• Proceedings of the KIEE Conference
• /
• 1995.07b
• /
• pp.714-716
• /
• 1995

This Paper presents a novel and systematic approach to a self-learning controller. The proposed controller is built on a neural network consisting of a standard back propagation (BNN) and approxinate reasoning (AR). The fuzzy inference and knowledge representation are carried out by the neural network structure and computing, instead of logic inference. An architecture similar to that used by traditional model reference adaptive control system (MRAC) is employed.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.6 no.4
• /
• pp.271-276
• /
• 2006

To implement a generalization of value functions in Adaptive Search Element (ASE)-reinforcement learning, CMAC (Cerebellar Model Articulation Controller) is integrated into ASE controller. ASE-reinforcement learning scheme is briefly studied to discuss how CMAC is integrated into ASE controller. Neighbourhood Sequential Training for CMAC is utilized to establish the look-up table and to produce discrete control outputs. In computer simulation, an ASE controller and a couple of ASE-CMAC neural network are trained to balance the inverted pendulum on a cart. The number of trials until the controllers are established and the learning performance of the controllers are evaluated to find that generalization ability of the CMAC improves the speed of the ASE-reinforcement learning enough to realize the cartpole control system.

• The Transactions of the Korean Institute of Electrical Engineers B
• /
• v.48 no.8
• /
• pp.426-430
• /
• 1999

In this paper, a speed sensorless algorithm for a high-speed induction motor is proposed. The proposed algorithm simply estimates rotor speed by integrating the deviation between the command current value of a controller and the real current value of the motor. To estimate rotor speed without a speed sensor, a fuzzy speed controller and a neural network speed estimator are applied. Computer simulation and implementation of the proposed system is described.

• Proceedings of the KIEE Conference
• /
• 1996.07b
• /
• pp.1205-1207
• /
• 1996

The computed-torque method (CTM) shows good trajectory tracking performance in controlling robot manipulator if there is no disturbance or modelling errors. But with the increase of a payload or the disturbance of a manipulator, the tracking errors become large. So there have been many researchs to reduce the tracking error. In this paper, we propose a new control algorithm based on the CTM that decreases a tracking error by generating new reference trajectory to the controller. In this algorithm we used a fuzzy system based on the rule bases. For the numerical simulation, we used a 2-link robot manipulator. To simulate the disturbance due to a modelling uncertainty, we added errors to each elements of the inertia matrix and the nonlinear terms and assumed a payload to the end-effector. In the simulations of several cases, our method showed better trajectory tracking performance compared with the CTM.

• Proceedings of the KIEE Conference
• /
• 1993.07a
• /
• pp.241-243
• /
• 1993

In this paper, a class of neural-network controllers with two inputs of error and error change, is applied to the position control of D.C. servo system. The proposed controller is learned by error back-propagating error information to compensate the weighting value using its previous derivatives and to decrease exponentially a series of self learning coefficients. Through the simulations and implementations, the effectiveness and superiority to the conventional fuzzy controller is proved.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.41.6-41
• /
• 2001

As autonomous mobile robot become more widely used in industry, the importance of navigation system is rising, But eh primary method of locomotion is with wheels, which cause man problems in controlling tracked mobile robots. In this paper, we discuss the used navigation control of tracked mobile robots with multiple sensors. The multiple sensors are composed of ultrasonic wave sensors and vision sensors. Vision sensors gauge distance using a laser and create visual images, to estimate robot position. The 80196 is used at close range and the vision board is used at long range. Data is managed in the main PC and management is distributed to ever sensor. The controller employs fuzzy logic.

• Advances in aircraft and spacecraft science
• /
• v.10 no.4
• /
• pp.369-391
• /
• 2023

In our research we have offered a solid solution for aeronautical analysis. which can guarantee the asymptotic stability of coupled nonlinear facilities. According to the theoretical solutions and methods presented, the engine of this aircraft is a small high-bypass turbofan engine. using the non-linear aero-motor control approach and this paper focuses on the power management function of the aero-motor control system. These include static controls and transient controls. A mathematical model of the high-bypass-ratio two-spool unmixed-flow aeroengine was developed through a set of nonlinear dynamic equations verified by experimental data. A single actuator using the displacement method is designed to maintain a certain level of thrust under steady-state conditions. and maintains repeatable performance during transient operation from the requested thrust phase to the next. A single controller can compensate for the effects of noise and harmonic noise at many performance points. And the dynamic performance of a single controller is satisfactory during the transient. for fairness Numerical and computer experiments are described in the perfection of the methods we offer in research.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.22 no.12
• /
• pp.1220-1226
• /
• 2012

This paper proposes a novel type of tactile device utilizing magnetorheological(MR) fluid which can be applicable for haptic master of minimally invasive surgery(MIS) robotic system. The salient feature of the controllability of rheological properties by the intensity of the magnetic field(or current) makes this potential candidate of the tactile device. As a first step, an appropriate size of the tactile device is designed and manufactured via magnetic analysis. Secondly, in order to determine proper input magnetic field the repulsive forces of the real body parts such as hand and neck are measured. Subsequently, the repulsive forces of the tactile device are measured by dividing 5 areas. The final step of this work is to obtain desired force in real implementation. Thus, in order to demonstrate this goal a neuro-fuzzy logic is applied to get the desired repulsive force and the error between the desired and actual force is evaluated.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.1
• /
• pp.76-84
• /
• 2003

Now days, It is based on GIS and GPS, it can search for the shortest path and estimation of arrival time by using the internet and cell phone to driver. But, even though good car navigation system does not create which is the shortest path when there average vehicle speed is 10 -20 Km. Therefore In order to reduce vehicle waiting time and average vehicle speed, we suggest optimal green time algorithm using fuzzy adaptive control, where there are different traffic intersection length and lane. In this paper, it will be able to forecast the optimal traffic information, estimation of destination arrival time, under construction road, and dangerous road using internet.

• 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.