• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.9 no.5
• /
• pp.1134-1138
• /
• 2008

In this paper, the synthesis of distributed impedance transformers is presented that is essential for power divider design, whereby a broadband power divider is designed. Transfer functions of distributed transformers are synthesized with Chebyshev approximation, and their element values are calculated for various minimum insertion losses(MIL) and ripples. Desired performance of transformers is obtained by optimizing MIL's and ripples of a transfer function. As an application example, a four-way power divider is designed that operates over 2 to 8GHz frequency range. Experimental results are shown to approach the design performance, so transformer design by distributed network synthesis proves to be useful to power divider design.

• Geophysics and Geophysical Exploration
• /
• v.10 no.4
• /
• pp.259-267
• /
• 2007

An efficient algorithm for inverting static-shifted magnetotelluric (MT) data has been proposed to produce a three-dimensional (3D) resistivity model. In the Gauss-Newton approach, computational costs associated with construction of a full sensitivity matrix usually make 3D MT inversion impractical. This computational difficulty may be overcome by using approximate sensitivities. We use four kinds of sensitivities in particular orders in the inversion process. These sensitivities are computed 1) analytically for an initial, homogeneous earth, 2) exactly for a current model, 3) approximately by the Broyden method, and 4) approximately using the previous adjoint fields. Inversion experiments with static-shifted synthetic and field MT data indicate that inversion results are highly dependent on characteristics of data and thus applying various combinations of sensitivities is helpful in obtaining a good image of the subsurface structure with reasonable computation time.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.237-241
• /
• 2001

본 논문에서는 TSK (Takagi-Sugeno-Kang) 형태의 퍼지모델을 유도하는데 있어서, 동적시스템의 비선형 미분방정식을 선형화시 off-equilibrium에서 발생할 수 있는 상수항을 배제하고, TSK 퍼지 모델의 전건부 소속함수들을 GA(Genetic Algorithm)을 이용하여 최적화한후 이를 퍼지를 이용하여 합성함으로써, 실제 동적시스템을 묘사하는 비선형 미분방정식에 최적 근사화된 TSK 퍼지 모델링기법을 제시한다.

• Proceedings of the Korean Society of Coastal and Ocean Engineers Conference
• /
• 2000.09a
• /
• pp.120-125
• /
• 2000

연안역의 토지이용이 극대화됨에 따라 연안구조물이나 항만구조물의 축조가 활발히 진행되고 있다. 이를 설계하는데 있어서 무엇보다 중요한 것은 천해설계파 등과 같은 파랑변형계산을 얼마만큼 실현상과 근사시켜 재현할 수 있는지의 여부에 달려 있다. 지금까지의 천해설계파 산정에 있어서는 주로 단일 주파수와 단일 방향에 대한 규칙파해석이 수행되어져 왔으나, 실제해역에서의 파랑은 수많은 주파수와 파향, 파고들이 합성되어 이루어진 다방향 불규칙파랑이므로 실제 해역의 자연현상이 잘 표현되었다고 보기는 어렵다. (중략)

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.227-230
• /
• 1999

잔향에 대한 인간의 주관적인 지각을 잔향기 설계에 객관적인 수치로써 반영하는 방법으로, 5 자유도 잔향 모델이 제안된 바 있다[1]. 5자유도 잔향 모델은 잔향에 대한 다섯 개의 객관적인 평가량들을 이용하여 시간에 따른 음 에너지 감쇠 곡선을 근사화한 것이다. 즉 5 자유도 잔향 모델을 이용하여 청취자가 원하는 특성을 갖는 잔향을 객관적으로 묘사할 수 있고, 이는 잔향을 합성할 때 잔향 필터의 설계 기준이 된다. 그러나 이 모델로부터 만들 수 있는 잔향 필터의 개수는 실로 무한하고, 그 중에는 인간이 듣기에 부자연스러운 합성음을 만들어 내는 경우도 있다. 즉 자연스러운 잔향을 만들기 위해서는 잔향 모델 외에도 부가적인 잔향 설계 기준이 필요하다. 시간, 주파수 영역에서 대표적인 특성을 갖는 몇 종류의 원음에 대한 청음실험을 통해, 필요한 잔향 설계 기준을 제시한다.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.7
• /
• pp.521-531
• /
• 2004

