• Journal of Practical Engineering Education
• /
• v.14 no.3
• /
• pp.491-498
• /
• 2022

In this paper, as a liberal arts course for non-major students, a case study of recurrent neural network SW practice, which is essential for designing a basic recurrent neural network subject curriculum, was developed. The developed SW practice case focused on understanding the operation principle of the recurrent neural network, and used a spreadsheet to check the entire visualized operation process. The developed recurrent neural network practice case consisted of creating supervised text completion training data, implementing the input layer, hidden layer, state layer (context node), and output layer in sequence, and testing the performance of the recurrent neural network on text data. The recurrent neural network practice case developed in this paper automatically completes words with various numbers of characters. Using the proposed recurrent neural network practice case, it is possible to create an artificial intelligence SW practice case that automatically completes by expanding the maximum number of characters constituting Korean or English words in various ways. Therefore, it can be said that the utilization of this case of basic practice of recurrent neural network is high.

• The Journal of Society for e-Business Studies
• /
• v.4 no.1
• /
• pp.87-101
• /
• 1999

To operate the automatic devices of manufacturing process more effectively and to solve the needs of the resource sharing, network technology is applied to the control devices located in common manufacturing zone and operated by connecting them. In this paper, functional standard of the network layers are set as physical and data link layer of IEEE 802.2, 802.4, and VMD application layer and ISO-CIM reference model. Then, they are divided as minimized architecture, designed as group objects which perform group management and service objects which organizes and operates the group. For the stability in this network, this paper measures the variation of data packet length and node number and analyzes the variated value of the waiting time for the network operation. For the method of the analysis, non-exhausted service method are selected, and the arrival rates of the each data packet to the nodes that are assumed to form a Poission distribution. Then, queue model is set as M/G/1, and the analysis equation for waiting time is found. For the evalution of the performance, the length of the data packet varies from 10 bytes to 100 bytes in the operation of the group management network, the variation of the wating time is less than 10 msec. Since the waiting time in this case is less than 10 msec, response time is fast enough. Furthermore, to evaluate the real time processing of the group management network, it shows if the number of nodes is less than 40, and the average arrival time is less than 40 packet/sec, it can perform stable operation even taking the overhead such as software delay time, indicated packet service, and transmissin safety margin.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2004.05a
• /
• pp.52-55
• /
• 2004

VMS (virtual manufacturing system) may be defined as a transparent interface/control mechanism to support human decision-making via simulation and monitoring of real operating situation through modeling of all activities in RMS (real manufacturing system). The three main layers in VMS are business process layer, manufacturing execution layer, and facility operation layer, and each layer is represented by a specific software system having its own input modeler module. The current version of these input modelers has been implemented based on its own 'local' framework, and as a result, there are no information sharing mechanism, nor a common user view among them. Proposed in this paper is a unified modeler framework covering the three VMS layers, in which the concept of PPR (product-process-resource) model is employed as a common semantics framework and a 2D graphic network model is used as a syntax framework. For this purpose, abstract class PPRObject and GraphicObject are defined and then a subclass is inherited from the abstract class for each application layer. This feature would make it easier to develop and maintain the individual software systems. For information sharing, XML is used as a common data format.

• Journal of the Korean Applied Science and Technology
• /
• v.17 no.3
• /
• pp.174-177
• /
• 2000

The effect of a-sexithiophene(${\alpha}-6T$) layers on the light emitting diode (LED) were studied. The ${\alpha}-6T$ was used for a buffer layer in electroluminescent (EL) devices. Enhanced carrier (hole) injection and improved emission efficiency were observed. Carrier injection characteristics were investigated as a function of ${\alpha}-6T$ later thickness. The efficiency of the electroluminescence was proportional to the thickness of ${\alpha}-6T$ layer. The highest efficiency was observed 600A of ${\alpha}-6T$ later, which was about 1.5 times higher than that of device without ${\alpha}-6T$ later. The device with a-6T showed an operation voltage lowered by 2V. The ${\alpha}-6T$ layer can substitute hole blocking layer, and control charge injection properties.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.5 no.5
• /
• pp.1001-1010
• /
• 2001

In TLM method, Discrete Green's function ABC have been used when improved the exactness of analyzing in wide frequency band. But this technology has a complicated process to apply absorbing boundary, which means it needs additional numerical analyzing process to obtain discrete Green's function data. so, In this paper, we propose new Green's absorbing layer for simple process to apply absorbing boundary. newly proposed Green's absorbing layer is produced by applying of loss operation, loading discrete Green's function with attenuation. A state of optimum absorbing would be obtained by relation between increasing rate of loss, attenuation constant and length of green's absorbing layer. and then Analysts of waveguide BPF is carried out using Green's absorbing layer within state of optimum absorbing, then this result is in corrective agreement with the result applying traditional discrete Green's function ABC.

