• Transactions of the Korean Society of Automotive Engineers
• /
• v.13 no.2
• /
• pp.72-82
• /
• 2005

The objective of this paper is to describe the optimization of design parameters in a large-sized commercial bus heater system by using CFD(computational fluid dynamics) analysis and Taguchi method. In order to obtain the best combination of each control factor which results in a desired performance of heater system, the parameter design of the Taguchi method is adopted for the robust design considering the dynamic characteristic. The research activity may be divided into four phases. The first one is analyzing the problem, i.e., ascertaining the influential factors. In the second phase the levels were set in such a way that their variation would significantly influence the response. In the third phase the experimental runs were designed. In the final phase the planned runs were carried out numerically to evaluate the optimal combination of factors which is able to provide the best response. In this study, eight factors were considered for the analysis: one with two level and seven with three level combinations comprising the $L_{18}(2^1{\times}3^7)$ orthogonal array. The results of this study can be summarized as follows ; (i)The optimum condition of control factor is a set of <$A_2\;B_1\;C_3\;D_3\;E_1\;F_2\;G_3\;H_2$> where A is shape of the outer fin, B is pitch of the outer fin, C is height of the outer fin, D is the inner fin number, E is the inner fin height, F is length of the flame guide, G is diameter of the heating element and H is clearance between air guide and heating element. (ii)The heat capacity of heated discharge air under the optimum condition satisfies the equation y=0.6M w here M is a signal factor. (iii)The warm-up rate improves about three times, more largely as com pared with the current condition, which results in about 9.2minutes reduction.

• Investigative Magnetic Resonance Imaging
• /
• v.20 no.4
• /
• pp.215-223
• /
• 2016

Purpose: The management of metal-induced field inhomogeneities is one of the major concerns of distortion-free magnetic resonance images near metallic implants. The recently proposed method called "Slice Encoding for Metal Artifact Correction (SEMAC)" is an effective spin echo pulse sequence of magnetic resonance imaging (MRI) near metallic implants. However, as SEMAC uses the noisy resolved data elements, SEMAC images can have a major problem for improving the signal-to-noise ratio (SNR) without compromising the correction of metal artifacts. To address that issue, this paper presents a novel reconstruction technique for providing an improvement of the SNR in SEMAC images without sacrificing the correction of metal artifacts. Materials and Methods: Low-rank approximation in each coil image is first performed to suppress the noise in the slice direction, because the signal is highly correlated between SEMAC-encoded slices. Secondly, SEMAC images are reconstructed by the best linear unbiased estimator (BLUE), also known as Gauss-Markov or weighted least squares. Noise levels and correlation in the receiver channels are considered for the sake of SNR optimization. To this end, since distorted excitation profiles are sparse, $l_1$ minimization performs well in recovering the sparse distorted excitation profiles and the sparse modeling of our approach offers excellent correction of metal-induced distortions. Results: Three images reconstructed using SEMAC, SEMAC with the conventional two-step noise reduction, and the proposed image denoising for metal MRI exploiting sparsity and low rank approximation algorithm were compared. The proposed algorithm outperformed two methods and produced 119% SNR better than SEMAC and 89% SNR better than SEMAC with the conventional two-step noise reduction. Conclusion: We successfully demonstrated that the proposed, novel algorithm for SEMAC, if compared with conventional de-noising methods, substantially improves SNR and reduces artifacts.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.38 no.3
• /
• pp.39-48
• /
• 2015

In this study, granting the optimal loan limit on SME (Small and Medium Enterprise) loans of financial institutions was proposed using the traditional newsvendor model. This study was the first domestic case study that applied the newsvendor model that was mainly used to calculate the optimum order quantity under some uncertain demands to the calculation of the loan limit (debt ceiling) of institutions. The method presented in this study made it possible to calculate the loan limit (debt ceiling) to maximize the revenue of a financial institution using probability functions, applied the newsvendor model setting the order volume of merchandise goods as the loan product order volume of the financial institution, and proposed, through the analysis of empirical data, the availability of additional loan to the borrower and the reduction of the debt ceiling and a management method for the recovery of the borrower who could not generate profit. In addition, the profit based loan money management model presented in this study also demonstrated that it also contributed to some extent to the prediction of the bankruptcy of the borrowing SME (Small and Medium Enterprise), as well as the calculation of the loan limit based on profit, by deriving the result values that the borrowing SME (Small and Medium Enterprise) actually went through bankruptcy at later times once the model had generated a signal of loan recovery for them during the validation of empirical data. accordingly, The method presented in this study suggested a methodology to generated a signal of loan recovery to reduce the losses by the bankruptcy.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.14 no.1
• /
• pp.109-116
• /
• 2014