In this paper we developed the Physical modeling of the Sanio Gayageum using the improved digital waveguide theory. The frequency characteristics of the Gayageum body is implemented by an inverse filtering and the impulse response of the body. We obtained the synthesis sounds of the unit sound for the Gayageum using the simulation of the straight-line fits by the changes of the fundamental frequencies depending on the Amok location. Finally. we could obtain the virtual Sanio Gayageum sounds similar to the actual Gayageum by tuning the Amok positions.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.280-282
• /
• 2020

최근 사물인터넷(IoT), 자율주행과 같이 기계 간의 통신이 요구되는 서비스가 늘어감에 따라, 기계 임무 수행에 최적화된 데이터의 생성 및 압축에 대한 필요성이 증가하고 있다. 또한, 사물인터넷과 인공지능(AI)이 접목된 기술이 주목을 받으면서 딥러닝 모델에서 추출되는 특징(feature)을 디바이스에서 클라우드로 전송하는 방안에 관한 연구가 진행되고 있으며, 국제 표준화 기구인 MPEG에서는 '기계를 위한 부호화(Video Coding for Machine: VCM)'에 대한 표준 기술 개발을 진행 중이다. 딥러닝으로 특징을 추출하는 가장 대표적인 방법으로는 합성곱 신경망(Convolutional Neural Network: CNN)이 있으며, 오토인코더는 입력층과 출력층의 구조를 동일하게 하여 출력을 가능한 한 입력에 근사시키고 은닉층을 입력층보다 작게 구성하여 차원을 축소함으로써 데이터를 압축하는 딥러닝 기반 이미지 압축 방식이다. 이에 본 논문에서는 이러한 오토인코더의 성질을 이용하여 CNN 기반의 이미지 분류 네트워크의 합성곱 신경망으로부터 추출된 feature에 오토인코더를 적용하여 압축하는 방안을 제안한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.6C
• /
• pp.572-578
• /
• 2007

To establish eye contact in video teleconferencing, it is necessary to synthesize a front view image by viewpoint interpolation. We can find the viewing direction of a user, and interpolate an image seen from that viewpoint, which will result in a face image observed from the front. There are two categories of previous research: image based method and model based method. The former is simple to calculate, however, it shows limited performance for complex objects. And the latter is robust to noise while it is computationally expensive. We propose to approximate face images as ellipses and match them to build an ellipsoid and then synthesize a new image from a given virtual camera position. We show that it is simple and robust from various experiments.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.46 no.11
• /
• pp.50-55
• /
• 2009

In this paper, a distributed amplifier is designed using distributed network synthesis that provides the optimum absorption capability of a capacitance. Transfer functions of filters, which consist of the amplifier, are synthesized by a low-pass Chebyshev approximation. Capacitances that a filter network can absorb are calculated as a function of its minimum insertion loss(MIL) and ripple. Active devices in a distributed amplifier are modeled as equivalent circuits by using their S-parameters, and their equivalent capacitances are absorbed into filter structures by properly adjusting the MIL and ripple of a transfer function. As an application example, a distributed amplifier with the gain of about 12.5dB is designed that operates over the frequency range between 0.1 and 7.5GHz. Experimental results prove that distributed network synthesis, which considers capacitance absorption capability, is useful to the design of distributed amplifiers.

• Journal of Wetlands Research
• /
• v.21 no.2
• /
• pp.152-156
• /
• 2019

The removal property of magnetic zeolite for the adsorption of aqueous ammonium ion was examined in this work. The surface modified magnetic zeolite was produced by hydrothermal synthesis. The complex of zeolite and $Fe_3O_4$ was established by the observation of SEM and XRD analysis and less than 12.6% of $Fe_3O_4$ content in magnetic zeolite was observed in the form of $Fe_3O_4$ particles. The optimum pH of adsorption was shown around 8 and the maximum adsorption linearly decreased with the increase of $Fe_3O_4$ content. The adsorption isotherm for aqueous ammonium ion was approximated by Langmuir equation. The developed surface modified magnetic zeolite adsorbent was recommended to control the nitrogen pollution for wetland environment system.