• Journal of Hydrogen and New Energy
• /
• v.27 no.2
• /
• pp.176-181
• /
• 2016

Recent years, supercapacitors have been attracting a growing attention as an efficient energy storage, due to their long-lifetime, device reliability, simple device structure and operation mechanism and, most importantly, high power density. Along with the increasing interest in flexible/stretchable electronics, the supercapacitors with compatible mechanical properties have been also required. A eutectic gallium-indium (EGaIn) liquid metal could be a strong candidate as a soft electrode material of the supercapacitors because of its insulating surface oxide layer for electric double layer formation. Here, we report the electrochemical study on the charging/reaction process at the interface of EGaIn liquid metal and electrolyte. Numerical fitting of the charging current curves provides the capacitance of EGaIn/insulating layer/electrolyte (${\sim}38F/m^2$). This value is two orders of magnitude higher than a capacitance of a general metal electrode/electrolyte interface.

• Journal of IKEEE
• /
• v.25 no.3
• /
• pp.493-500
• /
• 2021

In this paper, we propose the development of deep learning structure to improve quality of polygonal containers. The deep learning structure consists of a convolution layer, a bottleneck layer, a fully connect layer, and a softmax layer. The convolution layer is a layer that obtains a feature image by performing a convolution 3x3 operation on the input image or the feature image of the previous layer with several feature filters. The bottleneck layer selects only the optimal features among the features on the feature image extracted through the convolution layer, reduces the channel to a convolution 1x1 ReLU, and performs a convolution 3x3 ReLU. The global average pooling operation performed after going through the bottleneck layer reduces the size of the feature image by selecting only the optimal features among the features of the feature image extracted through the convolution layer. The fully connect layer outputs the output data through 6 fully connect layers. The softmax layer multiplies and multiplies the value between the value of the input layer node and the target node to be calculated, and converts it into a value between 0 and 1 through an activation function. After the learning is completed, the recognition process classifies non-circular glass bottles by performing image acquisition using a camera, measuring position detection, and non-circular glass bottle classification using deep learning as in the learning process. In order to evaluate the performance of the deep learning structure to improve quality of polygonal containers, as a result of an experiment at an authorized testing institute, it was calculated to be at the same level as the world's highest level with 99% good/defective discrimination accuracy. Inspection time averaged 1.7 seconds, which was calculated within the operating time standards of production processes using non-circular machine vision systems. Therefore, the effectiveness of the performance of the deep learning structure to improve quality of polygonal containers proposed in this paper was proven.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.36 no.2
• /
• pp.161-170
• /
• 2012

There are many factors to consider when attempting to improve the efficiency of fuel cell operation, such as the operation temperature, humidity, stoichiometry, operation pressure, geometric features, etc. In this paper, the effects of the operation pressure were investigated to find the current density and water saturation behavior on a cross section designated by the design geometry. A two-dimensional geometric model was established with a gas channel that can provide $H_2$ to the anode and $O_2$ and water vapor to the cathode gas diffusion layer (GDL). The results from this numerical modeling revealed that higher operation pressures would produce a higher current density than lower ones, and the water saturation behavior was different at operation pressures of 2 atm and 3 atm in the cathode GDL. In particular, the water saturation ratios are higher directly below the collector than in other areas. In addition, this paper presents the dependence of the velocity behavior in the cathode on pressure changes, and the velocity fluctuations through the GDL are higher in the output area than in inlet area. This conclusion will be utilized to design more efficient fuel cell modeling of real fuel cell operation.

• Proceedings of the IEEK Conference
• /
• 2002.07a
• /
• pp.560-563
• /
• 2002

This paper proposes a home control system architecture that have network configuration function. The proposed home control system architecture is implemented with partly hardware and software. For implementation of this system architecture, we developed ECONICS which is home automation controller. ECONICS consists of main board and communication modem. This communication modem supply the power line communication. The physical layer and the MAC layer software of EIA 709.1 standard protocol are implemented in communication modem. The upper layer software of EIA 709.1 standard protocol and home configuration software for home network installation, management, diagnostics, control and monitoring are implemented in main board of ECONICS. We verified the commercial feasibility of the proposed system through the home network configuration and operation. As a result, we have concluded that the proposed home control system architecture provides all the key function necessary to easily manage and control home network nodes.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2016.07a
• /
• pp.239-240
• /
• 2016

본 논문에서는 RPP(Rural Postman Problem)을 시뮬레이션 하기 위한 툴을 설계하고 구현하였다. RPP 문제의 해를 구하기 위해 유전 알고리즘을 시뮬레이션 툴 내부에 엔진으로 구현하였다. 시뮬레이션 툴의 구성은 유전 알고리즘의 파라미터를 설정하기 위한 사용자 인터페이스 부분과 시뮬레이션 결과를 그래프로 표현해 주는 Presentation Layer, 유전 알고리즘을 이용하여 경로탐색을 처리하는 Operation Layer, 유전 알고리즘에서 사용되는 염색체들을 저장 관리하는 Data Layer로 되어 있다. 본 논문에서 구현한 시뮬레이션 툴을 이용하여 다양한 RPP 문제를 파라미터의 설정만을 통해 해를 구할 수 있으며, 실험 결과를 그래프로 확인할 수 있다.