• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1990년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 26-27 Oct. 1990
• /
• pp.1145-1149
• /
• 1990

This paper proposes a numerical method for redundant manipulators using predetermined optimal resolution. In order to obtain optimal joint trajectories, it is desirable to formulate redundancy resolution as an optimization problem having an integral cost criterion. We predetermine the trajectories of redundant joints in terms of the Nth partial sum of the Fourier series, which lead to the solution in the desirable homotopy class. Then optimal coefficients of the Fourier series, which yield the optimal solution within the predetermined class, are searched by the Powell's method. The proposed method is applied to a 3-link planar manipulator for cyclic tasks in Cartesian space. As the results, we can obtain the optimal solution in the desirable homotopy class without topological liftings of the solution. To show the validity of the proposed method, we analyze both optimal and extremal solutions by the Fast Fourier Transform (FFT) and discuss joint trajectories on the phase plane.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2008년도 제39회 하계학술대회
• /
• pp.2255-2256
• /
• 2008

음성인식 시스템을 설계함에 있어서는 대표적으로 사람의 성도 특성을 모방한 LPC(Linear Predict Cording)방식과 청각 특성을 고려한 MFCC(Mel-Frequency Cepstral Coefficients)방식이 있다. 본 논문에서는 MFCC를 통해 특징파라미터를 추출하고 해당 영역에서의 수행된 작업을 매틀랩 알고리즘을 이용하여 그래프로 시현하였다. MFCC 방식의 추출과정은 최초의 음성신호로부터 전처리과정을 통해 아날로그 신호를 디지털 신호로 변환하고, 잡음부분을 최소화하며, 음성 부분을 강조한다. 이 신호는 다시 Windowing을 통해 음성의 불연속을 제거해 주고, FFT를 통해 시간의 영역을 주파수의 영역으로 변환한다. 이 변환된 신호는 Filter Bank를 거쳐 다수의 복잡한 신호를 몇 개의 간단한 신호로 간소화 할 수 있으며, 마지막으로 Mel-cepstrum을 통해 최종적으로 특징 파라미터를 얻고자 하였다.

• 전기학회논문지
• /
• 제63권11호
• /
• pp.1559-1563
• /
• 2014

Recently, the prevention of crime using CCTV draws special in accordance with the higher crime incidence rate. Therefore security systems like a CCTV with audio capability are developing for giving an instant alarm. This paper proposes a scream detecting engine from various ambient noises in real environment for surveillance systems. The proposed engine detects scream signals among the various ambient noises using the features extracted in time/frequency domain. The experimental result shows the performance of our engine is very promising in comparison with the traditional engines using the model based features like LPC, LPCC and MFCC. The proposed method has a low computational complexity by using FFT and cross correlation coefficients instead of extracting complex features like LPC, LPCC and MFCC. Therefore the proposed engine can be efficient for audio-based surveillance systems with low SNRs in real field.

• 대한수의학회지
• /
• 제46권1호
• /
• pp.47-55
• /
• 2006

To evaluate method for monitoring anesthetic depth with quantitative electroencephalography (q-EEG), we recorded processed EEG (raw EEG) and pain score till 100 minutes in beagle dogs anesthetized for 60 minutes with propofol (n = 5, PRO group), isoflurane (n = 5, ISO group) and propofol-ketaminefentanyl (n = 5, PFK group). Raw EEG was converted into 95% spectral edge frequency (SEF) by fast Fourier transformation (FFT) method. We investigated anesthetic depth by comparing relationship (Pearson's correlation) between q-EEG (95% SEF) and pain score. Pearson's correlation coefficients are +0.2372 (p = 0.0494, PRO group), +0.79506 (p < 0.001, ISO group) and +0.49903 (p = 0.0039, PFK group).

