• 응용통계연구
• /
• 제27권3호
• /
• pp.431-444
• /
• 2014

본 논문에서는 시간-진동수 영역에서 시계열을 여러 구성 성분으로 분해하는 방법인 경험적모드분해법(Empirical Mode Decomposition)을 소개하고, 이를 이용하여 한국의 주요 거시 경제 지표를 대상으로 순환변동과 추세 성분을 추출하고 예측에 활용한다. 그 효율성을 살펴보기 위하여, 추출된 구성 성분들의 변동성, 동행성, 지속성, 인과성, 비정상성 및 예측력을 계산하고, 가장 보편적으로 널리 사용되고 있는 Hodrick-Prescott 필터에 의한 결과와 비교한다.

• ETRI Journal
• /
• 제43권1호
• /
• pp.120-132
• /
• 2021

A frequency diverse array (FDA) multiple-input multiple-output (MIMO) radar employs a small frequency increment across transmit elements to produce an angle-range-dependent beampattern for target angle and range detection. The joint angle and range estimation problem is a trilinear model. The traditional trilinear alternating least square (TALS) algorithm involves high computational load due to excessive iterations. We propose a fast-convergence trilinear decomposition (FC-TD) algorithm to jointly estimate FDA-MIMO radar target angle and range. We first use a propagator method to obtain coarse angle and range estimates in the data domain. Next, the coarse estimates are used as initialized parameters instead of the traditional TALS algorithm random initialization to reduce iterations and accelerate convergence. Finally, fine angle and range estimates are derived and automatically paired. Compared to the traditional TALS algorithm, the proposed FC-TD algorithm has lower computational complexity with no estimation performance degradation. Moreover, Cramer-Rao bounds are presented and simulation results are provided to validate the proposed FC-TD algorithm effectiveness.

• 전기전자학회논문지
• /
• 제18권2호
• /
• pp.240-247
• /
• 2014

안압의 측정은 각막에 air puff 후 각막의 변형 즉, 각막의 두께나 변위 등 생역학적 성질을 관찰함으로써 이루어진다. 본 논문에서는 녹내장이나 라식의 진단을 위하여 사용되는 안압계의 air puff에 의한 각막의 변형을 주파수 영역에서 분석하였다. Air puff 후 각막의 중심부 주변의 변위에 대한 진동 주파수를 측정함으로써 환자와 정상인을 구별한다. 동영상으로부터 이진 영상을 구하고, 상하 변위 데이터와 곡선 정합의 차를 구하여 시간에 따른 각막 상하 진동 프로파일 데이터를 추출하였다. Fourier 변환으로 비정상인의 진동 주파수는 479.2Hz를 얻었고 정상인의 경우 단단함으로 인하여 702.8Hz의 고주파 성분을 볼 수 있었다. 또한 Hilbert-Huang 변환의 EMD 방법을 이용, 고유 모드 함수로 분해하여 국소적, 비선형, 비정상성을 가지는 데이터를 얻고 주파수와 전력을 분석하였다. 마지막으로 특정 고유 모드 함수에 대한 환자와 정상인의 전력비가 6배 이상 차이가 나는 것을 확인하였다.

• Structural Engineering and Mechanics
• /
• 제22권1호
• /
• pp.53-70
• /
• 2006

The paper summarizes the dynamic-based assessment of a reinforced concrete arch bridge, dating back to the 50's. The outlined approach is based on ambient vibration testing, output-only modal identification and updating of the uncertain structural parameters of a finite element model. The Peak Picking and the Enhanced Frequency Domain Decomposition techniques were used to extract the modal parameters from ambient vibration data and a very good agreement in both identified frequencies and mode shapes has been found between the two techniques. In the theoretical study, vibration modes were determined using a 3D Finite Element model of the bridge and the information obtained from the field tests combined with a classic system identification technique provided a linear elastic updated model, accurately fitting the modal parameters of the bridge in its present condition. Hence, the use of output-only modal identification techniques and updating procedures provided a model that could be used to evaluate the overall safety of the tested bridge under the service loads.

• 지능정보연구
• /
• 제5권1호
• /
• pp.103-123
• /
• 1999

Input filtering as a preprocessing method is so much crucial to get good performance in time series forecasting. There are a few preprocessing methods (i.e. ARMA outputs as time domain filters, and Fourier transform or wavelet transform as time-frequency domain filters) for handling time series. Specially, the time-frequency domain filters describe the fractal structure of financial markets better than the time domain filters due to theoretically additional frequency information. Therefore, we, first of all, try to describe and analyze specially some issues on the effectiveness of different filtering methods from viewpoint of the performance of a neural network based forecasting. And then we discuss about neural network model architecture issues, for example, what type of neural network learning architecture is selected for our time series forecasting, and what input size should be applied to a model. In this study an input selection problem is limited to a size selection of the lagged input variables. To solve this problem, we simulate on analyzing and comparing a few neural networks having different model architecture and also use an embedding dimension measure as chaotic time series analysis or nonlinear dynamic analysis to reduce the dimensionality (i.e. the size of time delayed input variables) of the models. Throughout our study, experiments for integration methods of joint time-frequency analysis and neural network techniques are applied to a case study of daily Korean won / U. S dollar exchange returns and finally we suggest an integration framework for future research from our experimental results.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1998년도 하계종합학술대회논문집
• /
• pp.745-748
• /
• 1998

