• The Journal of the Acoustical Society of Korea
• /
• v.25 no.7
• /
• pp.319-324
• /
• 2006

We Propose two new methods for non-keyword modeling to improve the performance of speaker- and vocabulary-independent keyword spotting system. The first method is decision tree clustering of monophone at the state level instead of monophone clustering method based on K-means algorithm. The second method is multi-state multiple mixture modeling at the syllable level rather than single state multiple mixture model for the non-keyword. To evaluate our method, we used the ETRI speech DB for training and keyword spotting test (closed test) . We also conduct an open test to spot 100 keywords with 400 sentences uttered by 4 speakers in an of fce environment. The experimental results showed that the decision tree-based state clustering method improve 28%/29% (closed/open test) than the monophone clustering method based K-means algorithm in keyword spotting. And multi-state non-keyword modeling at the syllable level improve 22%/2% (closed/open test) than single state model for the non-keyword. These results show that two proposed methods achieve the improvement of keyword spotting performance.

• Advances in nano research
• /
• v.2 no.1
• /
• pp.69-76
• /
• 2014

We have obtained opal-like photonic crystals based on opals and inverted opals exhibiting a shift of the selective reflection band toward longer and shorter wavelengths with respect to the diffraction band of the initial opal consisting of $SiO_2$ spheres. The contribution of frames forming three-dimensional periodic structures and that of fillers to the spectral arrangement of the diffraction bands has been determined.

• Journal of Sasang Constitutional Medicine
• /
• v.20 no.2
• /
• pp.11-20
• /
• 2008

1. Objectives The objective of this paper is to understand on treatment of in critical state patients based on the text book of Sasang Constitutional Medicine. We expect this study can be helpful in treating the critical state patients. 2. Methods The study is mainly based on the bibliographical research on the "Dongyisoosebowon" and on the "Principle of dosing for patient in critical state". 3. Results and conclusions Lee je-ma suggested a health scale which people can measure his health grade from a standard healthy state to severe diseases. We investigated the treatment of 'critical state' patients according to his scale. Lee je-ma also confirmed the standard dose of herbs and how to take herbs in compliance with the state which each patient in. We tried to find a rules of treating the critical state patients based on the text "Dongyisoosebowon" and we suggested brief principles in treating patients.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.27 no.4
• /
• pp.332-338
• /
• 2022

This study proposes a battery state-of-health estimation method by applying a feature extraction technique. The technique that can improve estimation performance is the process of identifying and extracting meaningful data. To apply a data-driven-based aging state estimation method to batteries, health indicators are used as training data. However, limitations occur in extracting health indicators from charge/discharge cycles. This study proposes a deep-learning-based battery state-of-health estimation method that applies feature extraction techniques to compensate for this problem. According to the performance evaluation result of the proposed method, it has a low estimation error of 0.3887% based on an absolute error evaluation method.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.11a
• /
• pp.465-470
• /
• 2003

In this study, the state-space Newmark method based on average velocity is presented to analyse the transient dynamic response for general dynamic system. The conventional Newmark method based on average acceleration cannot he directly to the first-order state-space differential equations introducing the state-space vector. To overcome this problem, the time-step integration algorithm, based on average velocity concept, suitable for the first-order state-space differential equations is proposed In results, the proposed method has %he numerical stability and order of accuracy, which is proved analytically, equal to those of the conventional Newmark method based on average acceleration. Also, the formulation for numerical solution is very simple and the calculation time Is nearly equal to that of the conventional Newmark method based on average acceleration in spite of an increase of two times over matrix size. This method will be look forward to applying the general dynamic system to calculate the transient dynamic response.

• Proceedings of the KSME Conference
• /
• 2007.05a
• /
• pp.1103-1108
• /
• 2007

This paper describes a robust model-based roll state estimator for application to the detection of impending vehicle rollover. The roll state estimator is based on a 2-D bicycle model and a roll model to estimate the maneuver-induced vehicle roll motion. The measurement signals are lateral acceleration, yaw rate, steering angle, and vehicle speed. Vehicle mass is adapted to obtain robust performance of the estimator. Computer simulation is conducted to evaluate the proposed roll state estimator by using a validated vehicle simulator. It is shown that the roll state estimator shows robust performance without exact vehicle mass information.

• MALSORI
• /
• no.45
• /
• pp.117-126
• /
• 2003

Decision tree-based state tying has been proposed in recent years as the most popular approach for clustering the states of context-dependent hidden Markov model-based speech recognition. The aims of state tying is to reduce the number of free parameters and predict state probability distributions of unseen models. But, when doing state tying, the size of a decision tree is very important for word independent recognition. In this paper, we try to construct optimized decision tree based on the average of feature vectors in state pool and the number of seen modes. We observed that the proposed optimal decision tree is effective in predicting the state probability distribution of unseen models.

• Journal of Electrical Engineering and Technology
• /
• v.7 no.6
• /
• pp.891-897
• /
• 2012

In this paper, the observer based nonlinear state feedback controller has been developed to control the pressures of the oxygen and the hydrogen in the PEM(Proton Exchange Membrane) fuel cell system. Nonlinear model of the PEM fuel cell system was introduced to study the design problems of the state observer and model based controller. A cascade observer using the filtering technique was used to estimate the pressure derivatives of the cathode and the anode in the system. In order to estimate the pressures of the cathode and the anode, the sliding mode observer was designed by using these pressure derivatives. To estimate the oxygen pressure and the hydrogen pressure in the system, the nonlinear state observer was designed by using the cathode pressure estimates and the anode it. These results will be very useful to design the state feedback controller. The validity of the proposed observers and the controller has been investigated by using the Lyapunov's stability analysis strategy.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.20 no.5
• /
• pp.688-699
• /
• 2017

Model-based test, a well-known method of the black box tests, is consisted of the following four steps : model construction using requirement, test case generation from the model, execution of a SUT (software under test) and detection failures. Among models constructed in the first step, state-based models such as UML standard State Machine are commonly used to design event-based embedded systems (e.g., weapon control systems). To generate test cases from state-based models in the next step, coverage-based techniques such as state coverage and transition coverage are used. Round-trip path coverage technique using W-Method, one of coverage-based techniques, is known as more effective method than others. However it has a limitation of low failure observability because the W-Method technique terminates a testing process when arrivals meet states already visited and it is hard to decide the current state is completely same or not with the previous in the case like the GUI environment. In other words, there can exist unrevealed faults. Therefore, this study suggests a Extended W-Method. The Extended W-Method extends the round-trip path to a final state to improve failure observability. In this paper, we compare effectiveness and efficiency with requirement-item-based technique, W-Method and our Extended W-Method. The result shows that our technique can detect five and two more faults respectively and has the performance of 28 % and 42 % higher failure detection probability than the requirement-item-based and W-Method techniques, respectively.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.3
• /
• pp.625-631
• /
• 2011

This paper presents a discrete-time output feedback consensus algorithm for Multi-Agent Systems (MAS). Under the assumption that an agent is aware of the relative state information about its neighbors, a state feedback consensus algorithm is designed based on Linear Matrix Inequality (LMI) method. In general, however, it is possible to obtain its relative output information rather than the relative state information. To reconcile this problem, an Unknown Input Observer (UIO) is employed in this paper. To this end, first it is shown that the relative state information can be estimated using the UIO and the measured relative output information. Then a certainty-equivalence type output feedback consensus algorithm is proposed by combining the LMI-based state feedback consensus algorithm with the UIO. Finally, simulation results are given to illustrate that the proposed method successfully achieves the state consensus.