• 전자공학회논문지B
• /
• 제29B권10호
• /
• pp.27-36
• /
• 1992

This paper proposes a numerical method of motion planning for manipulators using Foruier series. For a redundant manipulator, we predetermine the trajectories of redundant joints in terms of the Nth partial sum of the fourier series. then the optimal coefficients of the fourier series are searched by the Powell's method. For a nonredundant or redundant manipulator, CS02T-continuous smooth joint trajectory for a point-to-point task can be obtained while considering the frequency response. We apply the proposed method to the 3-link planar manipulator and the PUMA 560 manipulator. To show the validity of the proposed method, we analyze solutions by the Fast Fourier Transform (FFT). Also, several features are discussed to obtain an optimal solution.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제2권11호
• /
• pp.789-794
• /
• 2013

차량 추적 시스템(vehicle tracking system)은 교통 흐름 파악, 차량 감시, 사고 감지 등을 통하여 교통 정체에 따른 차량의 이동 경로를 유도할 수 있고, 교통사고를 사전에 방지할 수 있게 하는 시스템이다. 효과적인 차량 추적을 위해서는 먼저 연속된 영상 내의 각 객체의 특징 값을 추출하여 영상 내에 존재하는 차량 객체를 인지할 수 있어야 한다. 다음으로, 검출된 다중 객체에 대하여 영상 간 객체 매칭을 통해 연속된 프레임에 걸쳐 출현하는 동일한 차량을 인식함으로써 각 차량의 움직임을 추적할 수 있다. 본 논문에서는 차 영상의 이진화 및 레이블링(labeling)을 통하여 객체를 검출하고, 검출한 객체의 최소 외접 직사각형(minimum bounding rectangle: MBR)의 중심 좌표와 이 MBR의 가로, 세로 방향에 대한 라인(line)별 1D FFT(fast Fourier transform) 변환 결과의 평균 계수 값을 계산하여 객체의 특징 값을 구한다. 다음으로, 연속된 프레임에 걸쳐 출현하는 객체들 중 유사도가 가장 높은 객체 쌍을 동일한 객체로 인식하여 객체를 추적하는 방법을 제안한다. 실험 결과, 제안한 방법은 객체의 기하학적 특성에 기초한 기존 방법들에 비하여 정확한 추적이 가능함을 보여주었다.

• Smart Structures and Systems
• /
• 제4권3호
• /
• pp.343-365
• /
• 2008

In this study, the frequency domain method which utilizes the evaluation of changes in the structural mode shape is adopted to identify regions which contain localized damages. Frequency response function (FRF) values corresponding to the modal frequency, analogous to the mode shape coefficients, are used since change in natural frequency of the system is usually insignificant for localized damage. This method requires only few sensors to obtain the dynamic response of the structure at specific locations to determine the FRF via fast-Fourier transform (FFT). Numerical examples of an aluminum plate, which includes damages of varying severity, locations and combinations of multiple locations, are presented to demonstrate the feasibility of the method. An experimental verification of the method is also done using an aluminum plate with two different degrees of damage, namely a half-through notch and a through notch. The inconsistency in attaining the FRF values for practical applications due to varying impact load may be overcome via statistical averaging, although large variations in the loading in terms of the contact duration should still be avoided. Nonetheless, this method needs special attention when the damages induce notable changes in the modal frequency, such as when the damages are of high severity or cover more extensive area or near the boundary where the support condition is modified. This is largely due to the significant decrease in the frequency term compared to the increase in the vibration amplitude. For practical reasons such as the use of limited number of sensors and to facilitate automation, extending the resolution of this method of identification may not be efficient. Hence, methods based on wave propagation can be employed as a complement on the isolated region to provide an accurate localization as well as to trace the geometry of the damage.

