• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.35C no.3
• /
• pp.79-89
• /
• 1998

The Error Back-Propagation(EBP) algorithm is widely applied to train a multi-layer perceptron, which is a neural network model frequently used to solve complex problems such as pattern recognition, adaptive control, and global optimization. However, the EBP is basically a gradient descent method, which may get stuck in a local minimum, leading to failure in finding the globally optimal solution. Moreover, a multi-layer perceptron suffers from locking a systematic determination of the network structure appropriate for a given problem. It is usually the case to determine the number of hidden nodes by trial and error. In this paper, we propose a new algorithm to efficiently train a multi-layer perceptron. OUr algorithm uses stochastic perturbation in the weight space to effectively escape from local minima in multi-layer perceptron learning. Stochastic perturbation probabilistically re-initializes weights associated with hidden nodes to escape a local minimum if the probabilistically re-initializes weights associated with hidden nodes to escape a local minimum if the EGP learning gets stuck to it. Addition of new hidden nodes also can be viewed asa special case of stochastic perturbation. Using stochastic perturbation we can solve the local minima problem and the network structure design in a unified way. The results of our experiments with several benchmark test problems including theparity problem, the two-spirals problem, andthe credit-screening data show that our algorithm is very efficient.

• Journal of the Korea Safety Management & Science
• /
• v.16 no.2
• /
• pp.1-9
• /
• 2014

This study tried to propose plan to prevent human error of railroad driver among human error of railroad worker which takes great share in railroad accident. For this, in order to maintain correlation between the accident actually occurred after the opening of high-speed railroad and experience of accident that did not happened, survey on respondent was analyzed by conducting survey on KTX captain who is working in driving work of high-speed railroad, and instruction management team manager who manages KTX captain and captain. This thesis classified the factors by human factor, job factor, environment factor, organization factor, and established human error management model by comparing and analyzing how each factors have spatial interrelations with a railroad accident. The purpose of this study is to contribute to make safe railroad, and reliable railroad by preventing human error accident by minimizing human error of high-speed railroad drivers, and improving driving workers to cope accurately and fast with irregularities through various institutional improvement, improvement of driving facilities, improvement of operating room environment, and improvement of education system.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.11
• /
• pp.2176-2181
• /
• 2011

This paper estimates the life cycle cost(LCC) of a propulsion control system using the maintenance information in the high speed train(KTX-1). Life cycle costing is one of the most effective approaches for the cost analysis of long-life systems such as the KTX-1. Until now, most life cycle cost of the system has been studied as a whole system viewpoint. But in case of railway industry, LCC studies are needed on the subsystem like a propulsion control system because subsystems are developed continuously localization. This paper proposes the life cycle cost model which fitted to estimate life cycle cost (LCC) using maintenance information manual. As a result, LCC on propulsion control system increased moderately expect for periodical time when major parts are replaced at the same time. Results will be reflected in the development of domestic products.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.1
• /
• pp.451-459
• /
• 2018

This work presents an approach for modeling of electric vehicle considering the vehicle dynamics, drive train, rotational wheel and load dynamics. The system is composed of IPMSM (Interior Permanent Magnet Synchronous Motor) coupled with the wheels through a drive train. Generally, IPMSM is controlled by ordinary PID controllers. Performance of the ordinary PID controller is not satisfactory owing to the difficulties of optimal gain selections. To overcome this problem, a new type of fuzzy logic gain tuner for PID controllers of IPMSM is required. Therefore, in this paper fuzzy logic based gain tuning method for PID controller is proposed and compared with some previous control techniques for the better performance of electric vehicle with an optimal balance of acceleration, speed, travelling range, improved controller quality and response. The model was developed in MATLAB/Simulink, simulations were carried out and results were observed. The simulation results have proved that the proposed control system works well to remove the transient oscillations and assure better system response in all conditions.

• Proceedings of the KSR Conference
• /
• 2009.05b
• /
• pp.225-230
• /
• 2009

From the view point of railway vehicle dynamics, the interaction between wheel and rail have an huge effect on the behavior of the vehicle. This phenomenon is an unique motion, only for railway vehicles. Furthermore, close investigation of the backgrounds of the interaction is the key to estimate the dynamic behavior of the vehicle, successfully. To evaluate the model including flexible bodies such as car body and catenary system of the next generation express train, it is necessary to develop proper dynamic solver including a wheel rail contact module. In this study, wheel-rail contact module is developed using the general purpose dynamic solver. First of all, the procedure for calculation of the wheel-rail contact force has been established. Generally, yaw angle of the wheelset is ignored. Sets of information are summarized as tables and splined for further uses. With this information, normal force and creep coefficient can be extracted and used for FASTSIM algorithm, which has been shown good reliability over years. Normal force and longitudinal, lateral force at the contact surface are also calculated. Those data are verified by commercial railway simulation program 'VAMPIRE'. This procedure and program can offer a basic process for estimation of the dynamic behavior and wear of the wheel-rail system, even while running on the curved rail. Finally, multi-dimensional inspection tool will be developed including the prediction of the derailment.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.68 no.1
• /
• pp.199-206
• /
• 2019

