• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.08a
• /
• pp.93-96
• /
• 2000

본 논문에서는 신경망제어기의 출력을 보상하는 퍼지보상기를 갖는 신경망제어기에 관하여 제안하였다. 학습이 완료된 신경망제어기를 사용하더라도 예상치 못한 외란으로 인해 플랜트의 출력이 좋지 못한 경우가 있는데, 이것을 적절하게 조절해 주기 위해 퍼지보상기를 사용하여 원하는 결과를 얻을 수 있도록 하였다. 그리고, 플랜트의 동적 특성을 계속해서 학습할 수 있도록 시간이 경과함에 따라 신경망제어기의 성능이 향상되도록 하였다. 이것을 확인하기 위해서, 2차 플랜트에 적용하여 제안한 제어기의 성능을 확인하였다.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.14 no.3
• /
• pp.49-58
• /
• 2000

This paper deals with a systematic approach to neural network controller design for static VAR compensator (SVC) using a learning algorithm of error back propagation that accepts error and change of error as inputs, the momentum learning technique is used for reduction of learning time, to improve system stability. A SVC, one of the Flexible AC Transmission System(FACTS), constructed by a fixed capacitor(FC) and a thyristor controlled reactor(TCR), is designed and implemented to improve the damping of a synchronous generator, as well as controlling the system voltage.TO verify the robustness of the proposed method, we considered the dynamic response of generator rotor angle deviation, angular velocity deviation and generator terminal voltage by applying a power fluctuation and rotor angle fluctuation in initial point when heavy load and normal load. Thus, we prove the usefulness of proposed method to improve the stability of single machine-infinite bus with SVC system.

• Proceedings of the KIEE Conference
• /
• 1997.07b
• /
• pp.606-608
• /
• 1997

In this paper, Neuro-Fuzzy Controller(NFC), a fuzzy system realized using a neural network, is to adopt for the multivariable system. In the multivariable system, the interactive effects between the variables should be taken into account. A simple compensator, using the steady-state information can be obtained for open-loop stable systems, is presented to cope with this problem. However, it should be supposed that the plant is unknown to the control system designer, but an estimate of the DC gain has been obtained by carrying out experiments on the plant. Also, if the variables are not combinated completely, it is difficult to design the controller. Therefore, we design a neuro-fuzzy controller which controls a multivariable system with only input output informations, and compare its performance with that of a PI controller. In the proposed controller, the construction of the membership functions and rule base, which is highly heuristic, can be achieved using a training process. This allows the combination of knowledge of human experts and evidence from input-output data.

• Journal of the Institute of Convergence Signal Processing
• /
• v.5 no.2
• /
• pp.138-142
• /
• 2004

This paper proposes the design of the model following control system using the PID control structure. PID control system became model following control by inserting new pre-compensator in order to improve control performance in discrete-time region. Gain of the PID controller needs to be readjusted when response of system changes due to disturbance or load fluctuation. Performance of control system improves by joining neural network to PID control system because performance of control system depends largely on each PID gain in PID control system. And the games of the PID controller in the proposed control system are automatically adjusted by back-propagation algorithm of the neural network. Angular position of DC servo motor is selected as a plant in order to verify control performance in model following control. After it is applied to the position control system, it's performance is verified through computer experiment.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.18 no.6
• /
• pp.84-90
• /
• 2004

Many control techniques have been proposed in order to improve the control performance in the control system. In the feedback control system the output of controller is generally used as the input of a plant But the undesired noise is included in the output of a controller. Therefore, there is a need to use a precompensator for rejecting the undesired noise and improving the response characteristic of a system. In this paper, the design method of a precompensator is proposed for the model following control in the I-PD control system. The proposed precompensator is implemented with a neural network. The games of a precompensator are adjusted automatically to obtain a desired response of a system when the response characteristic of a system is changed under a condition.

