• Journal of the Korea Society for Simulation
• /
• v.5 no.1
• /
• pp.1-12
• /
• 1996

We propose a new approach to classifying a time series data into one of the autoregressive moving-average (ARMA) models. It is bases on two pattern recognition concepts for solving time series identification. The one is an extended sample autocorrelation function (ESACF). The other is a neural network-driven decision tree classifier(NNDTC) in which two pattern recognition techniques are tightly coupled : neural network and decision tree classfier. NNDTc consists of a set of nodes at which neural network-driven decision making is made whether the connecting subtrees should be pruned or not. Therefore, time series identification problem can be stated as solving a set of local decisions at nodes. The decision values of the nodes are provided by neural network functions attached to the corresponding nodes. Experimental results with a set of test data and real time series data show that the proposed approach can efficiently identify the time seires patterns with high precision compared to the previous approaches.

• Interaction and multiscale mechanics
• /
• v.5 no.4
• /
• pp.347-368
• /
• 2012

Measuring the bridge frequencies indirectly from an instrumented test vehicle is a potentially powerful technique for its mobility and economy, compared with the conventional direct technique that requires vibration sensors to be installed on the bridge. However, road surface roughness may pollute the vehicle spectrum and render the bridge frequencies unidentifiable. The objective of this paper is to study such an effect. First, a numerical simulation is conducted using the vehicle-bridge interaction element to demonstrate how the surface roughness affects the vehicle response. Then, an approximate theory in closed form is presented, for physically interpreting the role and range of influence of surface roughness on the identification of bridge frequencies. The latter is then expanded to include the action of an accompanying vehicle. Finally, measures are proposed for reducing the roughness effect, while enhancing the identifiability of bridge frequencies from the passing vehicle response.

• The Journal of the Acoustical Society of Korea
• /
• v.11 no.1
• /
• pp.35-41
• /
• 1992

A speaker identification scheme using the speaker-based VQ codecook of a sustained vowel is proposed and tested. With the pitch synchronous LPC vector of the sustained vowel /i/ as a feature vector, a VQ codebook size of 4 was found to be suitable to characterize each speaker's feature space. For 40 normal speakers (20 males, 20 females), we achieved the correct identification rate of 99.4% with a training data set, and 89.4% with a test data set with speech samples of only 50 pitch periods.

• Smart Structures and Systems
• /
• v.17 no.3
• /
• pp.431-444
• /
• 2016

This paper was written in the context of a benchmark study promoted by The Hong Kong Polytechnic University using data samples collected in an instrumented cable-stayed bridge. The main goal of the benchmark test was to study the identification of the bridge modes of vibration under different wind conditions. In this contribution, the tools developed at ViBest/FEUP for automated data processing of setups collected by dynamic monitoring systems are presented and applied to the data made available in the context of the benchmark study. The applied tools are based on parametric output only modal identification methods combined with clustering algorithms. The obtained results demonstrate that the proposed algorithms succeeded to automatically identify the modes with relevant contribution for the bridge response under different wind conditions.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.1002-1005
• /
• 1991

It is important to express the specific growth rate of a fermentation reaction as a function of substrate and product concentration in developing bioprocess automation techniques such as modeling of the reactor and controlling it via an advanced control scheme. Typical methods of identification utilize graphical representation of the rate constant data or nonlinear regression with an appropriate noise filter. But the former method fails when the data are erroneous and the latter are mathematically complicated to apply in the field. Neural network is another candidate for the identification from time series data since it is insensitive to the random data error and easy to implement. In this study, we will develop a neural network method of specific growth rate estimation from the time series state variable data and test the performance.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2001.05a
• /
• pp.786-792
• /
• 2001

In this paper, the identification of connections for a vibration system has been presented using FRF-based substructural sensitivity analysis. The substructural design sensitivity formula is derived and plugged into a commercial optimization program, MATLAB, to identify connection stiffness of an air-conditioner system of passenger car. The air-conditioner system, composed of a compressor and a bracket is analyzed by using FRF-based substructural(FBS) method. To obtain the FRFs, FE model is built for the bracket, and the impact hammer test is performed for the compressor. Obtained FRFs are combined to calculate the reaction force at the connection point and the system response. Connection element properties are determined by minimizing the difference between a target FRF and calculated one. It is shown that the proposed identification method is effective even for a real problem.

• Korean Journal of Pharmacognosy
• /
• v.30 no.1
• /
• pp.54-58
• /
• 1999

A method using coloric and spectrophotometric detection have been developed for the identification of the tertiary or quaternary alkaloids contained in Corydalis tuber and its preparations. The principle is based on the formation or decomposition of complex compounds. The complex compound of the tertiary and quaternary alkaloids have been formed by adding tetrathiocyanatocobaltate [II] ion to the test soln. Diverse crude drugs were screened using this method and the results indicated that isoquinoline, aconitine-type alkaloids in crude drugs can be readily detected. The method is simple, convenient, reproducible and applicable to the verification of the crude drug Corydalis tuber and its preparations.

• Journal of the Korean Society for Precision Engineering
• /
• v.15 no.3
• /
• pp.133-140
• /
• 1998

The neural network was applied to identify wear debris generated from the lubricated machine surface. The wear test was carried out under different experimental conditions. In order to describe characteristics of debris of various shapes and sizes, the four shape parameter(50％ volumetric diameter, aspect, roundness and reflectivity) of wear debris are used as inputs to the network and learned the friction condition of five values(material 3, applied load 1, sliding distance 1). It is shown that identification results depend on the ranges of these shape parameter learned. The three kinds of the wear debris had a different pattern characteristics and recognized the friction condition and materials very well by neural network.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 1997.04a
• /
• pp.102-109
• /
• 1997

A study to determine a proper sensor placement was developed to improve force identification. Improper selection of response position cause erroneous result in force identification problem. This paper presents two methods to improve the conditioning of the system's FRM(Frequency Response Matrix) which affects the accuracy of result. The basic strategy of the two methods in selecting the response position is to let the smallest singular value be as large as possible by maximizing the orthogonality of FRM. The suggested methods are tested numerically with a fixed-fixed beam model. The test results show that the proposed methods are very effective in dealing with the force identification problem.

• Phonetics and Speech Sciences
• /
• v.4 no.2
• /
• pp.79-86
• /
• 2012

In this paper, we propose a novel approach to improve histogram equalization for speaker identification. Our method collects all speech features of UBM training data to make a reference distribution. The ranks of the feature vectors are calculated in the sorted list of the collection of the UBM training data and the test data. We use the ranks to perform order-based histogram equalization. The proposed method improves the accuracy of the speaker recognition system with short utterances. We use four kinds of speech databases to evaluate the proposed speaker recognition system and compare the system with cepstral mean normalization (CMN), mean and variance normalization (MVN), and histogram equalization (HEQ). Our system reduced the relative error rate by 33.3% from the baseline system.