• 한국기계기술학회지
• /
• 제20권6호
• /
• pp.776-782
• /
• 2018

After CNN basic structure was introduced by LeCun in 1989, there has not been a major structure change except for more deep network until recently. The deep network enhances the expression power due to improve the abstraction ability of the network, and can learn complex problems by increasing non linearity. However, the learning of a deep network means that it has vanishing gradient or longer learning time. In this study, we proposes a CNN structure with MLP layer. The proposed CNNs are superior to the general CNN in their classification performance. It is confirmed that classification accuracy is high due to include MLP layer which improves non linearity by experiment. In order to increase the performance without making a deep network, it is confirmed that the performance is improved by increasing the non linearity of the network.

• Journal of Communications and Networks
• /
• 제14권6호
• /
• pp.653-661
• /
• 2012

High-availability seamless redundancy (HSR) is a redundancy protocol for Ethernet networks that provides two frame copies for each frame sent. Each copy will pass through separate physical paths, pursuing zero fault recovery time. This means that even in the case of a node or a link failure, there is no stoppage of network operations whatsoever. HSR is a potential candidate for the communications of a smart grid, but its main drawback is the unnecessary traffic created due to the duplicated copies of each sent frame, which are generated and circulated inside the network. This downside will degrade network performance and might cause network congestion or even stoppage. In this paper, we present two approaches to solve the above-mentioned problem. The first approach is called quick removing (QR), and is suited to ring or connected ring topologies. The idea is to remove the duplicated frame copies from the network when all the nodes have received one copy of the sent frame and begin to receive the second copy. Therefore, the forwarding of those frame copies until they reach the source node, as occurs in standard HSR, is not needed in QR. Our example shows a traffic reduction of 37.5%compared to the standard HSR protocol. The second approach is called the virtual ring (VRing), which divides any closed-loop HSR network into several VRings. Each VRing will circulate the traffic of a corresponding group of nodes within it. Therefore, the traffic in that group will not affect any of the other network links or nodes, which results in an enhancement of traffic performance. For our sample network, the VRing approach shows a network traffic reduction in the range of 67.7 to 48.4%in a healthy network case and 89.7 to 44.8%in a faulty network case, compared to standard HSR.

• Journal of Power Electronics
• /
• 제5권1호
• /
• pp.36-44
• /
• 2005

This paper presents an implementation of high performance control of a reluctance synchronous motor (RSM) using a neural network with a direct torque control. The equivalent circuit in a RSM, which considers iron losses, is theoretically analyzed. Also, the optimal current ratio between torque current and exiting current is analytically derived. In the case of a RSM, unlike an induction motor, torque dynamics can only be maintained by controlling the flux level because torque is directly proportional to the stator current. The neural network is used to efficiently drive the RSM. The TMS320C3l is employed as a control driver to implement complex control algorithms. The experimental results are presented to validate the applicability of the proposed method. The developed control system shows high efficiency and good dynamic response features for a 1.0 [kW] RSM having a 2.57 ratio of d/q.

• 한국정보통신학회논문지
• /
• 제7권1호
• /
• pp.158-163
• /
• 2003

다중 디스크 드라이브가 포함된 디스크 배열 시스템에서 디스크의 수가 증가 될 때 시스템 성능은 컨트롤러의 집중화 또는 버스로 사용되는 전송 경로의 병목현상에 의해 제한되어진다. 이러한 단점을 보완하기 위해 고성능 대용량의 RAID가 등장하였으며 RAID 시스템에서 컨트롤러 기능은 모든 디스크 드라이브에 분산되고 각 디스크는 그들의 임무를 수행하는 자율성을 가진 자율적 네트웍이 일반적 계층 시스템 보다 확장성이 좋고 시스템 자원을 보다 효율적으로 이용할 수 있어 디스크 수의 증가율에 따라 높은 읽기/쓰기 처리율의 성능을 제공한다.

• 한국콘텐츠학회:학술대회논문집
• /
• 한국콘텐츠학회 2004년도 추계 종합학술대회 논문집
• /
• pp.395-401
• /
• 2004

In the current high network-centralized computing and enterprising environment, it is getting essential to transmit data reliably at very high rates. Until now previous client/server model based NFS(Network File System) or AFS(Andrew's Files System) have met the various demands but from now couldn't satisfy those of the today's scalable high-performance computing environment. Not only performance but data sharing service redundancy have risen as a serious problem. In case of NFS, the locking issue and cache cause file system to reboot and make problem when it is used simply as ip-take over for H/A service. In case of AFS, it provides file sharing redundancy but it is not possible until the storage supporting redundancy and equipments are prepared. Lustre is an open source based cluster file system developed to meet both demands. Lustre consists of three types of subsystems : MDS(Meta-Data Server) which offers the meta-data services, OST(Objec Storage Targets) which provide file I/O, and Lustre Clients which interact with OST and MDS. These subsystems with message exchanging and pursuing scalable high-performance file system service. In this paper, we compare the transmission speed of gigabytes file between Lustre and NFS on the basis of concurrent users and also present the high availability of the file system by removing more than one OST in operation.

