• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.41 no.8
• /
• pp.537-544
• /
• 2017

The purpose of this study is to analyze the flow characteristics when a staggered hollow fiber membrane module is modeled as a porous medium. The pressure-velocity equation was used for modeling the porous medium, using pressure drop data. In terms of flow characteristics, we compared the case of the "porous medium" when the membrane module was modeled as a porous medium with the case of the "membrane module" when considering the original shape of the membrane module. The difference in pressure drop between the "porous medium" and "membrane module" was less than 0.6%. However, the maximum flow velocity and mean turbulent kinetic energy of the "porous medium" were 2.5 and 95 times larger than those of the "membrane module," respectively. Our results indicate that modeling the hollow fiber module as a porous medium is useful for predicting pressure drop, but not sufficient for predicting the maximum flow velocity and mean turbulent kinetic energy.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.288-290
• /
• 2022

As part of the SOC digitization for smart water management and flood prevention, the government reported that automatic and remote control system for drainage facilities (180 billion won) to 57% of national rivers and established a real-time monitoring system (30 billion won). In addition, they were also planning to establish a smart dam safety management system (15 billion won) based on big data at 11 regions. Therefore, research is needed for smart water management and flood prevention system that can accurately calculate the flow rate through real-time flow rate measurement of rivers. In particular, the most important thing to improve the system implementation and accuracy is to ensure the accuracy of real-time flow rate measurements. To this end, radar sensors for measuring the flow rate of electromagnetic waves in the United States and Europe have been introduced and applied to the system in Korea, but demand for improvement of the system continues due to high price range and performance. Consequently, we would like to propose an improved flow rate measurement and flood forecast system by developing a radar sensor for measuring the electromagnetic surface current meter for real-time flow rate measurement.

• Journal of IKEEE
• /
• v.27 no.4
• /
• pp.486-493
• /
• 2023

In this paper, we propose an automatic hand gesture area extraction and recognition technique using FMCW radar-based point cloud and LSTM. The proposed technique has the following originality compared to existing methods. First, unlike methods that use 2D images as input vectors such as existing range-dopplers, point cloud input vectors in the form of time series are intuitive input data that can recognize movement over time that occurs in front of the radar in the form of a coordinate system. Second, because the size of the input vector is small, the deep learning model used for recognition can also be designed lightly. The implementation process of the proposed technique is as follows. Using the distance, speed, and angle information measured by the FMCW radar, a point cloud containing x, y, z coordinate format and Doppler velocity information is utilized. For the gesture area, the hand gesture area is automatically extracted by identifying the start and end points of the gesture using the Doppler point obtained through speed information. The point cloud in the form of a time series corresponding to the viewpoint of the extracted gesture area is ultimately used for learning and recognition of the LSTM deep learning model used in this paper. To evaluate the objective reliability of the proposed technique, an experiment calculating MAE with other deep learning models and an experiment calculating recognition rate with existing techniques were performed and compared. As a result of the experiment, the MAE value of the time series point cloud input vector + LSTM deep learning model was calculated to be 0.262 and the recognition rate was 97.5%. The lower the MAE and the higher the recognition rate, the better the results, proving the efficiency of the technique proposed in this paper.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.16 no.3
• /
• pp.31-40
• /
• 2012

The irregularity of railway track affects not only the comfort of ride such as noise or vibration but also the safety of train operation. For this reason, it is an interesting research area to design a reliable and sustainable railway track system and to analyze the train movement mechanism based on systematic approaches considering reasons of track irregularity possible in a specific local environment. Irregularity data inspected by EM-120, an railway inspection system in Korea includes unavoidable incomplete and erratic information, so it is encountered lots of problem to analyse those data without appropriate pre-data-refining processes. In this research, for the efficient management and maintenance of railway system, progress rate of standard deviation of irregularity is quantified. During the computation, some important components of railways such as rail joint, ballast, roadbed, and fastener have been considered. Probabilistic distributions of irregularity growth with respect to time are computed to predict the remaining service life of railway track and to be adapted for the safety assessment.

