• Journal of IKEEE
• /
• v.22 no.4
• /
• pp.1115-1122
• /
• 2018

In this paper, we analyze the performance degradation of previous 1-Bit convolutional neural network method and introduce ways to mitigate it. Previous work applies 32-Bit operation to first and last layers. But our method applies 32-Bit operation to second layer too. We also show that nonlinear activation function can be removed after binarizing inputs and weights. In order to verify the method proposed in this paper, we experiment the object detection neural network for korean license plate detection. Our method results in 96.1% accuracy, but the existing method results in 74% accuracy.

• Journal of IKEEE
• /
• v.27 no.4
• /
• pp.617-623
• /
• 2023

With the evolution of communication networks, there is a growing demand for stable high-speed data connections to support services relying on large-capacity data. The increasing volume of packet data aggregated from user devices underscores the significance of quality diagnostics for the backhaul network, an intermediate link transmitting data to the core network. This paper conducts empirical research on techniques to diagnose issues within the backhaul network through practical case studies, through diagnosing various factors such as circuit bandwidth, speed disparities within switches, network segment-specific buffer sizes, routing policies, among other factors that could potentially cause RTT (Round Trip Time) delays and performance degradation.

• Proceedings of the KIEE Conference
• /
• 2006.10c
• /
• pp.473-475
• /
• 2006

FieldBus communication systems in industrial wired and wireless network may cause the degradation of the TCP performance due to the racket loss. TCP is particularly targeted at the wired networks, a packet loss is assumed to be caused by the network congestion. As a result, the performance of TCP decreases significantly when used over networks that exhibit a high bit error rate. In order to solve this problem, this paper designs and implements the WFSnoop mechanism which offers a fast local retransmission. The proposed mechanism does not require any changes in customer premises. Base on the simulation in the wired and wireless network environment, we analyzed the performance of the WFSnoop mechanism.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.2198-2202
• /
• 2005

In this paper we have presented a condition monitoring method of check valve using neural network. The acoustic emission sensor was used to acquire the condition signals of check valve in direct vessel injection (DVI) test loop. The acquired sensor signal pass through a signal conditioning which are consisted of steps; rejection of background noise, amplification, analogue to digital conversion, extract of feature points. The extracted feature points which represent the condition of check valve was utilized input values of fault diagnosis algorithms using pre-learned neural network. The fault diagnosis algorithm proceeds fault detection, fault isolation and fault identification within limited ranges. The developed algorithm enables timely diagnosis of failure of check valve’s degradation and service aging so that maintenance and replacement could be preformed prior to loss of the safety function. The overall process has been experimented and the results are given to show its effectiveness.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.5
• /
• pp.2362-2378
• /
• 2017

To mitigate the performance degradation caused by performance anomaly, a number of channel assignment algorithms have been proposed for multi-rate wireless mesh networks. However, network conditions have not been fully considered for routing process in these algorithms. In this paper, a joint scheme called Multi-rate Dijkstra's Shortest path - Rate Separated (MDSRS) is proposed, combining routing metrics and channel assignment algorithm. In MDSRS, the routing metric are determined through the synthesized deliberations of link costs and rate matches; then the rate separated channel assignment is operated based on the determined routing metric. In this way, the competitions between high and low rate links are avoided, and performance anomaly problem is settled, and the network capacity is efficiently improved. Theoretical analysis and NS-3 simulation results indicate that, the proposed MDSRS can significantly improve the network throughput, and decrease the average end-to-end delay as well as packet loss probability. Performance improvements could be achieved even in the heavy load network conditions.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.59-60
• /
• 2006

TCP applied to the wireless-wired integrated network is the one that was applied to the existing wired network. In the wireless-wired integrated network, both wireless and congestion loss can occur. When wireless packet losses occur, the congestion control of TCP causes performance degradation by reducing its transmission rate. In this paper, we propose the algorithm to distinguish the wireless packet loss from congestion packet loss using MIB of the 802.11 MAC which has been generally used recently in wireless links.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.6B
• /
• pp.563-572
• /
• 2003

When the Mobile IP is applied for mobility management protocol in the next-generation mobile communication networks, it can cause a serious performance degradation called micro-mobility problem. In this paper, we propose ANMM(Access Network Mobility Management) to efficiently support micro-mobility in the IP-based wireless access network. The ANMM scheme can reduce the handoff latency and signaling overhead resulting in performance enhancement by managing the mobility of mobile nodes within access network.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.19 no.2
• /
• pp.57-68
• /
• 2023

Because salt and pepper noise is a type of impulse, even a small amount of noise could cause a large image degradation. In this paper, we proposed a salt-and-pepper noise removal method using the convolutional neural network. It consists of four phases. In the first step, the proposed method reconstructs noisy image using a traditional salt-and-pepper noise reduction method, and in the second step, the result image of previous step is filtered with Gaussian low pass filter. After that, we reconstruct the filtered image using convolution neural network. In the last step, the pixels with salt-and-pepper noise are replaced with the result of previous phase. Simulation results show that the proposed method yields not only objective image qualities(PSNR, SSIM) but also subjective image qualities for all SAP noise ratios.

• Journal of KIISE:Information Networking
• /
• v.32 no.1
• /
• pp.20-28
• /
• 2005

TCP(Transmission Control Protocol) is one of the most important Internet protocols, which is widely used in wireless networks as well as wired networks. However, when TCP is deployed for wireless networks, it takes severe performance degradation because TCP was designed for wired network. To overcome this drawback, a number of TCP variants have been proposed in the literature. However, most previous schemes did not consider TCP enhancement over heterogeneous networks. In heterogeneous networks, an mobile node (MN) may move from one access network to another(i.e., vertical handover). In the case of vertical handover, an MN experiences a TCP performance degradation caused by the packet loss and the sudden change of link characteristics between two different access networks. In this work, we investigate the TCP performance degradation occurred in vortical handover across heterogeneous networks. First, we have conducted the measurement study over GPRS-WLAN testbed. In the measurement study. we observed the TCP performance degradation in the case of handover from WLAN to GPRS. In order to study more different TCP behaviors during vertical handover, we performed comprehensive simulations using a network simulator 2(ns-2). Based on measurement and simulation results, we investigated how to improve TCP performance in vertical handover and we concluded that the existing mechanisms cannot be perfect solutions and new mechanisms are strongly required.

• Advances in materials Research
• /
• v.1 no.4
• /
• pp.245-268
• /
• 2012

Results of isothermal torsional oscillation tests are reported on melts of linear low density polyethylene and isotactic polypropylene. Prior to rheological tests, specimens were annealed at various temperatures ranging from $T_a$ = 180 to $310^{\circ}C$ for various amounts of time (from 30 to 120 min). Thermal treatment induced degradation of the melts and caused pronounced decreases in their molecular weights. With reference to the concept of transient networks, constitutive equations are developed for the viscoelastic response of polymer melts. A melt is treated as an equivalent network of strands bridged by junctions (entanglements and physical cross-links). The time-dependent response of the network is modelled as separation of active strands from and merging of dangling strands with temporary nodes. The stress-strain relations involve three adjustable parameters (the instantaneous shear modulus, the average activation energy for detachment of active strands, and the standard deviation of activation energies) that are determined by matching the dependencies of storage and loss moduli on frequency of oscillations. Good agreement is demonstrated between the experimental data and the results of numerical simulation. The study focuses on the effect of molecular weight of polymer melts on the material constants in the constitutive equations.