• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10b
• /
• pp.261-264
• /
• 1992

This paper proposes a neuro-fuzzy adaptive controller which includes the procedure of initializing the identification neural network(INN) and that of learning the control neural network(CNN). The identification neural network is initialized with the informations of the plant which are obtained by a fuzzy controller and the control neural network is trained by the weight informations of the identification neural network during on-line operation.

• Proceedings of the KIEE Conference
• /
• 1997.07b
• /
• pp.675-678
• /
• 1997

In this paper, we present a fast and robust algorithm for the design of fuzzy controller and identifying fuzzy model from numerical data by combining the cluster estimation method with a linear least squares estimation procedure. The proposed method is compared with Adaptive Neuro-Fuzzy Inference System(ANFIS) as the standard example of neuro-fuzzy model. Finally we will show its usefulness and effectiveness for the design of fuzzy controller of a cart-pole system and fuzzy modeling for the coagulant dosing of a water purification system.

• Journal of Power Electronics
• /
• v.3 no.1
• /
• pp.10-16
• /
• 2003

Implementation and experimental verification of a simple neuro-controller (SNC) as a speed controller for a brush less DC (BLDC) motor is presented. The SNC with one weight and a linear hard limit activation function is trained on-line using the back propagation algorithm. A modified error function is used to ensure good performance during the on-line training, which has been used without previous off-line training. The SNC has been implemented using a computer-interface card mounted on a PC. The driving system performance has been investigated by a number of experimental tests for a variety of input reference speed trajectories.

• Journal of the Institute of Convergence Signal Processing
• /
• v.4 no.3
• /
• pp.41-48
• /
• 2003

This paper presents an indirect adaptive neuro controller using modified chaotic neural networks(MCNN) for nonlinear dynamic system. A modified chaotic neural networks model is presented for simplifying the traditional chaotic neural networks and enforcing dynamic characteristics. A new Dynamic Backpropagation learning method is also developed. The proposed MCNN paradigm is applied to the system identification of a MIMO system and the indirect adaptive neuro controller. The simulation results show good performances, since the MCNN has robust adaptability to nonlinear dynamic system.

• Journal of Power Electronics
• /
• v.4 no.1
• /
• pp.28-38
• /
• 2004

In this paper, the position control of a detuned indirect field oriented control (IFOC) induction motor drive is studied. A proposed Simple-Neuro-Controllers (SNCs) are designed and analyzed to achieve high-dynamic performance both in the position command tracking and load regulation characteristics for robotic applications. The proposed SNCs are trained on-line based on the back propagation algorithm with a modified error function. Four SNCs are developed for position, speed and ｄ-ｑ axes stator currents respectively. Also, a synchronous proportional plus integral-derivative (PI-D) two-degree-of-freedom (2DOF) position controller and PI-D speed controller are designed for an ideal IFOC induction motor drive with the desired dynamic response. The performance of the proposed SNCs and synchronous PI-D 2DOF position controllers for detuned field oriented induction motor servo drive is investigated. Simulation results show that the proposed SNCs controllers provide high-performance dynamic characteristics which are robust with regard to motor parameter variations and external load disturbance. Furthermore, comparing the SNC position controller with the synchronous PI-D 2DOF position controller demonstrates the superiority of the proposed SNCs controllers due to attain a robust control performance for IFOC induction motor servo drive system.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.1
• /
• pp.58-66
• /
• 1992

This paper proposes a neuro-controller for position control of one-link flexible robot arm. Basically the controller consists of a multi-layer neural network and a conventional PD controller. Two controller are parallelly connected. Neural network is traind by the conventional error back propagation learning rules. During learning period, the weights of neural network are adjusted to minimize the position error between the desired hub angle and the actual one. Finally the effectiveness of the proposed approach will be demonstrated by computer simulation.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.999-1004
• /
• 2005

