• Journal of information and communication convergence engineering
• /
• v.13 no.2
• /
• pp.123-131
• /
• 2015

Training neural networks is a complex task with great importance in the field of supervised learning. In the training process, a set of input-output patterns is repeated to an artificial neural network (ANN). From those patterns weights of all the interconnections between neurons are adjusted until the specified input yields the desired output. In this paper, a new hybrid algorithm is proposed for global optimization of connection weights in an ANN. Dynamic swarms are shown to converge rapidly during the initial stages of a global search, but around the global optimum, the search process becomes very slow. In contrast, the gradient descent method can achieve faster convergence speed around the global optimum, and at the same time, the convergence accuracy can be relatively high. Therefore, the proposed hybrid algorithm combines the dynamic particle swarm optimization (DPSO) algorithm with the backpropagation (BP) algorithm, also referred to as the DPSO-BP algorithm, to train the weights of an ANN. In this paper, we intend to show the superiority (time performance and quality of solution) of the proposed hybrid algorithm (DPSO-BP) over other more standard algorithms in neural network training. The algorithms are compared using two different datasets, and the results are simulated.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.28 no.1
• /
• pp.79-87
• /
• 2014

With the increase of demand for electricity power, new construction and expansion of transmission lines for transport have been required. However, it has been difficult to be realized by such opposition from environmental groups and residents. Therefore, the development of techniques for effective use of existing transmission lines is more needed. In this paper, the major variables to affect the allowable transmission capacity in an overhead transmission lines were selected and the dynamic line rating (DLR) method using artificial neural networks reflecting unique environment-heat properties was proposed. To prove the proposed method, the analyzed results using the artificial neural network were compared with the ones obtained from the existing method. The analyzed results using the proposed method showed an error of 0.9% within ${\pm}$, which was to be practicable.

• Journal of the Institute of Convergence Signal Processing
• /
• v.13 no.4
• /
• pp.233-237
• /
• 2012

In this paper, We propose a control system which compensate a output of a main Neual Network using a RTDNN(Recurrent Time Delayed Neural Network) with a FLC(Fuzzy Logic Controller)After a learn of main neural network, it can occur a Over shoot or Under shoot from a disturbance or a load variations. In order to adjust above case, we used the fuzzy compensator to get an expected results. And the weight of main neural network can be changed with the result of learning a inverse model neural network of plant, so a expected dynamic characteristics of plant can be got. We can confirm good response characteristics of proposed neural network controller by the results of simulation.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.2
• /
• pp.259-267
• /
• 1997

The advent of B-ISDN using ATM(asynchronous transfer mode) made possible a variety of new multimedia services, however it also created a problem of congestion control due to bursty nature of various traffic sources. To tackle this problem, UPC/NPC(user parameter control/network parameter control) have been actively studied and DRLB(dynamic rate leaky bucket) algorithm, in which the token generation rate is changed according to states of data source andbuffer occupancy, is a good example of the UPC/NPC. However, the DRLB algorithm has drawbacks of low efficiency and difficult real-time implementation for bursty traffic sources because the determination of token generation rate in the algorithm is based on the present state of network. In this paper, we propose a more plastic and effective congestion control algorithm by combining the DRLB algorithm and neural network based prediction to remedy the drawbacks of the DRLB algorithm, and verify the efficacy of the proposed method by computer simulations.

• Journal of information and communication convergence engineering
• /
• v.7 no.2
• /
• pp.125-130
• /
• 2009

A robust friction compensation scheme is proposed in this paper. The recurrent fuzzy neural network and friction parameter observer are developed with sliding mode based controller in order to obtain precise position tracking performance. For a servo system with incomplete identified friction parameters, a proposed control scheme provides a satisfactory result via some experiment.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.253-253
• /
• 2000

Each robot plays a role of its own behavior in dynamic robot-soccer environment. One of the most necessary conditions to win a game is control of robot movement. In this paper we suggest a win strategy using Cellular Neural Network to set optimal path and cooperative behavior, which divides a soccer ground into grid-cell based ground and has robots move a next grid-cell along the optimal path to approach the moving target.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.6
• /
• pp.809-817
• /
• 2020

In the design and development process of Small-Sat power distribution and transmission module, the stability of dynamic resources was evaluated by a deep learning algorithm. The requirements for the stability evaluation consisted of the power distribution function of the power distribution module and demand module to the SAR radar in Small-Sat. To verify the performance of the switching power components constituting the power module PDM, the reliability was verified using a dynamic neural network. The adoption material of deep learning for reliability verification is the power distribution function of the payload to the power supplied from the small satellite main body. Modeling targets for verifying the performance of this function are output voltage (slew rate control), voltage error, and load power characteristics. First, to this end, the Coefficient Structure area was defined by modeling, and PCB modules were fabricated to compare stability and reliability. Second, Levenberg-Marquare based Two-Way NARX neural network Sigmoid Transfer was used as a deep learning algorithm.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.368-368
• /
• 2000

In this paper, we present forecasting ozone concentration with decision support system. Since the mechanism of ozone concentration is highly complex, nonlinear, and nonstationary, modeling of ozone prediction system has many problems and results of prediction are not good performance so far. Forecasting ozone concentration with decision support system is acquired to information from human knowledge and experiment data. Fuzzy clustering method uses the acquisition and dynamic polynomial neural network gives us a good performance for ozone prediction with ability of superior data approximation and self-organization.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1996.10a
• /
• pp.227-233
• /
• 1996

The Inverted Pendulum has been one of most popular nonlinear dynamic systems for the exploration of control techniques. This paper presents a new linear optimal control techniques and nonlinear neural network learning methods. The multiayered neural networks are used to add nonlinear effects on the linear optimal regulator(LQR). The new regulator can compensate nonlinear system uncertainties that are not considered in the LQR design, and can tolerated a wider range of uncertainties than the LQR alone. The new regulator has two neural networks for modeling and control. The neural network for modeling is used to obtain a more accurate model than the given mathematical equations. The neural network for control is used to overcome deficiencies by adding corrections to the linear coefficients of the LQR and by adding nonlinear effects on the LQR. Computer simulations are performed to show the applicability and a more robust regulator than the LQR alone.

• Journal of the Korean Society for Precision Engineering
• /
• v.16 no.6
• /
• pp.133-140
• /
• 1999

In this study, a monitoring method to detect chip disposal state in drilling system based on neural network was proposed and its performance was evaluated. If chip flow is bad during drilling, not only the static component but also the fluctuation of dynamic component of drilling. Drilling torque is indirectly measured by sensing spindle motor power through a AC spindle motor drive system. Spindle motor power being measured drilling, four quantities such as variance/mean, mean absolute deviation, gradient, event count were calculated as feature vectors and then presented to the neural network to make a decision on chip disposal state. The selected features are sensitive to the change of chip disposal state but comparatively insensitive to the change of drilling condition. The 3 layerd neural network with error back propagation algorithm has been used. Experimental results show that the proposed monitoring system can successfully recognize the chip disposal state over a wide range of drilling condition even though it is trained under a certain drilling condition.