• International Journal of Quality Innovation
• /
• 제9권2호
• /
• pp.99-117
• /
• 2008

High-tech industry has been the principal economic source for Taiwan in recent years. The characteristics of high-tech industries in Taiwan are changeable product markets, short product life cycles and high company attrition rate. In the globalization trend, the high-tech industry has gradually increased corporate competitiveness and reached the goal of sustainable operations through knowledge management, knowledge sharing and new product research and development. Firms have aggressively strengthened and integrated their internal and external resources and enhanced knowledge sharing to increase industry operational performance. Effectively strengthening the knowledge management operation and performance evaluation of knowledge sharing in Taiwan's high-tech industry has become a critical issue. In the selection of knowledge sharing Key Performance Indicators (KPI), this research divided the knowledge sharing indicators into representative strategic indicators such as organizational knowledge learning, organizational knowledge resources, organizational information capacity and organizational knowledge performance through screening using Factor Analysis. The characteristics of the constructs were interdependent. This research calculated and adjusted the correlation among the key performance knowledge sharing indicators using ANP and determined the relative weight of knowledge sharing.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2005년도 추계학술대회 및 정기총회
• /
• pp.361-364
• /
• 2005

As increasing the Internet traffic, many researches on access network are reported for end-to-end high-speed broadband network, where E-PON(Ethernet-Passive Active Network) is one of reasonable candidate fur the network with respect to cost and bandwidth utilization. For the high-speed access network, E-PON need an efficient bandwidth allocation method. This paper suggests a WFA(Water-Filling Allocation) algorithm for the efficient bandwidth allocation with various simulation test.

• 대한전자공학회논문지TC
• /
• 제47권11호
• /
• pp.97-104
• /
• 2010

본 논문은 고성능32비트 AES구조를 제시한다. 재배열 구조는 5단 파이프라인을 사용한다. 그 안에 ShiftRows/InvShiftRows 모듈은 4단 파이프라인을 사용하고 MixColumn/InvMixColumn 모듈은 1단 파이프라인을 사용한다. Shift rows와 inverse shift rows 같은 구조를 사용한다. Mix column 과 inverse mix column 도 같은 구조를 사용한다. 그리고 RCON구조를 단순화 하여 사이즈를 줄였다. 제안된 구조는 verilogHDL 을 이용하여 구현 하였다. 이 회로의 처리량은 415Mbits/s 이고 크기는 0.18um CMOS 공정에서 13,764 게이트 이다. 재배열 구조는 무선 네트워크 라우터에서 사용할 수 있다.

• 한국산학기술학회논문지
• /
• 제12권11호
• /
• pp.4801-4811
• /
• 2011

본 연구는 사회적 기업의 내부 및 외부 환경요인과 기업가정신이 사회적 기업의 성과에 영향을 미치는 요인과 네트워크 활동의 조절효과를 검증하는 것이 목적이다. 연구의 목적을 달성하기 위해서 사회적 기업의 성과, 내부 및 외부 환경, 기업가정신과 네트워크 활동에 관한 이론적 배경을 바탕으로 연구모형과 가설을 설정하였다. 본 연구에서 제안된 연구모형을 사회적 기업가를 대상으로 분석하였다. 분석결과, 첫째, 조직문화가 기업가정신에 정(+)의 영향을 미치는 것으로 나타났다. 둘째, 환경 불확실성이 기업가정신에 정(+)의 영향을 미치는 것으로 나타났으며, 셋째, 기업가정신이 사회적 기업의 경영성과 및 사회적 성과를 향상시키는 것으로 나타났다. 마지막으로, 네트워크 활동에 따라 기업가정신에 미치는 영향력에 차이가 있는 것으로 나타났다. 네트워크 활동이 높은 집단이 낮은 집단 보다 조직문화가 기업가정신에 미치는 영향력이 더 큰 것으로 나타났으며, 네트워크 활동이 낮은 집단이 높은 집단 보다 환경 불확실성이 기업가정신에 미치는 영향력이 더 큰 것으로 나타났다.

• 대한임베디드공학회논문지
• /
• 제2권1호
• /
• pp.17-23
• /
• 2007

The home network has been revitalized due to the emergence of the high-speed broadband network and explosive increases of the home network services. Especially multicast based high-definition (HD) IPTV service is on high demand from both Internet service provider (ISP) and users. In order to provide high quality HD IPTV service without packet loss and delay, the ISP utilizes various network QoS mechanisms. Due to the misleading fact, that the home gateway is considered as an end terminal in terms of the ISP's viewpoint, the ISPs can not guarantee the end-to-end QoS within the scope of the current home network architecture. Therefore, an independent QoS guaranteed mechanism should be considered within the home network domain. As the home gateway manages both private home and public access network within the home network, we propose and implement a novel QoS mechanism for the multicast based HD IPTV service over the home gateway. In order to provide QoS guaranteed HD IPTV service, the proposed mechanism classifies the packets, manages home network bandwidth resources, and executes call admission control using the channel join message. Also, we utilize layer 2 multicast packet forwarding scheme to improve the overall performance of the home network services. Through the experiments of implementing the multicast based HD IPTV services, we will verify the efficiency of the proposed mechanism.