This paper presents a high-efficiency digital class-D audio amplifier using a composite interpolation filter for portable audio devices. The proposed audio amplifier is composed of an interpolation filter, a delta-sigma modulator, and a class-D output stage. To reduce power consumption, the designed interpolation filter has an optimized composite structure that uses a direct-form symmetric and Lagrange FIR filters. Compared to the filters with homogeneous structures, the hardware cost and complexity are reduced by about half by the optimization. The coefficients of the digital delta-sigma modulator are also optimized for low power consumption. The class-D output stage has gate driver circuits to reduce shoot-through current. The implemented class-D audio amplifier exhibited a high efficiency of 87.8 % with an output power of 57 mW at a load impedance of $16{\Omega}$ and a power supply voltage of 1.8 V. An outstanding signal-to-noise ratio of 90 dB and a total harmonic distortion plus noise of 0.03 % are achieved for a single-tone input signal with a frequency of 1 kHz.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.2
• /
• pp.285-297
• /
• 2015

In this paper, Ulta-WideBand (UWB) channel measurement and modeling based on compressive sampling (CS) are proposed. The sparsity of the channel impulse response (CIR) of the UWB signal in frequency domain enables the proposed channel measurement to have a low-complexity and to provide a comparable performance compared with the existing approaches especially used for the indoor geo-localization purpose. Furthermore, to improve the performance under noisy situation, the soft thresholding method is also investigated in solving the optimization problem for signal recovery of CS. Via numerical results, the proposed channel measurement and modeling are evaluated with the real measured data in terms of location estimation error, bandwidth, and compression ratio for indoor geo-localization using UWB system.

• The KIPS Transactions:PartB
• /
• v.11B no.2
• /
• pp.207-212
• /
• 2004

This paper presents a reinforcement learning framework for stock trading systems. Trading system parameters are optimized by Q-learning algorithm and neural networks are adopted for value approximation. In this framework, cooperative multiple agents are used to efficiently integrate global trend prediction and local trading strategy for obtaining better trading performance. Agents Communicate With Others Sharing training episodes and learned policies, while keeping the overall scheme of conventional Q-learning. Experimental results on KOSPI 200 show that a trading system based on the proposed framework outperforms the market average and makes appreciable profits. Furthermore, in view of risk management, the system is superior to a system trained by supervised learning.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.11
• /
• pp.2576-2583
• /
• 2015

In a full-duplex (FD) wireless cellular network, uplink (UL) users induce the severe inter-user interference to downlink (DL) users. Therefore, a user scheduling that makes a pair of DL user and UL user to use the same radio resource simultaneously influences the system performances significantly. In this paper, we first formulate an optimization problem for user scheduling to minimize the occurrence of outage, aiming to guarantee the quality of service of users, and then we propose a suboptimal user scheduling algorithm with low complexity. The proposed scheduling algorithm is designed in a way where the DL user with a worse signal quality has a higher priority to choose its UL user that causes less interference. Simulation results show that the FD system using the proposed user scheduling algorithm achieves the optimal performance and significantly decreases the outage probability compared with the conventional half-duplex cellular system.

• Journal of Digital Convergence
• /
• v.17 no.10
• /
• pp.251-258
• /
• 2019

In the future, various products are created in various fields using artificial intelligence. In this age, it is a very important problem to know the operation principle of artificial intelligence learning method and to use it correctly. This paper introduces artificial intelligence learning methods that have been known so far. Learning of artificial intelligence is based on the fixed point iteration method of mathematics. The GD(Gradient Descent) method, which adjusts the convergence speed based on the fixed point iteration method, the Momentum method to summate the amount of gradient, and finally, the Adam method that mixed these methods. This paper describes the advantages and disadvantages of each method. In particularly, the Adam method having adaptivity controls learning ability of machine learning. And we analyze how these methods affect digital signals. The changes in the learning process of digital signals are the basis of accurate application and accurate judgment in the future work and research using artificial intelligence.

• Journal of Digital Convergence
• /
• v.19 no.10
• /
• pp.253-263
• /
• 2021

Machine learning is the process of constructing a cost function using learning data used for learning and an artificial neural network to predict the data, and finding parameters that minimize the cost function. Parameters are changed by using the gradient-based method of the cost function. The more complex the digital signal and the more complex the problem to be learned, the more complex and deeper the structure of the artificial neural network. Such a complex and deep neural network structure can cause over-fitting problems. In order to avoid over-fitting, a weight decay regularization method of parameters is used. We additionally use the value of the cost function in this method. In this way, the accuracy of machine learning is improved, and the superiority is confirmed through numerical experiments. These results derive accurate values for a wide range of artificial intelligence data through machine learning.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.10
• /
• pp.49-55
• /
• 2019

In this paper, we propose a simple precoding vector design scheme for multi-user multiple-input single-output (MISO) interference channel when there are multiple eavesdroppers. We aim to obtain a mathematical closed-form solution of the secrecy rate optimization problem. For this goal, we design the precoding vector based on the signal-to-leakage plus noise ratio (SLNR). More specifically, the proposed precoding vector is designed to completely eliminate a wiretap channel capacity for refraining the eavesdroppers from detecting the transmitted information, and to maximize the transmitter-receiver link achievable rate. We performed simulation for the performance investigation. Simulation results show that the proposed scheme has better secrecy rate than the conventional scheme over all signal-to-noise ratio (SNR) range even though the special condition among the numbers of transmit antennas, transmitter-receiver links, and eavesdroppers is not satisfied.