• Journal of Advanced Navigation Technology
• /
• v.17 no.5
• /
• pp.475-483
• /
• 2013

In this paper, the performances of modeling methods for combining corrections from multiple reference stations for network user were compared and analyzed. The longer the distance between reference station and user is, the more the GPS errors are decorrelated. Based on this point, multiple corrections from reference stations which is constituting a network should be combined properly to be applied for user observation to eliminate GPS errors. There are many widely used conventional modeling methods and they are applied for Compact Network RTK users and user position accuracy is predicted by using residual errors in observation of user. Compact Network RTK is a technique of generating corrections which was developed by Seoul National University. As a result, the horizontal and vertical accuracies were within about 5 cm and 7 cm respectively with 95 % probability for all conventional modeling methods. In addition, we analyzed condition for reference station constellation for modeling method using height information.

• Journal of Korea Port Economic Association
• /
• v.25 no.3
• /
• pp.251-268
• /
• 2009

This study analyzes price competition between two ports theoretically using a game model. We considered two cases: in Case I, consignors are located in a given position, and in Case II, consignors are distributed uniformally between two ports. The results are as follows. In Case I, the higher consignors' preferences for quality are, the more two ports' prices increase. As the locations of consignors are closer to Port H, the price of Port H increases and that of the low quality port(Port L) decreases. In addition, when transportation cost increases, the price of Port L decreases and the price of Port H tends to increase. If the quality of Port H improves, the price of the port H increases but the price of Port L is not clearly determined. In Case II, the higher consignors' preferences for quality are, the more two ports' prices increase. As transportation cost increases, the prices of both ports decrease but the price of Port L decreases twice as fast as that of Port H. In addition, if the quality of Port H improves, the price of Port H increases but the price of Port L decreases when transportation cost is high.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.10
• /
• pp.391-398
• /
• 2021

Since the leased line is a structure that exclusively uses two connected areas for data transmission, a stable quality level and security are ensured, and despite the rapid increase in the number of switched lines, it is a line method that is continuously used a lot in companies. However, because the cost is relatively high, one of the important roles of the network operator in the enterprise is to maintain the optimal state by properly arranging and utilizing the resources of the network leased line. In other words, in order to properly support business service requirements, it is essential to properly manage bandwidth resources of leased lines from the viewpoint of data transmission, and properly predicting and managing leased line usage becomes a key factor. Therefore, in this study, various prediction models were applied and performance was evaluated based on the actual usage rate data of leased lines used in corporate networks. In general, the performance of each prediction was measured and compared by applying the smoothing model and ARIMA model, which are widely used as statistical methods, and the representative models of deep learning based on artificial neural networks, which are being studied a lot these days. In addition, based on the experimental results, we proposed the items to be considered in order for each model to achieve good performance for prediction from the viewpoint of effective operation of leased line resources.

• KIPS Transactions on Computer and Communication Systems
• /
• v.10 no.7
• /
• pp.191-198
• /
• 2021

Since the size of training dataset become large and the model is getting deeper to achieve high accuracy in deep learning, the deep neural network training requires a lot of computation and it takes too much time with a single node. Therefore, distributed deep learning is proposed to reduce the training time by distributing computation across multiple nodes. In this study, we propose hybrid allreduce strategy that considers the characteristics of each layer and communication and computational overlapping technique for synchronization of distributed deep learning. Since the convolution layer has fewer parameters than the fully-connected layer as well as it is located at the upper, only short overlapping time is allowed. Thus, butterfly allreduce is used to synchronize the convolution layer. On the other hand, fully-connecter layer is synchronized using ring all-reduce. The empirical experiment results on PyTorch with our proposed scheme shows that the proposed method reduced the training time by up to 33% compared to the baseline PyTorch.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.32 no.6
• /
• pp.481-495
• /
• 2020

After the resonator on the basis of the wave-filter theory was designed to control the waves with a specific frequency range surging into the harbor, the several case with the use of resonator have been reported in some part of sea, including the port of Long Beach, USA, and yacht harbor at Rome, Italy in order to control the long-period wave motion from the vessels. Recently, the utility and applicability of the resonator has been sufficiently verified in respect of the control of tsunami approximated as the solitary wave and/or the super long-period waves. However, the case with the application of tsunami in the real sea have not been reported yet. In this research, the respective case with the use of existing resonator at the port of Mukho and Imwon located in the eastern coast of South Korea were studied by using the numerical analysis through the COMCOT model adapting the reduction rate of 1983 Central East Sea tsunami and 1993 Hokkaido Southwest off tsunami. Consequently, the effectiveness of resonator against tsunami in the real sea was confirmed through the reduction rate of maximum 40~50% at the port of Mukho, and maximum 21% at the port of Imwom, respectively. In addition, it was concluded that it is necessary to study about the various case with application of different shape, arrangement, and size of resonator in order to design the optimal resonator considering the site condition.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.32 no.6
• /
• pp.496-505
• /
• 2020

In the previous research, the effectiveness of resonator was confirmed through the numerical analysis on two cases with the use of existing resonator at the Mukho and Imwon ports located in the eastern coast of South Korea by discussing the reduction rates of 1983 Central East Sea tsunami, and 1993 Hokkaido Southwest off tsunami, respectively. In this study, the reduction rates of tsunami height with three different resonators, Type I, II-1, and II-2, at the Samcheok port were examined respectively through the numerical analysis using COMCOT model under the same condition as the previous study. It was discussed the spatial distribution of maximum height of tsunami, change of water level, and effectiveness of resonator with the presence of new types of resonator, and change of their sizes. As a result, the effectiveness of resonator was verified through the application of new types of resonator reducing about maximum 40% of tsunami height. In order to design the optimal resonator for the variety of site condition, it is necessary to research about the various cases applying different shape, arrangement, and size of resonator as further study.

