• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.27 no.7
• /
• pp.625-634
• /
• 2016

A full digital receiving active phased array antenna generates final receiving beams by digital beam forming of received digital signals in element-level that makes difficult to use typical near-field measurement method. Thus in this paper, a modified near-field measurement method which is suitable for full digital receiving active phased array antenna is proposed. Also the measured results of receiving beam pattern and G/T using developed L-band full digital receiving active phased array antenna are shown for the verification of proposed method.

• Journal of Digital Convergence
• /
• v.17 no.1
• /
• pp.123-130
• /
• 2019

One of the most commonly used methods of web recommendation techniques is collaborative filtering. Many studies on collaborative filtering have suggested ways to improve accuracy. This study proposes a method of movie recommendation using Word2Vec and an ensemble convolutional neural networks. First, in the user, movie, and rating information, construct the user sentences and movie sentences. It inputs user sentences and movie sentences into Word2Vec to obtain user vectors and movie vectors. User vectors are entered into user convolution model and movie vectors are input to movie convolution model. The user and the movie convolution models are linked to a fully connected neural network model. Finally, the output layer of the fully connected neural network outputs forecasts of user movie ratings. Experimentation results showed that the accuracy of the technique proposed in this study accuracy of conventional collaborative filtering techniques was improved compared to those of conventional collaborative filtering technique and the technique using Word2Vec and deep neural networks proposed in a similar study.

• Journal of Korean Society of Steel Construction
• /
• v.19 no.5
• /
• pp.503-513
• /
• 2007

The seismic responses of a building are affected by the connection characteristics that have effects on structural stiffness. In this study, push-over analysis and time history analysis were performed to estimate structural behavior of 2D eight-story unbraced steel structures with partially restrained composite connections using a nonlinear dynamic analysis program. Nonlinear $M-{\theta}$characteristics of connection and material inelastic characteristics of composite beam and steel column were considered. The idealization of composite semi-rigid connection as fully rigid connection yielded an increase in initial stiffness and ultimate strength in the push-over analysis. In time history analysis, the stiffness and hysteretic behavior of connections have effects on base-shear force, maximum story-drift and maximum moment in beams and columns. For seismic waves with PGA of 0.4 g, the structure with the semi-rigid composite connections shows the maximum story-drift with less than the life safety criteria by FEMA 273 and no inelastic behavior of beam and column, whereas in the structure with rigid connections, beams and columns have experienced inelastic behaviors.

• Journal of Korean Society of Steel Construction
• /
• v.20 no.4
• /
• pp.493-504
• /
• 2008

This paper deals with the experimental analysis of the flexural behaviour of encased composite beams with hollow core slabs and channels. The shear force between steel beams and hollow core slabs are transferred by channels. Three full-scale specimens were constructed and tested with different steel beam heights, which were compared with those of previous studies. Based on observation of the experiments, the encased composite beams exhibited full shear connection behaviour without any other shear connectors due to their inherent mechanical and chemical bond stress. Experimental results show a behaviour similar to steel-concrete composite beams with classical connectors: elastic and yield domains, great ductility, flexural failure mode (plastic hinge), low relative movement at steel-concrete interface and all specimens failed in a very ductile manner. Consequently, this study enables the validation of the proposed connection device under static loading and shows that it meets modern structural requirements.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.1081-1088
• /
• 2020

In this paper, the method of detecting sea fog through CCTV image is proposed based on convolutional neural networks. The study data randomly extracted 1,0004 images, sea-fog and not sea-fog, from a total of 11 ports or beaches (Busan Port, Busan New Port, Pyeongtaek Port, Incheon Port, Gunsan Port, Daesan Port, Mokpo Port, Yeosu Gwangyang Port, Ulsan Port, Pohang Port, and Haeundae Beach) based on 1km of visibility. 80% of the total 1,0004 datasets were extracted and used for learning the convolutional neural network model. The model has 16 convolutional layers and 3 fully connected layers, and a convolutional neural network that performs Softmax classification in the last fully connected layer is used. Model accuracy evaluation was performed using the remaining 20%, and the accuracy evaluation result showed a classification accuracy of about 96%.