• The Journal of Society for e-Business Studies
• /
• v.25 no.3
• /
• pp.1-13
• /
• 2020

Machine learning refers to a model generation technique that can solve specific problems from the generalization process for given data. In order to generate a high performance model, high quality training data and learning algorithms for generalization process should be prepared. As one way of improving the performance of model to be learned, the Ensemble technique generates multiple models rather than a single model, which includes bagging, boosting, and stacking learning techniques. This paper proposes a new Ensemble technique with multiple stacking that outperforms the conventional stacking technique. The learning structure of multiple stacking ensemble technique is similar to the structure of deep learning, in which each layer is composed of a combination of stacking models, and the number of layers get increased so as to minimize the misclassification rate of each layer. Through experiments using four types of datasets, we have showed that the proposed method outperforms the exiting ones.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.16 no.6 s.97
• /
• pp.615-621
• /
• 2005

In this paper, a model for simulating synthetic aperture radar(SAR) images of earth surfaces. The earth surfaces include forest area, rice crop field, other agricultural fields, grass field, road, and water surface. At first, the backscattering models are developed for bare soil surfaces, water surfaces, short vegetation fields such as rice fields and grass field, other agriculture areas, and forest areas. Then, the SAR images are generated from the digital elevation model(DEM) and digital terrain map. The DTM includes ten parameters, such as soil moisture, surface roughness, canopy height, leaf width, leaf length, leaf density, branch length, branch density, trunk length, and trunk density, if applicable. The scattering models are verified with measurements, and applied to generate an SAR image for an area.

• The Journal of the Korea Contents Association
• /
• v.18 no.7
• /
• pp.495-502
• /
• 2018

The advent of deep learning technology has made rapid progress in handwritten letter recognition in many languages. Handwritten Chinese recognition has improved to 97.2% accuracy while handwritten Japanese recognition approached 99.53% percent accuracy. Hanguel handwritten letters have many similar characters due to the characteristics of Hangeul, so it was difficult to recognize the letters because the number of data was small. In the handwritten Hanguel recognition using Hybrid Learning, it used a low layer model based on lenet and showed 96.34% accuracy in handwritten Hanguel database PE92. In this paper, 98.64% accuracy was obtained by organizing deep CNN (Convolution Neural Network) in handwritten Hangeul recognition. We designed a new network for handwritten Hangeul data based on GoogLenet without using the data augmentation or the multitasking techniques used in Hybrid learning.

• Journal of Korea Multimedia Society
• /
• v.8 no.11
• /
• pp.1483-1495
• /
• 2005

TCP/IP is the most commonly used protocol to communicate among servers, and is used in a wide range of applications. Unfortunately, Data transmission through TCP/IP places a very heavy burden on host CPUs. And it hardly makes another job to be processed. So, TOE(TCP/IP Offload Engine) is considered in many servers. But, most of TOE modules tends to not support binary compatibility for standard socket interfaces. So, it has problems that existing applications should be modified and recompiled to get advantage of TOE device. In this paper, to resolve upper problems, we suppose design and implementation of TOE module supporting binary compatibility for standard socket interfaces. Also, it can make a usage of multiple TOEs and NICs simultaneously.

• Journal of IKEEE
• /
• v.23 no.4
• /
• pp.1302-1308
• /
• 2019

This paper presents an approach for classifying image made by acquired position data from a 2D Lidar sensor with a convolutional neural network (CNN). Lidar sensor has been widely used for unmanned devices owing to advantages in term of data accuracy, robustness against geometry distortion and light variations. A CNN algorithm consists of one or more convolutional and pooling layers and has shown a satisfactory performance for image classification. In this paper, different types of CNN architectures based on training methods, Gradient Descent(GD) and Levenberg-arquardt(LM), are implemented. The LM method has two types based on the frequency of approximating Hessian matrix, one of the factors to update training parameters. Simulation results of the LM algorithms show better classification performance of the image data than that of the GD algorithm. In addition, the LM algorithm with more frequent Hessian matrix approximation shows a smaller error than the other type of LM algorithm.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.54 no.3
• /
• pp.22-28
• /
• 2017

In modern electronic warfare, a variety of devices are being operated in the fields for the purposes of communication and surveillance. Therefore, if such devices work in the same band, interference may occur and affect each other. Regarding L-band in which various devices including radar systems are operating, interference from existing devices may affect new ones in the band. In this paper, we estimate a probability distribution of the interference environment in L-band from the selected measurement data, which is fundamental for the mathematical analysis. After selecting the candidates of probability distribution, we suggest the best one from the group. The results of this study are expected to be utilized as fundamental data for the mathematical approach to the L-band interference environment.