• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.11
• /
• pp.5427-5445
• /
• 2019

While deep neural networks have achieved remarkable performance in representation learning, a huge amount of labeled training data are usually required by supervised deep models such as convolutional neural networks. In this paper, we propose a new representation learning method, namely generative adversarial networks (GAN) based bagging deep convolutional autoencoders (GAN-BDCAE), which can map data to diverse hierarchical representations in an unsupervised fashion. To boost the size of training data, to train deep model and to aggregate diverse learning machines are the three principal avenues towards increasing the capabilities of representation learning of neural networks. We focus on combining those three techniques. To this aim, we adopt GAN for realistic unlabeled sample generation and bagging deep convolutional autoencoders (BDCAE) for robust feature learning. The proposed method improves the discriminative ability of learned feature embedding for solving subsequent pattern recognition problems. We evaluate our approach on three standard benchmarks and demonstrate the superiority of the proposed method compared to traditional unsupervised learning methods.

• Journal of Korea Multimedia Society
• /
• v.21 no.8
• /
• pp.864-875
• /
• 2018

In this paper, we propose a measurement method for height of white light scanning interferometer using deep learning. In order to measure the fine surface shape, a three-dimensional surface shape measurement technique is required. A typical example is a white light scanning interferometer. In order to calculate the surface shape from the measurement image of the white light scanning interferometer, the height of each pixel must be calculated. In this paper, we propose a neural network for height calculation and use virtual data generation method to train this neural network. The accuracy was measured by inputting 57 actual data to the neural network which had completed the learning. We propose two new functions for accuracy measurement. We have analyzed the cases where there are many errors among the accuracy calculation values, and it is confirmed that there are many errors when there is no interference fringe or outside the learned range. We confirmed that the proposed neural network works correctly in most cases. We expect better results if we improve the way we generate learning data.

• Proceedings of the KIEE Conference
• /
• 2009.04a
• /
• pp.10-12
• /
• 2009

Electromagnetic Interference(EMI) in electric railway operation has become increasingly important. The components within very high power electronic, and the circuits for treating low-level signals, comprise complex system that must coexist and be highly reliable. To study it, It were included how much the HEMU-400X generates EMI and it has an effect on the power conversion units which resulted from Power Line Disturbance (PLD) phenomenon by contact loss during its running. In this study, the dynamic characteristic of a contact wire and pantograph suppling electrical power to high-speed trains are investigated. The analysis of the loss of contact based on Power Simulator program software is performed to develop power line disturbance model suitable for high speed operation. It is confirmed that a contact wire and pantograph model are necessary for studying the dynamic behavior of the pantograph system.

• Proceedings of the KSR Conference
• /
• 2005.05a
• /
• pp.522-529
• /
• 2005

The commercial service of KTX from April 1st 2004 has realized that whole country has shrunk into half-day life zone and it opened the times of speed. However, some technical problems unexpected before service have been raised and, among them, the noise generation at train passing was claimed by the residents living nearby the track and it is often publicized by the broadcastings and newspapers. In case where the residential area is close to the track, the installation of noise barrier at trackside is the general measures for noise reduction on bridge section and earthwork section. In case of KTX project, such measures were actually taken without any exception; however, the noise level is still high. In this study, analysis for the reason of higher noise level in spite of which the noise barriers are being installed was carried out, the sample of insulation material effective for noise barrier was manufactured and the performance of the insulation material was evaluated to verify its effectiveness

• Proceedings of the KSR Conference
• /
• 2010.06a
• /
• pp.133-138
• /
• 2010

This paper describes on aerodynamic characteristics of pantograph system for Next generation high speed train(HEMU-400). The pantograph which supports electric power is located on the roof. Because of this, it generate high drag, severe acoustic noise and vibration which induced unstable flow due to complex configuration. Therefore, the design of high efficient pantograph needs to increase operational speed. In this research, wind tunnel tests were performed to design a high efficient pantograph system using 1/4 scaled model which were KTX-II pantograph, single arm pantograph and periscope type pantograph with square cylinder shape panhead and optimized shape panhead. For real operational condition, flow directions were adapted by rotation of pantograph. From this results of wind tunnel, it is checked that the pantograph with optimized panhead and single arm type or periscope type has better aerodynamic performance. In addition, lift control device and spoiler in pantograph were tested to investigate the validity of application.

