• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.7 no.10
• /
• pp.935-943
• /
• 2017

CNN (Convolution neural network), which is used for image classification and speech recognition among neural networks learning based on positive data, has been continuously developed to have a high performance structure to date. There are many difficulties to utilize in an embedded system with limited resources. Therefore, we use GPU (General-Purpose Computing on Graphics Processing Units), which is used for general-purpose operation of GPU to solve the problem because we use pre-learned weights but there are still limitations. Since CNN performs simple and iterative operations, the computation speed varies greatly depending on the thread allocation and utilization method in the Single Instruction Multiple Thread (SIMT) based GPGPU. To solve this problem, there is a thread that needs to be relaxed when performing Convolution and Pooling operations with threads. The remaining threads have increased the operation speed by using the method used in the following feature maps and kernel calculations.

• Proceedings of the KIEE Conference
• /
• 2009.07a
• /
• pp.1139_1141
• /
• 2009

Distributed Generation (DG) is predicted to play a important role in electric power system in the near future. insertion of DG system into existing distribution network has great impact on real-time system operation and planning. It is widely accepted that micro turbine generation (MTG) systems are currently attracting lot of attention to meet customers need in the distributed power generation market In order to investigate the performance of MT generation systems, their efficient modeling is required. This paper presents the modeling and simulation of a MT generation system suitable for grid-connected operation. The system comprises of a permanent magnet synchronous generator driven by a MT. A brief description of the overall system is given, and mathematical models for the MT and permanent magnet synchronous generator are presented. Also, the use of power electronics in conditioning the power output of the generating system is demonstrated. Simulation studies with MATLAB/Simulink have been carried out in grid-connected operation mode of a DG system. The control strategies for grid connected operation mode of DG system is also presented.

• Journal of the Korea Society of Computer and Information
• /
• v.4 no.4
• /
• pp.94-100
• /
• 1999

This paper concerns the implementation of a digital neural network which performs the high speed operation of Hopfield model's arithmetic operation. It is also designed to use a look-up table and produce floating point arithmetic of nonlinear function with high speed operation. The arithmetic processing of Hopfleld is able to describe the matrix-vector operation, which is adaptable to design the array processor because of its recursive and iterative operation .The proposed method is expected to be applied to the field of real neural networks because of the realization of the current VLSI techniques.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.4
• /
• pp.586-591
• /
• 2015

This paper proposes a force feedback system of telepresence robot for remote operation. The ultrasonic sensors attached at the robot detect the obstacles, and generate the force to the operation joystick. In order to consider the network delay, we developed the fuzzy control system using ultrasonic data and robot speed. The method to calculate the force vector from the ultrasonic data is also presented to operate the robot more accurately. The simulation and experimental results are presented to verify the safe and accurate operation of the proposed system.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.5
• /
• pp.3071-3080
• /
• 2015

The increased demand on the transport of both passengers and goods through rail systems implies higher traffic intensity and congestion on the railways, resulting in greater likelihood of accidents and also degraded passenger services. To cope with the issues, development of an integrated communication network for rails has attracted great deal of attention lately. GSM-R is such an example developed in Europe, which seems to have restrictions in providing various communication services due to network speed limit. For the reason, an LTE-based approach is under study in Korea. After the network development, operation management of the network is necessary. Design of operation management systems has been studied little and thus is the objective of this paper. To do so, a conceptual design has been carried out based on model-based approach. Specifically, a context model has first been created using the use case diagram. Then, SysML models of operational scenarios were developed for the management system. The SysML models have alternatively been expressed as EFFBD models to simulate and verify them. Consequently, the result of the conceptual system design for the operation management of the integrated wireless network is expected to be used as a basis for the detailed design later.

• Journal of Practical Engineering Education
• /
• v.15 no.2
• /
• pp.381-388
• /
• 2023

Through artificial neural network education using spreadsheets, non-major undergraduate students can understand the operation principle of artificial neural networks and develop their own artificial neural network software. Here, training of the operation principle of artificial neural networks starts with the generation of training data and the assignment of correct answer labels. Then, the output value calculated from the firing and activation function of the artificial neuron, the parameters of the input layer, hidden layer, and output layer is learned. Finally, learning the process of calculating the error between the correct label of each initially defined training data and the output value calculated by the artificial neural network, and learning the process of calculating the parameters of the input layer, hidden layer, and output layer that minimize the total sum of squared errors. Training on the operation principles of artificial neural networks using a spreadsheet was conducted for undergraduate non-major students. And image training data and basic artificial neural network development results were collected. In this paper, we analyzed the results of collecting two types of training data and the corresponding artificial neural network SW with small 12-pixel images, and presented methods and execution results of using the collected training data for Orange machine learning model learning and analysis tools.