• KIPS Transactions on Computer and Communication Systems
• /
• v.9 no.12
• /
• pp.291-306
• /
• 2020

Nowadays, Data-Network-AI (DNA)-based intelligent services and applications have become a reality to provide a new dimension of services that improve the quality of life and productivity of businesses. Artificial intelligence (AI) can enhance the value of IoT data (data collected by IoT devices). The internet of things (IoT) promotes the learning and intelligence capability of AI. To extract insights from massive volume IoT data in real-time using deep learning, processing capability needs to happen in the IoT end devices where data is generated. However, deep learning requires a significant number of computational resources that may not be available at the IoT end devices. Such problems have been addressed by transporting bulks of data from the IoT end devices to the cloud datacenters for processing. But transferring IoT big data to the cloud incurs prohibitively high transmission delay and privacy issues which are a major concern. Edge computing, where distributed computing nodes are placed close to the IoT end devices, is a viable solution to meet the high computation and low-latency requirements and to preserve the privacy of users. This paper provides a comprehensive review of the current state of leveraging deep learning within edge computing to unleash the potential of IoT big data generated from IoT end devices. We believe that the revision will have a contribution to the development of DNA-based intelligent services and applications. It describes the different distributed training and inference architectures of deep learning models across multiple nodes of the edge computing platform. It also provides the different privacy-preserving approaches of deep learning on the edge computing environment and the various application domains where deep learning on the network edge can be useful. Finally, it discusses open issues and challenges leveraging deep learning within edge computing.

• Journal of the Korea Convergence Society
• /
• v.12 no.3
• /
• pp.9-15
• /
• 2021

With the recent development of 5G and artificial intelligence technologies, it is interested in AIOT technology to collect, process, and analyze information in cloud edge environments. AIIoT technology is being applied to various smart environments, but research is needed to perform fast response processing through accurate analysis of collected information. In this paper, we propose a technique to minimize bandwidth and processing time by blocking the connection processing between AIOT information through fast processing and accurate analysis/forecasting of information collected in the smart environment. The proposed technique generates seeds for data indexes on AIOT devices by multipointing information collected by blockchain, and blocks them along with collection information to deliver them to the data center. At this time, we deploy Deep Neural Network (DNN) models between cloud and AIOT devices to reduce network overhead. Furthermore, server/data centers have improved the accuracy of inaccurate AIIoT information through the analysis and predicted results delivered to minimize latency. Furthermore, the proposed technique minimizes data latency by allowing it to be partitioned into a layered multilayer network because it groups it into blockchain by applying weights to AIOT information.

• Clean Technology
• /
• v.27 no.1
• /
• pp.61-68
• /
• 2021

Selective catalytic reduction (SCR) is widely used as a method of removing nitrogen oxide in large-capacity thermal power generation systems. Uniform mixing of the injected ammonia and the inlet flue gas is very important to the performance of the denitrification reduction process in the catalyst bed. In the present study, a computational analysis technique was applied to the ammonia injection system design process of a denitrification facility. The applied model is the denitrification facility of an 800 MW class coal-fired power plant currently in operation. The flow field to be solved ranges from the inlet of the ammonia injection system to the end of the catalyst bed. The flow was analyzed in the two-dimensional domain assuming incompressible. The steady-state turbulent flow was solved with the commercial software named ANSYS-Fluent. The nozzle arrangement gap and injection flow rate in the ammonia injection system were chosen as the design parameters. A total of four (4) cases were simulated and compared. The root mean square of the NH3/NO molar ratio at the inlet of the catalyst layer was chosen as the optimization parameter and the design of the experiment was used as the base of the optimization algorithm. The case where the nozzle pitch and flow rate were adjusted at the same time was the best in terms of flow uniformity.

• Journal of Wetlands Research
• /
• v.23 no.1
• /
• pp.74-84
• /
• 2021

Active stormwater management is essential to minimize the impact of urban development and improve the hydrological cycle system. In recent years, the Low Impact Development (LID) technique for urban stormwater management is attracting attention as a reasonable alternative. The Storm Water Management Model (SWMM) is actively used in urban hydrological cycle improvement projects as it provides simulation functions for various GI (Green Infra) facilities through its LID module. However, in order to simulate GI facilities using SWMM, there are many difficulties in setting up complex watersheds and deploying GI facilities. In this study, a model that can evaluate the performance of GI facilities is proposed while implementing the core hydrological process of GI facilities. Since the proposed model operates based on hydrological routing, it can not only reflect the infiltration, storage, and evapotranspiration of GI facilities, but also quantitatively evaluate the effect of improving urban hydrological cycle by GI facilities. The applicability of the proposed model was verified by comparing the results of the proposed model with the results of SWMM. In addition, a discussion of errors occurring in the SWMM's permeable pavement system simulation is included.