• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.77-78
• /
• 2014

A new MCMC method for optimization is presented in this paper, which is called the scanline block Gibbs sampler. Due to its slow convergence speed, traditional Markov chain Monte Carlo (MCMC) is not widely used. In contrast to the conventional MCMC method, it is more convenient to parallelize the scanline block Gibbs sampler. Since The main part of the scanline block Gibbs sampler is to calculate message between each edge, in order to accelerate the calculation of messages passing in scanline sampler, it is parallelized in GPU. It is proved that the implementation on GPU is faster than on CPU based on the experiments on the OpenGM2 benchmark.

• Journal of information and communication convergence engineering
• /
• v.8 no.1
• /
• pp.95-98
• /
• 2010

The Electro Magnetic (EM) Acceleration and Deceleration (ACC/DCC) system of the Ferromagnetic Ball(FB) is the linear motor's final structural development which can be used for devices that conserve energy, gaming or rail gun. By accelerating the FB within the coil structure, it is difficult to utilize the FB's magnetizing feature via the ACC/DCC system. There is much monopole space inside the monopole coil. By using this particular feature of the FB, starting coil and Monopole Coil Structure (MCS) can be structurally separated and another simple electric related control system can be experimented for further development. For the purpose of development a review is needed of the control system of both basic stepper motor and BLDC motor.

• Journal of applied mathematics & informatics
• /
• v.3 no.2
• /
• pp.193-204
• /
• 1996

Overflow queuing models are ofter analyzed by explicitly solving a large sparse singular linear systems arising from Kolmogorov balance equation. The system is often converted into an eigenvalue problem the dominant eigenvector of which is the desired null vector. In this paper we convert an overflow queuing problem the dominant eigenvector of which is the desired null vector. In this paper we convert an overflow queuing problem into an overflow queuing problem into an eigen-value problem into an eigen-value problem of size 1/2 of the original. Then we devise an orthogonal projector that enhances its convergence by removing unsanted eigen-components effectively. Numerical result with some suggestion is given at the end.

• Journal of information and communication convergence engineering
• /
• v.8 no.3
• /
• pp.323-327
• /
• 2010

Recently, the television market trend is change to HD television and the need of the study on HD image enhancement is increased rapidly. To enhancement of image quality, the retinex algorithm is commonly used. That's why we studied how to accelerate the retinex algorithm with CUDA on GPGPU (general purpose graphics processing unit). Calculating average part in retinex algorithm is similar to pyramidal calculation. We parallelize this recursive pyramidal average calculating for all layers, map the average data into the 2D plane and reduce the calculating time dramatically. Sequential C code takes 8948ms to get the average values for all layers in $1024{\times}1024$ image, but proposed method takes only only about 0.9ms for the same image. We are going to study about the real-time HD video rendering and image enhancement.

• Journal of Korea Multimedia Society
• /
• v.20 no.2
• /
• pp.179-187
• /
• 2017

The fourth industrial revolution is a phenomenon where productivity is improved in each field by the convergence of IT technology and existing industries. In the dental treatment process, prosthetic treatment time is drastically shortened through AI and expert software. Oral imaging, prosthesis design, and prosthesis manufacturing are performed continuously, so the treatment can be completed in a few hours. In this paper, we introduce the research trend of multimedia technology in the prosthetic process. We also propose a new method for accelerating the fusion of surface data during the optical impression. Proposed method enables high-speed optical impression and accelerates the overall automated production process of dental prosthesis.

• Proceedings of the KSME Conference
• /
• 2000.11a
• /
• pp.475-480
• /
• 2000

The transient dynamic-response analysis of fuel-storage tanks of flying vehicles accelerating in the vertical direction is achieved with finite element method. A fuel-storage tank is a representative example of the fluid-structure interaction problem, in which structure and fluid media interact strongly. For the accurate analysis of this complicated fluid-structure system, we employed ALE(arbitrary Lagrangian-Eulerian) coupling method. Two types of fuel-storage tanks, one with two baffles and the other without baffle, are considered to examine the effect of baffles. The fuel-storage tank with baffles shows more uniform hydrodynamic pressure distribution, resulting effective stress in structural region and faster convergence from transient to steady states. MSC/Dytran, a commercial FEM software for the 3D coupled dynamic analysis, is used for this analysis.

• Journal of information and communication convergence engineering
• /
• v.5 no.3
• /
• pp.239-244
• /
• 2007

In the telematics system, a reasonable path search time should be guaranteed from a great number of user's queries, even though the optimal path with minimized travel time might be continuously changed by the traffic flows. Thus, the path search method should consider traffic flows of the roads and the search time as well. However, the existing path search methods are not able to cope efficiently with the change of the traffic flows and to search rapidly paths simultaneously. This paper proposes a new path search method for fast computation. It also reflects the traffic flows efficiently. Especially, in order to simplify the computation of variable heuristic values, it employs a simplification method for estimating values of traffic-adaptable heuristics. The experiments are carried out with the $A^*$ algorithm and the proposed method in terms of the execution time, the number of node accesses and the accuracy. The results obtained from the experiments show that the method achieves very fast execution time and the reasonable accuracy as well.

• Journal of information and communication convergence engineering
• /
• v.10 no.3
• /
• pp.242-247
• /
• 2012

In the determination of suitable frame sizes associated with dynamic framed slotted Aloha used in radio frequency identification tag identification processes, the widely imposed constraint $L=2^Q$ often yields inappropriate values deviating far from the optimal values, while a straightforward use of the estimated optimal frame sizes causes frequent restarts of read procedures, both resulting in long identification delays. Taking a trade-off, in this paper, we propose a new method for determining effective frame sizes where the effective frame size changes in a multiple of a predetermined step size, namely ${\Delta}$. Through computer simulations, we show that the proposed scheme works fairly well in terms of identification delay.

• International journal of advanced smart convergence
• /
• v.8 no.2
• /
• pp.30-38
• /
• 2019

Recently, the spread of personal mobility has been rapidly increasing due to the development of environmentally friendly alternative transportation means. In addition, the level of battery technology is also rapidly developing, accelerating the popularization of personal mobility. Such personal mobility has convenience of location transfer, amusement, and high portability compared to other transportation devices. Most personal mobility, however, is made up of small wheels, which cannot overcome obstacles such as rugged roads or obstacles on the road. In this paper, to solve these problems, we tried to devise a device that can easily overcome obstacles by combining wheels with small moving means. The wheel size can be mounted on the front wheel of the small moving means in a protruding manner so that obstacles can be encountered before the front wheels and the safety and ride comfort of the running can be improved.

• KEPCO Journal on Electric Power and Energy
• /
• v.6 no.4
• /
• pp.399-404
• /
• 2020

There are many attempts to apply AI technology to diagnose facilities or improve the work efficiency of the power industry. The emergence of new machine learning technologies, such as deep learning, is accelerating the digital transformation of the power sector. The problem is that traditional power systems face security risks when adopting state-of-the-art AI systems. This adoption has convergence characteristics and reveals new cybersecurity threats and vulnerabilities to the power system. This paper deals with the security measures and implementations of the power system using machine learning. Through building a commercial facility operations forecasting system using machine learning technology utilizing power big data, this paper identifies and addresses security vulnerabilities that must compensated to protect customer information and power system safety. Furthermore, it provides security guidelines by generalizing security measures to be considered when applying AI.