• ETRI Journal
• /
• v.40 no.3
• /
• pp.355-365
• /
• 2018

The increase in data traffic calls for investment in mobile networks; however, the saturating revenue of mobile broadband and increasing capital expenditure are discouraging mobile operators from investing in next-generation mobile networks. Mobile network sharing is a viable solution for operators and regulators to resolve this dilemma. This research uses a difference-in-differences analysis of 33 operators (including 11 control operators) to empirically evaluate the cost reduction effect of mobile network sharing. The results indicate a reduction in overall operating expenditure and short-term capital expenditure by national roaming. This finding implies that future technology and standards development should focus on flexible network operation and maintenance, energy efficiency, and maximizing economies of scale in radio access networks. Furthermore, mobile network sharing will become more viable and relevant in a 5G network deployment as spectrum bands are likely to increase the total cost of ownership of mobile networks and technical enablers will facilitate network sharing.

• Korean Management Science Review
• /
• v.24 no.2
• /
• pp.127-142
• /
• 2007

This study deals with the optimal design of mobile access network connecting base stations(BSs) and mobile switching centers(MSCs). Generally mobile operators constitute their access networks by leasing communication lines. Using the characteristic of leased line rate based on administration region, we build an optimization model for mobile access network design which has much smaller number of variables than the existing researches. And we develop a GUI based optimization tool integrating the well-known softwares such as MS EXCEL. MS VisualBasic, MS PowerPoint and Ip_solve, a freeware optimization software. Employing the current access network configuration of a Korean mobile carrier, this study using the optimization tool obtain an optimal solution for both single MSC access network and nation-wide access network. Each optimal access network achieves 7.45% and 9.49% save of lease rate, respectively. Considering the monthly charge and total amount of lease line rate, our optimization tool provides big amount of save in network operation cost. Besides the graphical representation of access networks makes the operator easily understand and compare current and optimal access networks.

• Journal of IKEEE
• /
• v.20 no.3
• /
• pp.303-306
• /
• 2016

In this paper, we propose a method to accelerate convolutional neural network by utilizing a GPGPU. Convolutional neural network is a sort of the neural network learning features of images. Convolutional neural network is suitable for the image processing required to learn a lot of data such as images. The convolutional layer of the conventional CNN required a large number of multiplications and it is difficult to operate in the real-time on the embedded environment. In this paper, we reduce the number of multiplications through Winograd convolution operation and perform parallel processing of the convolution by utilizing SIMT-based GPGPU. The experiment was conducted using ModelSim and TestDrive, and the experimental results showed that the processing time was improved by about 17%, compared to the conventional convolution.

• Nuclear Engineering and Technology
• /
• v.54 no.11
• /
• pp.3975-3984
• /
• 2022

Distributed control system architecture is adopted for I&C systems of Prototype Fast Breeder Reactor, where the geographically distributed control systems are connected to centralized servers & display stations via switched Ethernet networks. TCP/IP communication plays a significant role in the successful operations of this architecture. The communication tasks at control nodes are taken care by TCP/IP offload modules; local area switched network is realized using layer-2/3 switches, which are finally connected to network interfaces of centralized servers & display stations. Safety, security, reliability, and fault tolerance of control systems used for safety-related applications of nuclear power plants is ensured by indigenous design and qualification as per guidelines laid down by regulatory authorities. In the case of commercially available components, appropriate suitability analysis is required for getting the operation clearances from regulatory authorities. This paper details the proposed approach for the suitability analysis of TCP/IP communication nodes, including control systems at the field, network switches, and servers/display stations. Development of test platform using commercially available tools and diagnostics software engineered for control nodes/display stations are described. Each TCP link behavior with impaired packets and multiple traffic loads is described, followed by benchmarking of the network switch's routing characteristics and security features.