• Journal of Biomedical Engineering Research
• /
• v.2 no.1
• /
• pp.47-54
• /
• 1981

This paper deals with a new approach to the modelling of neural interactions in the cerebellar cortex to construct a general purpose electronic simulation model. Since physiological data show that cerebellar neural activity changes in an approximately pulse manner in response to pulse stimulation, the differences in timing between excitation and inhibition of cerebellar cells will be treated as pure time delays and the transfer functions of the cells will be presented by pure gains. The parameters to be discussed in this paper are the coupling coefficients between a cell and its several inputs, the magnitude of a coupling coefficient which is presented as a measure of how much influnce a particular has on its target cell. And also this paper has been proposed that the cerrbellum engaged in improving the overall performance of the motor control system, i.e., the cerebellum is a compensator.

• Proceedings of the KIEE Conference
• /
• 2008.10b
• /
• pp.187-192
• /
• 2008

A saturation and deadzone compensator is designed for robot systems using fuzzy logic (FL) and neural network (NN). The classification property of FL system and the function approximation ability of the NN make them the natural candidate for the rejection of errors induced by the saturation and deadzone. The tuning algorithms are given for the fuzzy logic parameters and the NN weights, so that the saturation and deadzone compensation scheme becomes adaptive, guaranteeing small tracking errors and bounded parameter estimates. Formal nonlinear stability proofs are given to show that the tracking error is small. The NN saturation and FL deadzone compensator is simulated on a robot system to show its efficacy.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.5
• /
• pp.602-608
• /
• 2011

In this study, a method of designing a neurointerface using neural network (NN) is proposed for controlling nonholonomic mobile robots. According to the concept of virtual master-slave robots, in particular, a partially stable inverse dynamic model of the master robot is acquired online through the NN by applying a feedback-error learning method, in which the feedback controller is assumed to be based on a PD compensator for such a nonholonomic robot. The NN for the online feedback-error learning can composed that the input layer consists of six units for the inputs $x_i$, i=1~6, the hidden layer consists of two hidden units for hidden outputs $o_j$, j=1~2, and the output layer consists of two units for the outputs ${\tau}_k$, k=1~2. A tracking control problem is demonstrated by some simulations for a nonholonomic mobile robot with two-independent driving wheels. The initial q value was set to [0, 5, ${\pi}$].

• Journal of Institute of Control, Robotics and Systems
• /
• v.14 no.10
• /
• pp.1014-1021
• /
• 2008

This paper proposes an intelligent sliding mode control method for robotic systems with the unknown bound of model uncertainties. In our control structure, the unknown bound of model uncertainties is used as the gain of the sliding controller. Then, we employ the function approximation technique to estimate the unknown nonlinear function including the width of boundary layer and the uncertainty bound of robotic systems. The adaptation laws for all parameters of the self-recurrent wavelet neural network and those for the reconstruction error compensator are derived from the Lyapunov stability theorem, which are used for an on-line control of robotic systems with model uncertainties and external disturbances. Accordingly, the proposed method can not only overcome the chattering phenomenon in the control effort but also have the robustness regardless of model uncertainties and external disturbances. Finally, simulation results for the five-link biped robot are included to illustrate the effectiveness of the proposed method.

• Proceedings of the KIPE Conference
• /
• 2010.07a
• /
• pp.483-484
• /
• 2010

$3{\times}3$ 매트릭스 컨버터(matrix converter)는 3상 입력 전원이 3상 부하에 직접 연결되는 에너지 변환 장치이다. 기존의 AC-DC-AC 전력변환 장치와는 달리 매트릭스 컨버터는 직류단의 전해 커패시터가 존재하지 않기 때문에 불평형의 입력전원은 왜곡된 출력전류를 발생시킨다. 왜곡된 출력전류를 보상하기 위해 본 논문에서는 신경망 기반 전류 보상기를 제안한다. 제안된 기법의 타당성을 시뮬레이션을 통해 증명한다.