Recently, image and voice processing part is using wavelet transform. We propose thining algorithm using wavelet tranform. Wavelet transform consists of low frequency and high frequency in the spatial and frequency domain. After the wavelet decomposition, more than 90 percents of energy are contained in lowest frequency band. Therefor, for images with large difference of gray value between foreground and background like character images, we can more accurately in the lowest frequency band. Lowest frequency band has wavelet transform significant coefficient(WTS) that is required for the thinning algorithm we proposed Paper [3][5][7][8] can not separate consonants and vowels of korean characters. Becuase korean characters have structural feature. This paper can separate consonants and vowels. Simulation executed low frequency image and data compression can reduce 1/4$^{n}$ with level n. we can redcue time complexity 3/8.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2013년도 추계학술대회 논문집
• /
• pp.63-64
• /
• 2013

Induction motors' faults detection is almost a popular topic among researchers. Monitoring the output of motors is a key factor in detecting these faults. (Short-time) Fourier, (continuous, discrete) wavelet, and extended Park vector transformations are among the methods for fault detection. One major deficiency of these methods is not being able to detect the severity of faults that carry low energy information, e.g. in ball bearing system failure, there is absolutely no way to detect the severity of fault using Fourier or wavelet transformations. In this paper, the authors have applied the Discrete Wavelet Transform (DWT) frequency-domain analysis to detect bearing faults in an induction motor. In other words, in discrete transform which the output signal is decomposed in several steps and frequency resolution increases considerably, the frequency-band analysis is performed and it will be verified that first of all, fault sidebands become more recognizable for detection in higher levels of decomposition, and secondly, the inner race bearing faults turn out easier in these levels; and all these matter because of eliminating the not-required high energy components in lower levels of decomposing.

• Computers and Concrete
• /
• 제10권3호
• /
• pp.277-294
• /
• 2012

The aim of the study is to determine the modal parameters of a prototype damaged arch dam by operational modal analysis (OMA) method for some damage scenarios. For this purpose, a prototype arch dam-reservoir-foundation model is constructed under laboratory conditions. Ambient vibration tests on the arch dam model are performed to identify the modal parameters such as natural frequency, mode shape and damping ratio. The tests are conducted for four test-case scenarios: an undamaged dam with empty reservoir, two different damaged dams with empty reservoirs, and a damaged dam with full reservoir. Loading simulating random impact effects is applied on the dam to crack. Cracks and fractures occurred at the middle of the upper part of the dams and distributed through the abutments. Sensitivity accelerometers are placed on the dams' crests to collect signals for measurements. Operational modal analysis software processes the signals collected from the ambient vibration tests, and enhanced frequency domain decomposition and stochastic subspace identification techniques are used to estimate modal parameters of the dams. The modal parameters are obtained to establish a basis for comparison of the results of two techniques for each damage case. Results show that approximately 35-40% difference exists between the natural frequencies obtained from Case 1 and Case 4. The natural frequencies of the dam considerably decrease with increasing cracks. However, observation shows that the filled reservoir slightly affected modal parameters of the dam after severe cracking. The mode shapes obtained are symmetrical and anti-symmetrical. Apparently, mode shapes in Case 1 represent the probable responses of arch dams more accurately. Also, damping ratio show an increase when cracking increases.

• 대한의용생체공학회:의공학회지
• /
• 제25권6호
• /
• pp.589-598
• /
• 2004

산모의 흉부와 복부로부터 측정된 다채널 심전도에서 태아 심전도를 추출하는 새로운 알고리듬을 제안한다. 산모의 복부 심전도로부터 태아 심전도를 추출하기 위하여, 시간 영역에서 특이값 분해를 근간으로한 방법이 일반적으로 사용되었다. 그러나 이 방법은 산모와 태아의 심전도 벡터 방향이 서로 직교해야 하는 가정과 많은 연산량을 요구하는 단점이 있다. 제안한 알고리듬은 이산연현변환 영역에서 특이값 분해를 이용하여 이러한 단점을 극복한다. 적은 연산량으로 특이값 분해를 하기 위하여 이산여현변환 계수의 특성과 태아 심전도의 주파수 특성에 기초하여 고주파 수 성분에 해당하는 이산여현변환 계수를 제거하였다. 또한 산모와 태아의 심전도 벡터 방향에 의한 영향을 덜 받으면서 순수한 태아 심전도를 추출하기 위하여, 산모 복부 심전도에서 산모 심전도가 억압된 새로운 세 개의 채널을 만들고 이들을 다채널 심전도에 추가하였다. 모의 신호와 실제 신호를 이용하여 기존의 시간 영역에서 특이값 분해를 근간으로한 방법과 제안한 알고리듬의 성능을 비교하였다. 제안한 알고리듬은 기존 방법보다 적은 연산량으로 순수한 태아 심전도를 얻을 수 있음을 실험적으로 확인되었다.

• 한국컴퓨터정보학회논문지
• /
• 제26권7호
• /
• pp.127-132
• /
• 2021

본 논문은 건화물시장과 탱커시장의 운임지수 예측에 관하여 머신러닝을 적용하였으며 신호분해법인 웨이블릿 분해와 EMD분해를 데이터 전처리 과정에 반영하여 시간의 영역의 정보와 주파수 영역의 정보를 모두 반영할 수 있는 운임예측모형을 구축하였다. 건화물 시장의 경우 웨이블릿으로 분해한 예측모형이 우수하였으며 탱커시장의 EMD분해로 예측한 모형이 우수하였으며 실무적으로 각 운송시장 참여자들에게 새로운 단기예측 방법론을 제시하였다. 이러한 연구는 운송시장에서 양적으로 가장 중요한 건화물 시장과 탱커시장에 대한 다양한 예측방법론을 확대하고 새로운 방법론을 제시하였다는 측면에서 중요하며, 변동성이 큰 운임시장에서 과학적인 의사결정 방법에 대한 실무적인 요구를 반영할 수 있을 뿐만 아니라 가장 빈번한 스팟거래에 합리적인 의사결정이 이뤄질 수 있는 기초가 될 것으로 기대된다.