• 한국신재생에너지학회:학술대회논문집
• /
• 2008.10a
• /
• pp.353-356
• /
• 2008

HEV(Hybrid Electrical Vehicle) is considered as one of the next generation vehicles. To develop the HEV, there must be a reliable simulator, by which the capacities of the power resources are tested, and the parameters of the HEV are optimized before developing the real model of the HEVs. This process can save the money for designing the HEV system and improve the system without experiments. Matlab Simulink is familiar to mechanical engineers and the program can simultaneously provide a system model and a controller in one program. Nowadays, the Simdriveline toolbox which is used for analysis a power-train system is applied to build a dynamic model for a HEV system. In this study, we make a HEV simulator with the Simdriveline toolbox and develop a controller. There are two simple strategies, applied to the controller. One strategy includes a power split ratio and a shift map which are created by user. Other strategy calculated an appropriate amount of resource's torque along specific results, and this is useful when users can't develop a fitting controller. The methodologies for configuring the simulator and its control system are presented in this paper.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.19 no.6
• /
• pp.113-118
• /
• 2011

Fuel Cell Hybrid Vehicle (FCHV) is one of the most promising candidates for the next generation of transportation. It has many outstanding advantages such as higher energy efficiency and much lower emissions than internal combustion engine vehicles. It also has the ability of recovering braking energy. In order to design an FCHV drive train, we need to determine the size of the electric motor, the Fuel Cell System (FCS), and the battery. In this paper, the methodology for the sizing of these components is introduced based on the driveability constraints of the FCHV. A power management strategy is also presented because the battery energy capacity depends on it. The warm-up time of the FCS is also considered in the power management strategy and the simulation result is compared to that without considering the warm-up time.

• International Journal of Aeronautical and Space Sciences
• /
• v.9 no.1
• /
• pp.100-110
• /
• 2008

It is not easy to monitor and identify all engine faults and conditions using conventional fault detection approaches like the GPA (Gas Path Analysis) method due to the nature and complexity of the faults. This study therefore focuses on a model based diagnostic method using Neural Network algorithms proposed for fault detection on a turbo shaft engine (PW 206C) selected as the power plant for a tilt rotor type unmanned aerial vehicle (Smart UAV). The model based diagnosis should be performed by a precise performance model. However component maps for the performance model were not provided by the engine manufacturer. Therefore they were generated by a new component map generation method, namely hybrid method using system identification and genetic algorithms that identifies inversely component characteristics from limited performance deck data provided by the engine manufacturer. Performance simulations at different operating conditions were performed on the PW206C turbo shaft engine using SIMULINK. In order to train the proposed BPNN (Back Propagation Neural Network), performance data sets obtained from performance analysis results using various implanted component degradations were used. The trained NN system could reasonably detect the faulted components including the fault pattern and quantity of the study engine at various operating conditions.

• Proceedings of the KSR Conference
• /
• 2008.06a
• /
• pp.206-213
• /
• 2008

Mo-Am arch bridge is only the long-span bridge (with 125m span) in the Kyong-Bu high-speed line in service, while other bridges are PSC box girder bridges and steel composite bridges with span lengths of $25\sim50m$. However, in foreign high-speed lines, special cable-supported bridges like cable-stayed bridges and extradosed bridges are being adopted in earnest with technical specifications. The cable supported bridge is recognized as one of the indices of technology in civil engineering field, and thus it is being adpoted with a sense of rivalry in countries with advanced technology in railway engineering. In this paper, to apply the top-level cable-supported bridge technology to the domestic high-speed line up to 400km/h by establishing the technical specifications on cable-supported bridges including span length, the requirements for securing the dynamic stability and running safety of high speed train are analyzed through case studies for domestic and foreign cases.

• Proceedings of the KSR Conference
• /
• 2008.06a
• /
• pp.913-918
• /
• 2008

When developing railway vehicle system, investigation of the dynamical stability is essential as a virtual prototyping process. Not only the verification using the commercial analysis tools, systematic analysis using customized tools is also necessary, because these can give other points of view in stability, which is sometimes unable to evaluate in the former one. As a solver platform for customization, it is important to derive basic theory about flexible bodies and build flexible structure, which enables easy module insertion of user-created functions. In the paper, a flexible dynamic analysis system is developed, using absolute cartesian coordinate, modal coordinate and absolute nodal coordinate. Each coordinate system is verified by respective examples for every system. This solver system will play an important role for building the basic platform for analysis system, keeping pace with the concurrent development of the modules, such as wheel-contact force, constraints and user-defined force modules. Using the information from the analysis, the evaluation of the dynamic behavior of the train and its stability analysis will be available.