An algorithm for a hybrid controller consists of a sliding mode control part and a fuzzy logic part which ar purposely for nonlinear systems. The sliding mode part of the solution is based on "eigenvalue/vector"-type controller is used as the backstepping approach for tracking errors. The fuzzy logic part is a Mamdani fuzzy model. This is designed by applying sliding mode control (SMC) method to the dynamic model. The main objective is to keep the update dynamics in a stable region by used SMC. After that the plant behavior is presented to train procedure of adaptive neuro-fuzzy inference systems (ANFIS). ANFIS architecture is determined and the relevant formulation for the approach is given. Using the error (e) and rate of error (de), occur due to the difference between the desired output value (yd) and the actual output value (y) of the system. A dynamic adaptation law is proposed and proved the particularly chosen form of the adaptation strategy. Subsequently VSC creates a sliding mode in the plant behavior while the parameters of the controller are also in a sliding mode (stable trainer). This study considers the ANFIS structure with first order Sugeno model containing nine rules. Bell shaped membership functions with product inference rule are used at the fuzzification level. Finally the Mamdani fuzzy logic which is depends on adaptive neuro-fuzzy inference systems structure designed. At the transferable stage from ANFIS to Mamdani fuzzy model is adjusted for the membership function of the input value (e, de) and the actual output value (y) of the system could be changed to trapezoidal and triangular functions through tuning the parameters of the membership functions and rules base. These help adjust the contributions of both fuzzy control and variable structure control to the entire control value. The application example, control of a mass-damper system is considered. The simulation has been done using MATLAB. Three cases of the controller will be considered: for backstepping sliding-mode controller, for hybrid controller, and for adaptive backstepping sliding-mode controller. A numerical example is simulated to verify the performances of the proposed control strategy, and the simulation results show that the controller designed is more effective than the adaptive backstepping sliding mode controller.

• Journal of the Korean Institute of Navigation
• /
• v.21 no.3
• /
• pp.55-66
• /
• 1997

In this paper, applications of neural networks to vibration control of flexible single link robot manipulator are ocnsidered. The architecture of neural networks is a hidden layer, which is comprised of self-recurrent one. Tow neural networks are utilized in a control system ; one as an identifier is called neuro identifier and the othe ra s a controller is called neuro controller. The neural networks can be used to approximate any continuous function to any desired degree of accuracy and the weights are updated by dynamic error-backpropagation algorithm(DEA). To guarantee concegence and to get faster learning, an approach that uses adaptive learning rates is developed by introducing a Lyapunov function. When a flexible manipulator is ratated by a motor through the fixed end, transverse vibration may occur. The motor torque should be controlle dinsuch as way, that the motor is rotated by a specified angle. while simulataneously stabilizing vibration of the flexible manipulators so that it is arrested as soon as possible at the end of rotation. Accurate vibration control of lightweight manipulator during the large body motions, as well as the flexural vibrations. Therefore, dynamic models for a flexible single link manipulator is derived, and LQR controller and nerual networks controller are composed. The effectiveness of the proposed nerual networks control system is confirmed by experiments.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.8 no.3
• /
• pp.9-15
• /
• 1998

This paper presents an adaptive learning control method for unknown nonlinear systems by combining neuro control and iterative learning control techniques. In the present control system, an iterative learning controller (ILC) is used for a process of short term memory involved in a temporary adaptive and learning manipulation and a short term storage of a specific temporary action. The learning gain of the iterative learning law is estimated by using a neural network for an unknown system except relative degrees. The control informations obtained by ILC are transferred to a long term memory-based feedforward neuro controller (FNC) and accumulated in it in addition to the previously stored infonnations. This scheme is applied to a two link robot manipulator through simulations.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2005.11a
• /
• pp.373-376
• /
• 2005

본 논문에서는 이와 같은 요청에 부합되는 강인한 처지제어기를 얻고자, 다층 신경회로망을 이용하여 퍼지제어기 멤버쉽 함수의 전건부 및 후건부 파라미터들을 시스템에 알맞게 자기 조정하기 위해 최급구배법(Steepest Gradient Method)에 근거한 오차 역전파 알고리즘으로 적응 학습시킬 수 있는 뉴로-퍼지제어기 (Neuro-Fuzzy Control : NFC)의 구조 및 알고리즘을 제안하였다.