• 인터넷정보학회논문지
• /
• 제8권3호
• /
• pp.33-43
• /
• 2007

본 논문에서는 하다마드 변환을 이용하여 워터마크 정보를 삽입 및 추출하는 손실압축에 강인한 워터마킹 방법을 제안한다. 하다마드 행렬은 1과 -1로 구성된 행렬이므로 계산이 빠르고 또한 역변환이 가능한 행렬이므로 워터마크 구현에 사용 가능하다. 워터마크 삽입과정에서, 워터마크는 하다마드 계수중의 중간 주파수 계수 10개를 선택하여 워터마크 패턴을 이용하여 삽입하였다. 워터마크 추출과정에서는, 삽입 시 사용했던 워터마크 패턴을 이용하여 비교하는 방법을 사용하여 워터마크 정보를 추출하였다. 실험결과, 하다마드 도메인에서 40%비트의 이진 로고영상을 삽입하였을 때 PSNR(Peok Signal To Noise Rate)이 $36{\sim}46dB$사이에서 BER(Bit Error Rate)이 $3.9{\sim}12.5%$에 달하는 성능을 나타내었고 JPEG 압축에 대해서는 QF(Quality Factor)가 30에서부터 육안으로 구분할 수 있을 정도의 로고를 추출해 낼 수 있었다. 본 논문에서는 하다마드 도메인에서의 성능을 증명하기 위하여 DCT, FFT, DWT등과 비교하여 실험한 결과 하마다드 도메인에서 가장 좋은 성능을 나타냄을 알 수 있었다.

• 한국의류학회지
• /
• 제24권4호
• /
• pp.605-615
• /
• 2000

This study was carried out to investigate sound characteristics including sound parameters and subjective sensation, and primary hand values related with sound of fabrics for blouse, and furthermore to predict subjective sound sensation with mechanical properties and sound parameters. Sound of specimens was analyzed by FFT. Level pressure of total sound(LPT), loudness(Z), coefficients of autoregressive(AR) functions for fitting the spectra, and sound color factors(ΔL and Δf) were obtained as sound parameters. Primary hand values for women's thin dress were calculated by using KES-FB. Subjective sensation for sound including softness, loudness, sharpness, clearness, roughness, highness, and pleasantness was evaluated by free modulus magnitude estimation. The results were as follows; 1. Fabrics for blouse showed similar spectral shapes to one another in that amplitude values were lower in most ranges of frequencies than fabrics for other uses. 2. It was found that fabrics for blouse were less louder because LPT, loudness(Z), and ARC values were lower than other fabrics. 3. Primary hand values indicated that specimens were soft-touched, flexible, and less crisp. Among primary hands related with sound, Shari values were higher for silk fabrics than for synthetic ones, while the values for Kishimi were similar, 4. Fabrics for blouse were rated more highly for softness, clearness, and pleasantness than for loudness, sharpness. roughness, and highness. Silk fabrics were evaluated more pleasant than synthetic fabrics. 5. Subjective sensation for sound of blouse fabrics were predicted with mechanical properties and physical sound parameters.

• 대한전기학회논문지
• /
• 제40권8호
• /
• pp.782-791
• /
• 1991

In the fields of pattern recognition such as speech recognition, several new techniques using Artifical Neural network Models have been proposed and implemented. In particular, the Multilayer Perception Model has been shown to be effective in static speech pattern recognition. But speech has dynamic or temporal characteristics and the most important point in implementing speech recognition systems using Artificial Neural Network Models for continuous speech is the learning of dynamic characteristics and the distributed cues and contextual effects that result from temporal characteristics. But Recurrent Multilayer Perceptron Model is known to be able to learn sequence of pattern. In this paper, the results of applying the Recurrent Model which has possibilities of learning tedmporal characteristics of speech to phoneme recognition is presented. The test data consist of 144 Vowel+ Consonant + Vowel speech chains made up of 4 Korean monothongs and 9 Korean plosive consonants. The input parameters of Artificial Neural Network model used are the FFT coefficients, residual error and zero crossing rates. The Baseline model showed a recognition rate of 91% for volwels and 71% for plosive consonants of one male speaker. We obtained better recognition rates from various other experiments compared to the existing multilayer perceptron model, thus showed the recurrent model to be better suited to speech recognition. And the possibility of using Recurrent Models for speech recognition was experimented by changing the configuration of this baseline model.