• Journal of the Korean Geotechnical Society
• /
• v.19 no.5
• /
• pp.35-47
• /
• 2003

In the conventional axial load transfer analysis for composite piles (i.e., steel pipe pile filled with concrete), it was assumed that the concrete's strain is same as the measured steel's strain and the elastic modulus of the steel and the concrete calculated by formular as prescribed by specification is used in calculation of pile axial load. But, the pile axial load calculated by conventional method had some difference with the actual pile load. So, the behavior of a composite pile could not be analyzed exactly. Thus, the necessity to measure the strain for each pile components was proposed. In this study, the verification test for DRS (Deformable Rod Sensor) developed to measure the strain of each pile component (i.e., the steel and the concrete) was performed. In the calculation of pile axial load using the DRS, elastic modulus of concrete could be determined by the uniaxial compression test for the concrete cylinder samples made in the test site and an average tangential modulus in the stress range of (0.2∼0.6)f$_ck$ was taken.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.7
• /
• pp.213-220
• /
• 2020

The convolutional neural network is composed of convolutional layers and fully connected layers. The nonlinear activation function is used in each layer of the convolutional layer and the fully connected layer. The activation function being used in a neural network is a function that simulates the method of transmitting information in a neuron that can transmit a signal and not send a signal if the input signal is above a certain criterion when transmitting a signal between neurons. The conventional activation function does not have a relationship with the loss function, so the process of finding the optimal solution is slow. In order to improve this, an agile activation function that generalizes the activation function is proposed. The agile activation function can improve the performance of the deep neural network in a way that selects the optimal agile parameter through the learning process using the primary differential coefficient of the loss function for the agile parameter in the backpropagation process. Through the MNIST classification problem, we have identified that agile activation functions have superior performance over conventional activation functions.

• Journal of Korean Society of Steel Construction
• /
• v.15 no.5 s.66
• /
• pp.499-507
• /
• 2003

As the load is increased on the steel-concrete composite structure, its interface begins to show nonlinear behavior due to the reduction of interaction, micro-crack, slip and separation, and it causes slip-softening, Therefore, it is essential to consider the partial-interaction analysis technique. Until now, however, full-interaction or, in some instances, the linear-elastic model, which are insufficient to simulate accurate behavior, are assumed in the analysis of composite structure since the analysis method and nonlinear model for interface are very difficult and complicated. Therefore, the design of composite structure is followed by the experimental method which is inefficient-because a number of tests have to be carried out according to the design environments. In this study, we carried out the nonlinear analysis according to various interface nonlinear models by interaction magnitude, and analyzed more accurate structural behavior and performance by maximum tangential traction and slip-softening at the interface. As a result of this study. we were able to prove that the nonlinear model of interface more exactly represents behavior after yielding, such as ultimate load: that initial tangential stiffness of interface has a significant effect on the yielding load of structural members or part: and that the maximum tangential traction and slip-softening mainly effects structural yielding and ultimate load. Therefore, the structural performance of composite structure is highly dependent on the steel-concrete interface or interaction, which may result in initial tangential stiffness, maximum tangential traction and slip-softening in nonlinear model.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.26-27
• /
• 2021

본 논문에서는 자막방송 제공을 위해 방송콘텐츠를 이해하는 방법으로 잔차 합성곱 순환신경망 기반 음향 사건 분류 기법을 제안한다. 제안된 기법은 잔차 합성곱 신경망과 순환 신경망을 연결한 구조를 갖는다. 신경망의 입력 특징으로는 멜-필터벵크 특징을 활용하고, 잔차 합성곱 신경망은 하나의 스템 블록과 5개의 잔차 합성곱 신경망으로 구성된다. 잔차 합성곱 신경망은 잔차 학습으로 구성된 합성곱 신경망과 기존의 합성곱 신경망 대비 특징맵의 표현 능력 향상을 위해 합성곱 블록 주의 모듈로 구성한다. 추출된 특징맵은 순환 신경망에 연결되고, 최종적으로 음향 사건 종류와 시간정보를 추출하는 완전연결층으로 연결되는 구조를 활용한다. 제안된 모델 훈련을 위해 라벨링되지 않는 데이터 활용이 가능한 평균 교사 모델을 기반으로 훈련하였다. 제안된 모델의 성능평가를 위해 DCASE 2020 챌린지 Task 4 데이터 셋을 활용하였으며, 성능 평가 결과 46.8%의 이벤트 단위의 F1-score를 얻을 수 있었다.

• Korean Journal of Optics and Photonics
• /
• v.15 no.2
• /
• pp.130-136
• /
• 2004

We investigate characteristics of viewing resolution improvement in three-dimensional integral imaging, when a dynamic lens array method is adopted. We show that the viewing resolution changes for different moving directions and distances of the lens array through computer-synthesized integral imaging. From this study, optimal moving conditions of the lens array for efficient viewing resolution improvement can be determined.