• Journal of the Korean Institute of Intelligent Systems
• /
• v.10 no.5
• /
• pp.449-458
• /
• 2000

In this paper, we propose a feed identification method using neural network to predict feed in crude distillation unit. The proposed FINN(feed identifier by neural network) is functionally composed of two modes-training mode and prediction mode. Also, we implement a neural network-based soft sensor system using Borland C++(3.0) Builder. The effectiveness of the proposed neural network-based feed identification method is shown by simulation results.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.20 no.5
• /
• pp.413-420
• /
• 2009

We propose a novel feed network for a $2{\times}2$ array antenna to form a sectoral conical beam. The proposed feed network, which is a symmetrical structure, consists of four $90^{\circ}$ hybrids, a crossover, and four $90^{\circ}$ delay lines. To verify the performance of the feed network a $2{\times}2$ array antenna and the feed network are fabricated on a microstrip structure, and the radiation patterns are measured at the center frequency of 2.57 GHz. The maximum radiation is measured at the $45^{\circ}$ elevation angle and at the $45^{\circ}$, $135^{\circ}$, $225^{\circ}$, and $315^{\circ}$ azimuth angles depending on the choice of the input port of the feed network.

• Journal of Digital Contents Society
• /
• v.8 no.3
• /
• pp.393-401
• /
• 2007

P2P systems have gained a lot of research interests and popularity over the years and have the capability to unleash and distribute awesome amounts of computing power, storage and bandwidths currently languishing - often underutilized - within corporate enterprises and every Internet connected home in the world. Since there is no central control over resources or devices and no before hand information about the resources or devices, device discovery remains a substantial problem in P2P environment. In this paper, we cover some of the current solutions to this problem and then propose our feed forward neural network (FFNN) based solution for device discovery in mobile P2P environment. We implements feed forward neural network (FFNN) trained with back propagation (BP) algorithm for device discovery and show, how large computation task can be distributed among such devices using agent technology. It also shows the possibility to use our architecture in home networking where devices have less storage capacity.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.19 no.8
• /
• pp.927-932
• /
• 2008

We propose a novel $4{\times}4$ Hadamard matrix feed network for a $4{\times}1$ array antenna to form a dual beam. If each element of the array is excited following the elements in a row of the Hadamard matrix, a two-lobed antenna beam can be obtained. The angle between the two lobes can be controlled. The Hadamard matrix feed network consists of four $90^{\circ}$ hybrids, a crossover and four $90^{\circ}$ phase shifters. The array, including the Hadamard matrix feed network, was fabricated on a microstip structure. The measured beam directions of the two lobes are $0^{\circ}$, ${\pm}15^{\circ}$, ${\pm}33^{\circ}$, ${\pm}45^{\circ}$ depending on the choice of the input port of the feed network.

• Journal of Multimedia Information System
• /
• v.4 no.4
• /
• pp.225-232
• /
• 2017

Understanding actions in videos is an important task. It helps in finding the anomalies present in videos such as fights. Detection of fights becomes more crucial when it comes to sports. This paper focuses on finding fight scenes in Hockey sport videos using blur & radon transform and convolutional neural networks (CNNs). First, the local motion within the video frames has been extracted using blur information. Next, fast fourier and radon transform have been applied on the local motion. The video frames with fight scene have been identified using transfer learning with the help of pre-trained deep learning model VGG-Net. Finally, a comparison of the methodology has been performed using feed forward neural networks. Accuracies of 56.00% and 75.00% have been achieved using feed forward neural network and VGG16-Net, respectively.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.24 no.4
• /
• pp.426-435
• /
• 2013

In this paper, we propose the MIMO circular polarization feed network to enhance the communication performances from the previous $2{\times}2$ MIMO channel to $4{\times}4$ channel for Land Mobile Satellite communication system. The only possibility to extend the communication channel is to use the additional satellite because of the limitation of satellite spaces to install additional antennas. For overcoming this problems, we propose the MIMO circular polarization feed network to secure the isolation characteristics without the distant antenna space. The port isolation characteristics and each port impedance matching conditions are numerically verified and we suggest the $4{\times}4$ MIMO channel model of the proposed system and the performances are verified. The fabricated circular polarization patch antennas with the proposed feed network are measured in the reverberation chamber and 7~10 dB of diversity gain and 80 % increasement of channel capacity are obtained.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.45 no.1
• /
• pp.20-27
• /
• 2008

In this paper, a nonlinear classifier of a feed-forward neural network is implemented on an FPGA chip. The feedforward neural network is implemented in hardware for fast parallel processing. After off line training of neural network, weight values are saved and used to perform forward propagation of neural processing. As an example, AND and XOR digital logic classification is conducted in off line, and then weight values are used in neural network. Experiments are conducted successfully and confirmed that the FPGA neural network hardware works well.

• Journal of Power System Engineering
• /
• v.6 no.4
• /
• pp.62-67
• /
• 2002

In this paper, the cause of the discrepancy of the inlet and outlet flow of the lubricating oil feed pump was analyzed by the flow measurement and the analysis of the flow network. At first, we thought that the flow difference was induced by a leak in the middle of the flow network. But, through the flow measurement using ultrasonic flow meter and the performance analysis of the pump, we knew that the cause of the flow difference was due to a drop in efficiency of the pump according to the pressure drop of the outlet. Also, we knew that the shape of the piping had no effect on the efficiency of the pump.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2008.03a
• /
• pp.732-736
• /
• 2008

The traditional method to calculate the gravity feed is to assume that only one tank in fuel system supplies the needed fuel to the engine, and then calculated for the single branch. Actually, all fuel tanks compete for supplying oil. Our method takes into consideration all fuel tanks and therefore, we believe, our method is intrinsically superior to traditional methods and is closer to understanding the real seriousness of the oil supply situation. Firstly, the thesis gives the mathematical model for fuel flow pipe, pump, check valve and the simulation model for fuel tank. On the basis of flow network theory and time difference method, we established a new calculation method for gravity feed oil of aeroplane fuel system, secondly. This model can solve the multiple-branch and transient process simulation of gravity feed oil. Finally, we give a numerical example for a certain type of aircraft, achieved the variations of oil level and flow mass per second of each oil tanks. In addition, we also obtained the variations of the oil pressure of the engine inlet, and predicted the maximum time that the aeroplane could fly safely under gravity feed. These variations show that our proposed method of calculations is satisfactory.

• Journal of KIISE:Software and Applications
• /
• v.29 no.5
• /
• pp.289-294
• /
• 2002

It is known that restructuring feed-forward neural network affects generalization capability and efficiency of the network. In this paper, we introduce a new approach to restructure a neural network using abstraction of the hidden knowledge that the network has teamed. This method involves extracting local rules from non-input nodes and aggregation of the rules into global rule base. The extracted local rules are used for pruning unnecessary connections of local nodes and the aggregation eliminates any possible redundancies arid inconsistencies among local rule-based structures. Final network is generated by the global rule-based structure. Complexity of the final network is much reduced, compared to a fully-connected neural network and generalization capability is improved. Empirical results are also shown.