• Transactions on Control, Automation and Systems Engineering
• /
• 제2권1호
• /
• pp.40-46
• /
• 2000

In this paper, a transient identification based on a Hidden Markov Model (HMM) has been suggested and evaluated experimentally for the classification of transients in the dynamic process. The transient can be identified by its unique time dependent patterns related to the principal variables. The HMM, a double stochastic process, can be applied to transient identification which is a spatial and temporal classification problem under a statistical pattern recognition framework. The HMM is created for each transient from a set of training data by the maximum-likelihood estimation method. The transient identification is determined by calculating which model has the highest probability for the given test data. Several experimental tests have been performed with normalization methods, clustering algorithms, and a number of states in HMM. Several experimental tests have been performed including superimposing random noise, adding systematic error, and untrained transients. The proposed real-time transient identification system has many advantages, however, there are still a lot of problems that should be solved to apply to a real dynamic process. Further efforts are being made to improve the system performance and robustness to demonstrate reliability and accuracy to the required level.

• 한국지진공학회:학술대회논문집
• /
• 한국지진공학회 2002년도 춘계 학술발표회 논문집
• /
• pp.51-58
• /
• 2002

The relative site amplification factors in southern Korea were determined from coda waves using coda normalization method. The seismograms of 15 events at 79 stations were used in this study. Seismogram envelopes were obtained by the Hilbert transform of bandpass-filtered velocity seismograms with frequency bands at 1-2Hz, 2-4Hz, 4-8Hz, 8-l6Hz and 16-32Hz. The envelopes were stabilized by application of moving-average scheme with time window of 1 second. The relative amplitudes of seismogram envelope were computed by dividing the amplitude of seismogram envelope at one site by the amplitude of seismogram envelope at reference site. The relative site amplification factors were obtained by taking averages of the relative amplitude. Values of relative site amplification factors in southern Korea are generally low in western area and high in eastern area.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1991년도 하계학술대회 논문집
• /
• pp.900-904
• /
• 1991

In this paper, our study for the recognition of Hand-written Korean characters, Arabic numbers and alphabets by neural netwoks. This System extracts feature of character by using the MESH feature point of handwritten character, Arabic numbers and alphabets. To reduce the input image data, features are extracted from each input images. A MLP(multi-layer perceptron) with one hidden layer was trained with a modified BEP(back error propagation) algorithm. This method extracts feature sets of the characters directly from the scanner and can enhance computation speed without using the special preprocesses such as size normalization, smoothing, and thinning.

• 한국군사과학기술학회지
• /
• 제10권4호
• /
• pp.185-194
• /
• 2007

This paper presents a motor selection technique for a manipulator design that is used in a multipurpose dog-horse robot. Since the dynamics of a manipulator and its servo drives are closely related to each other, it requires a repetitive analysis to determine a suitable motor. In order to simplify this procedure, Straete et al. proposed a simple normalization method to separate the load dependent terms and the motor dependent terms. This technique is adopted in this paper for selecting a motor in designing a manipulator.

• 한국경영과학회:학술대회논문집
• /
• 대한산업공학회/한국경영과학회 1996년도 춘계공동학술대회논문집; 공군사관학교, 청주; 26-27 Apr. 1996
• /
• pp.172-178
• /
• 1996

Databases are critical to business information systems, and RDBMS is most widely usded for the database system. Normalization has been designed to control various anomalies(insert, update, and delete anomalies). However, normalized databese design does not account for the tradeoffs necessary for the performance. In this research, we develop a model for database desin by denormalization of duplicating attributes in order to reduce frequent join processes. In this model, we consider insert, update, and delete costs. The anomalies are treated by additional disk I/O which is necessary for each insert and update transaction. We propose a branch and bound method for this model, and show considerable cost reduction.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1997년도 춘계학술대회
• /
• pp.453-456
• /
• 1997

In this paper, we present the developement of Biofeedback system. Biofeedback is a method for learned control of physiological response of the body. It can be used to control certain biological response that cause health problems. Over the past 2 decades, biofeedback has received much attention from medicine as a cure-all for a variety of disease. It has been heralded as a technique that can provide normalization of disturbed physiology without the need for drugs or surgery and their attendant negative side effects. Biofeedback system has been implemented by using PIC microprocessor. We have displayed 5-channel value on PC and managed patient information by using Delphi.

• 제어로봇시스템학회논문지
• /
• 제16권12호
• /
• pp.1137-1142
• /
• 2010

This paper studied the speech parameters less affected by the human emotion for the development of the robust speech recognition system. For this purpose, the effect of emotion on the speech recognition system and robust speech parameters of speech recognition system were studied using speech database containing various emotions. In this study, mel-cepstral coefficient, delta-cepstral coefficient, RASTA mel-cepstral coefficient and frequency warped mel-cepstral coefficient were used as feature parameters. And CMS (Cepstral Mean Subtraction) method were used as a signal bias removal technique. Experimental results showed that the HMM based speaker independent word recognizer using vocal tract length normalized mel-cepstral coefficient, its derivatives and CMS as a signal bias removal showed the best performance of 0.78% word error rate. This corresponds to about a 50% word error reduction as compare to the performance of baseline system using mel-cepstral coefficient, its derivatives and CMS.

• International Journal of Advanced Culture Technology
• /
• 제5권2호
• /
• pp.63-73
• /
• 2017

In this paper, we propose algorithm for determining the static and dynamic posture using the acceleration sensor of smartphone. The measured acceleration values are then analyzed according to a preprocessing to the respective axis (X, Y, Z) and posture (standing, sitting, lying) presents static posture determination criterion. The proposed static posture determination condition is used for static posture determination and dynamic posture determination. The dynamic posture is determined by using regression linear equations. In addition, transition state can be grasped by SVM change in dynamic posture determination. Experimental results are presented using data and app. Experiments were performed using data collected from 10 adults.

• 한국콘크리트학회:학술대회논문집
• /
• 한국콘크리트학회 2006년도 춘계학술발표회 논문집(I)
• /
• pp.494-497
• /
• 2006

The objectives of this study are to develop and evaluate the Neural Network algorithm which can predict the inelastic shortening such as the creep strain and the drying shrinkage strain of reinforced concrete members using the previous test data. New learning algorithms for the prediction of creep strain and the drying shrinkage strain are proposed focusing on input layer components and a normalization method for input data and their validity is examined through several test data. In Neural Network algorithm, the main input data to be trained are the compressive strength of the concrete, volume to surface ratio, curing condition, relative humidity, and the applied load. The results show that the new algorithms proposed herein successfully predict creep strain and the drying shrinkage strain.

• Journal of Electrical Engineering and information Science
• /
• 제3권1호
• /
• pp.80-85
• /
• 1998

A theoretical calculation and an optical measurement method of normal-incidence absorption are proposed. By using a waveguide structure, optical interference and the problem of low level signal can be avoided in the measurement of normal-incidence absorption. The oscillator strength of intersubband absorption for a waveguide structure is calculated in Si(001), Si(110), and Ge(001) quantum wells. The polarization angle dependence of the measured and the calculated absorption strength can be obtained with the same waveguide structure, and be compared after normalization. The normal-incidence absorption in Si(110) and Ge(001) quantum wells is shown theoretically, and can be observed in the optical measurement using waveguide structures at the polarization angle of 90$^{\circ}$.