Recently, an expansion of safety facilities has been widely applied to effectively manage the safety of train operation due to increase of the high-speed section of the general railway and the introduction of high-speed train. Accordingly, performance improvement, upgrading and high reliability of existing safety devices are required. the dragging detector, one of the safety devices, is an analogue system that consists of closed circuit with an electric current flows and operates when the closed circuit is opened by the impact of the dragging object. Such method has unreasonable problem that should be replaced after being detected. It is need to replace with an automatic return type dragging detector which is easy to maintain. In addition, it is necessary to develop a dragging detector that applicable to general railway and urban railway including high-speed railway, in accordance with the speeding up and densification of trains, although it is currently applied only to high-speed railway. In this paper, we propose an automatic return type dragging detector which has versatility and excellent maintainability with digital sensor and real time monitoring.

• Journal of Korean Port Research
• /
• v.13 no.1
• /
• pp.101-112
• /
• 1999

Traffic accidents increase with the increase of the vehicles in operation on the street. Especially big traffic accidents composed of over 3 killed or 20 injured accidents with the property damage become one of the serious problems to be solved in most of the cities. The purpose of this study is to build the discrimination model on big traffic accidents using the Quantification II theory for establishing the countermeasures to reduce the big traffic accidents. The results are summarized as follows. 1)The existing traffic accident related model could not explain the phenomena of the current traffic accident appropriately. 2) Based on the big traffic accident types vehicle-vehicle, vehicle-alone, vehicle-pedestrian and vehicle-train accident rates 73%, 20.5% 5.6% and two cases respectively. Based on the law violation types safety driving non-fulfillment center line invasion excess speed and signal disobedience were 48.8%, 38.1% 2.8% and 2.8% respectively. 3) Based on the law violation types major factors in big traffic accidents were road and environment, human, and vehicle in order. Those factors were vehicle, road and environment, and human in order based on types of injured driver’s death. 4) Based on the law violation types total hitting and correlation rates of the model were 53.57% and 0.97853. Based on the types of injured driver’s death total hitting and correlation rates of the model were also 71.4% and 0.59583.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.4
• /
• pp.1489-1503
• /
• 2018

Spectrum reuse has attracted much concern of researchers and scientists, however, the dynamic spectrum access is challenging, since an individual secondary user usually just has limited sensing abilities. One key insight is that spectrum usage forecasting among secondary users, this inspiration enables users to obtain more informed spectrum opportunities. Therefore, spectrum usage forecasting is vital to cognitive radio networks (CRNs). With this insight, a spectrum usage forecasting model for the occurrence of primary users prediction is derived in this paper. The proposed model is based on auto regressive enhanced primary user emergence reasoning (AR-PUER), which combines linear prediction and primary user emergence reasoning. Historical samples are selected to train the spectrum usage forecasting model in order to capture the current distinction pattern of primary users. The proposed scheme does not require the knowledge of signal or of noise power. To verify the performance of proposed spectrum usage forecasting model, we apply it to the data during the past two months, and then compare it with some other sensing techniques. The simulation results demonstrate that the spectrum usage forecasting model is effective and generates the most accurate prediction of primary users occasion in several cases.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.56 no.3
• /
• pp.632-640
• /
• 2007

The low force estimation method of skeletal muscle was proposed by using ICA(independent component analysis) and neuro-transmission model. An EMG decomposition is the procedure by which the signal is classified into its constituent MUAP(motor unit action potential). The force index of electromyography was due to the generation of MUAP. To estimate low force, current analysis technique, such as RMS(root mean square) and MAV(mean absolute value), have not been shown to provide direct measures of the number and timing of motoneurons firing or their firing frequencies, but are used due to lack of other options. In this paper, the method based on ICA and chemical signal transmission mechanism from neuron to muscle was proposed. The force generation model consists of two linear, first-order low pass filters separated by a static non-linearity. The model takes a modulated IPI(inter pulse interval) as input and produces isometric force as output. Both the step and random train were applied to the neuro-transmission model. As a results, the ICA has shown remarkable enhancement by finding a hidden MAUP from the original superimposed EMG signal and estimating accurate IPI. And the proposed estimation technique shows good agreements with the low force measured comparing with RMS and MAV method to the input patterns.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.7
• /
• pp.2116-2124
• /
• 2000

In this paper we propose combustion control of refuse incineration plant using fuzzy model and genetic algorithm. At first fuzzy modelling is performed to obtain fuzzy model of the refuse incineration plant and obtained fuzzy model predicts outputs of the plant when inputs are given. Fuzzy model ca be used to obtain control strategy, and train and enhance operators' skill by simulating the plant. Then genetic algorithms search and find out optimal control inputs over all possible solutions in respect to desired outputs and these are inserted to plant. In order to testify proposed control method, computer simulation was carried out. As a result, ISE of fuzzy model of refuse incineration plant is 0.015 and ITAE of control by proposed method, 352 which is better than that by manual operation.