• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.41 no.4
• /
• pp.417-428
• /
• 2021

Since future maintenance cost estimation of infrastructure involves uncertainty, it is important to make use of a failure prediction model. However, it is difficult for local governments to develop accurate failure prediction models applicable to infrastructure due to a lack of budget and expertise. Therefore, this study estimated the life expectancy of asphalt road pavement of national highways using the Bayesian Markov Mixture Hazard model. In addition, in order to accurately estimate life expectancy, environmental variables such as traffic volume, ESAL (Equivalent Single Axle Loads), SNP (Structural Number of Pavement), meteorological conditions, and de-icing material usage were applied to retain reliability of the estimation results. As a result, life expectancy was estimated from at least 13.09 to 19.61 years by region. By using this approach, it is expected that it will be possible to estimate future maintenance cost considering local failure characteristics.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.15 no.5
• /
• pp.3-11
• /
• 2005

This paper presents a performance analysis model based on an M/M/1 queue and Poisson distribution of input data traffic. The simulation on a pipelined AES system with processing rate of 10 rounds per clock shows $4.0\%$ higher performance than a non-pipelined version consuming 10 clocks per transaction. Physical implementation of pipelined AES with FPGA takes 3.5 times bigger gate counts than the non-pipelined version whereas the pipelined version yields only $3.5\%$ performance enhancement. The proposed analysis model can be used to optimize cost-performance of AES hardware designs.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.6
• /
• pp.1053-1065
• /
• 2020

With the recent development of hardware performance and artificial intelligence technology, sophisticated fake videos that are difficult to distinguish with the human's eye are increasing. Face synthesis technology using artificial intelligence is called Deepfake, and anyone with a little programming skill and deep learning knowledge can produce sophisticated fake videos using Deepfake. A number of indiscriminate fake videos has been increased significantly, which may lead to problems such as privacy violations, fake news and fraud. Therefore, it is necessary to detect fake video clips that cannot be discriminated by a human eyes. Thus, in this paper, we propose a deep-fake detection model applied with Bidirectional Convolution LSTM and Attention Module. Unlike LSTM, which considers only the forward sequential procedure, the model proposed in this paper uses the reverse order procedure. The Attention Module is used with a Convolutional neural network model to use the characteristics of each frame for extraction. Experiments have shown that the model proposed has 93.5% accuracy and AUC is up to 50% higher than the results of pre-existing studies.

• Composites Research
• /
• v.34 no.6
• /
• pp.426-433
• /
• 2021

The carbon fiber reinforced plastic manufacturing process has a problem in that a dimensional error occurs due to thermal deformation such as residual stress, spring-in, and warpage. The main causes of thermal deformation are various, including the shape of the product, the chemical shrinkage, thermal expansion of the resin, and the mold effect according to the material and surface condition of the mold. In this study, a viscoelastic model was applied to the plate model to predict the thermal deformation. The effects of chemical shrinkage and thermal expansion of the resin, which are the main causes of thermal deformation, were analyzed, and the analysis technique of the 3-D viscoelastic model with and without mold was also studied. Then, the L-shaped mold effect was analyzed using the verified 3D viscoelastic model analysis technique. The results show that different residual deformation occurs depending on the surface condition even when the same mold is used.

• Journal of Platform Technology
• /
• v.11 no.3
• /
• pp.3-11
• /
• 2023

Real-time control systems are growing rapidly as infrastructure technologies such as IoT and mobile communication develop and services that value real-time such as factory management and vehicle operation checks increase. Various solutions have been proposed to increase the time sensitivity of this system, but most real-time control systems are currently composed of local servers and multiple clients located in control stations, which are transmitted to local servers where control systems are located. In this paper, we proposed an edge computing-based real-time control model that can reduce the time it takes for the bus information system, one of the real-time control systems, to provide the information to the user at the time it collects the information. Simulating the existing model and the edge computing model, the edge computing model confirmed that the cost for users to receive data is reduced from at least 10% to up to 80% compared to the existing model.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.5
• /
• pp.307-314
• /
• 2023

As supercomputing and hardware technology advances, climate prediction models are improving. The Korean Meteorological Administration adopted GloSea5 from the UK Met Office and now operates an updated GloSea6 tailored to Korean weather. Universities and research institutions use Low-GloSea6 on smaller servers, improving accessibility and research efficiency. In this paper, profiling Low-GloSea6 on smaller servers identified the tri_sor_dp_dp subroutine in the tri_sor.F90 atmospheric model as a CPU-intensive hotspot. Applying linear regression, a type of machine learning, to this function showed promise. After removing outliers, the linear regression model achieved an RMSE of 2.7665e-08 and an MAE of 1.4958e-08, outperforming Lasso and ElasticNet regression methods. This suggests the potential for machine learning in optimizing identified hotspots during Low-GloSea6 execution.

