antarctic Peninsula in December of 2002 and has been operated over one year. Here, we analyze turbulent characteristics to determine quality control criteria for turbulent sensible heat flux data as well as to diagnose the possibility of long term eddy covariance measurement under extreme weather conditions of the Antarctic Peninsula. We also report the preliminary result on sensible heat flux. Based on the analyses on turbulent characteristics such as integral turbulence characteristics of vertical velocity (w) and heat (T), stationarity test and investigation of correlation coefficient, they fallow the Monin-Obukhov similarity and eddy covariance flux data were reliable. ${\sim}47%$ of total retrieved sensible heat flux data could be used for further analysis. Daytime averaged sensible heat flux showed a pronounced seasonal variation, with a maximum of up to $300Wm^{-2}$ in summer. In conclusion, continuous and long-term eddy covariance measurement may be possible at the study site and the land surface may influence the atmosphere significantly through heat transport in summer.

• Journal of Advanced Navigation Technology
• /
• 제21권1호
• /
• pp.138-144
• /
• 2017

A method for improving the integrated navigation performance in the INS/GPS navigation system by the considering that the condition that the geometric arrangement of the satellite is degraded due to limitation of the line of sight of the satellite such as geographic feature and GPS signal jamming is proposed. A variable covariance extended Kalman filter (VCEKF) that correlated to the measured covariance to the DOP of GPS is suggested. The navigation performance of integrated navigation system using EKF and VCEKF is analyzed by Monte-Carlo simulations. The result is verified that VCEKF has better estimation performance than EKF using fixed covariance on condition that DOP value is larger than the smaller value.

• International Journal of Aeronautical and Space Sciences
• /
• 제11권4호
• /
• pp.326-337
• /
• 2010

This work addresses problems regarding trajectory planning for unmanned aerial vehicle sensors. Such sensors are used for taking measurements of large nonlinear systems. The sensor investigations presented here entails methods for improving estimations and predictions of large nonlinear systems. Thoroughly understanding the global system state typically requires probabilistic state estimation. Thus, in order to meet this requirement, the goal is to find trajectories such that the measurements along each trajectory minimize the expected error of the predicted state of the system. The considerable nonlinearity of the dynamics governing these systems necessitates the use of computationally costly Monte-Carlo estimation techniques, which are needed to update the state distribution over time. This computational burden renders planning to be infeasible since the search process must calculate the covariance of the posterior state estimate for each candidate path. To resolve this challenge, this work proposes to replace the computationally intensive numerical prediction process with an approximate covariance dynamics model learned using a nonlinear time-series regression. The use of autoregressive time-series featuring a regularized least squares algorithm facilitates the learning of accurate and efficient parametric models. The learned covariance dynamics are demonstrated to outperform other approximation strategies, such as linearization and partial ensemble propagation, when used for trajectory optimization, in terms of accuracy and speed, with examples of simplified weather forecasting.