• Journal of the Society of Disaster Information
• /
• v.19 no.4
• /
• pp.968-975
• /
• 2023

Purpose: We would like to confirm that the false positive rate of flames/smoke is high when detecting fires. Propose a method and dataset to recognize and classify fire situations to reduce the false detection rate. Method: Using the video as learning data, the characteristics of the fire situation were extracted and applied to the classification model. For evaluation, the model performance of Yolov8 and Slowfast were compared and analyzed using the fire dataset conducted by the National Information Society Agency (NIA). Result: YOLO's detection performance varies sensitively depending on the influence of the background, and it was unable to properly detect fires even when the fire scale was too large or too small. Since SlowFast learns the time axis of the video, we confirmed that detects fire excellently even in situations where the shape of an atypical object cannot be clearly inferred because the surrounding area is blurry or bright. Conclusion: It was confirmed that the fire detection rate was more appropriate when using a video-based artificial intelligence detection model rather than using image data.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.4
• /
• pp.39-46
• /
• 2024

Many researchers make efforts to evaluate water quality using various models. Such models require a dataset without missing values, but in real world, most datasets include missing values for various reasons. Simple deletion of samples having missing value(s) could distort distribution of the underlying data and pose a significant risk of biasing the model's inference when the missing mechanism is not MCAR. In this study, to explore the most appropriate technique for handing missing values in water quality data, several imputation techniques were experimented based on existing KNN and MICE imputation with/without the generative neural network model, Autoencoder(AE) and Denoising Autoencoder(DAE). The results shows that KNN and MICE combined imputation without generative networks provides the closest estimated values to the true values. When evaluating binary classification models based on support vector machine and ensemble algorithms after applying the combined imputation technique to the observed water quality dataset with missing values, it shows better performance in terms of Accuracy, F1 score, RoC-AuC score and MCC compared to those evaluated after deleting samples having missing values.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.34 no.10
• /
• pp.1427-1435
• /
• 2010

A mixed finite element model was developed using the classical plate theory to analyze the nonlinear bending of a plate. The appropriate weight functions for the constraints integrated over the domain were determined by the Lagrange multiplier method by using the principle of minimum virtual energy; which provides the constitutive relations between force-like variables and strains. All of detail terms of element wise coefficient matrices and associate tangent matrices to be used in the Newton iterative method are presented. Then, the linear solutions of the current model and those of the traditional displacement model under the SS (simple support) boundary conditions were compared with the existing analytical solution. The post-processed images of the nonlinear results of the force-like variables are presented to show the continuity of the solutions at the joint of the element boundaries. Finally, the converged nonlinear finite element solutions of the current model are compared with those of existing traditional displacement model.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.4
• /
• pp.297-303
• /
• 2021

When sound waves propagate over long distances in shallow water, measured transmission loss is greater than predicted one using underwater acoustic model with the Rayleigh reflection model due to inhomogeneity of the bottom. Accordingly, the US Navy predicts sound wave propagation by applying the empirical formula-based High Frequency Bottom Loss (HFBL) model. In this study, the measurement and analysis of transmission loss was conducted using mid-frequency (2.3 kHz, 3 kHz) in the shallow water of the East Sea in summer. BELLHOP eigenray tracing output shows that only sound waves with lower grazing angle than the critical angle propagate long distances for several kilometers or more, and the difference between the predicted transmission loss based on the Rayleigh reflection model and the measured transmission loss tend to increase along the propagation range. By comparing the Rayleigh reflection model and the HFBL model at the high grazing angle region, the bottom province, the input value of the HFBL model, is estimated and BELLHOP transmission loss with HFBL model is compared to measured transmission loss. As a result, it agrees well with the measurements of transmission loss.