• 제어로봇시스템학회논문지
• /
• 제10권12호
• /
• pp.1295-1304
• /
• 2004

In this paper, the Bayesian recurrent neural network is proposed to predict time series data. A neural network predictor requests proper learning strategy to adjust the network weights, and one needs to prepare for non-linear and non-stationary evolution of network weights. The Bayesian neural network in this paper estimates not the single set of weights but the probability distributions of weights. In other words, the weights vector is set as a state vector of state space method, and its probability distributions are estimated in accordance with the particle filtering process. This approach makes it possible to obtain more exact estimation of the weights. In the aspect of network architecture, it is known that the recurrent feedback structure is superior to the feedforward structure for the problem of time series prediction. Therefore, the recurrent neural network with Bayesian inference, what we call Bayesian recurrent neural network (BRNN), is expected to show higher performance than the normal neural network. To verify the proposed method, the time series data are numerically generated and various kinds of neural network predictor are applied on it in order to be compared. As a result, feedback structure and Bayesian learning are better than feedforward structure and backpropagation learning, respectively. Consequently, it is verified that the Bayesian reccurent neural network shows better a prediction result than the common Bayesian neural network.

• 대한기계학회논문집A
• /
• 제23권6호
• /
• pp.931-942
• /
• 1999

This paper develops a 3 DOF vehicle model which includes lateral, roll and yaw motion to study a 4WS vehicle. The model is used for the simulation of a 4WS vehicle behavior, and to derive a control algorithm for rear wheel steering. This paper uses a feedforward plus feedback control scheme to compute a rear wheel steering angle. The feedforward control scheme for computing the first rear wheel steering angle uses a gain which is acquired by multiplying a proper value on a gain to maintain a zero sideslip angle. The feedback control scheme for computing the second rear wheel steering angle uses fuzzy logic and model following control scheme. A linear 2 DOF model is used as a reference model for model following control, and is derived from the developed 3 DOF model by neglecting sprung mass roll motion. A reference state variable is yaw rate, and is computed using the linear 2 DOF model. J-turn and lane change maneuver simulation are performed to show the effectiveness of the developed control scheme. The simulation results show that the 4WS vehicle with the developed control scheme has much better performance in yaw rate, lateral acceleration, roll angle, and sideslip angle than the 2WS vehicle. Also, the results show that the performance of the developed control is close to the one of an optimal control which assumes all states are perfect.

• 대한기계학회논문집A
• /
• 제21권3호
• /
• pp.502-510
• /
• 1997

A disturbance rejection controller using eccentricity filtering and LQ control techniques is proposed to alleviate the effecto of major roll eccentricity in multivariable cold-rolling processes. Fundamental problems in multivariable cold-rolling processes such as process time delay inherent in exit thickness measurement and non-stationary characteristics of roll eccentricity signals can be overcome by the proposed control method. The filtered instantaneous estimate of roll eccentricity may be exploited to improve instantaneous estimate of the exit thickness variation based on roll force and roll gap measurements, and a feedforward compensator is augmented as a reference for a gaugemeter thickness estimator. LQ feedback controller is combined with eccentricity filter for the attenuation of the exit thickness variation due to the entry thickness variation. The simulation results show that the roll eccentricity disturbance is significantly eliminated and other disturbances also are attenuated.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 추계학술대회 논문집 학회본부 B
• /
• pp.475-478
• /
• 1999

In this paper, we propose the robust performance design of multivariable systems within the framework of Quantitative Feedback Theory(QFT) using ICD. The ICD(Individual Channel Design) is a multivariable control method based on the classical frequence response. It is considered to apply feedforward controller for compensating the effect of interconnection between channels. Performance of the proposed method are demonstrated by simulations in appling gas turbine model.

• 한국지열·수열에너지학회논문집
• /
• 제7권1호
• /
• pp.38-44
• /
• 2011

In this study, heating performance and energy consumption characteristics with control strategies for central heating system were researched by the simulation. The simulation analysis is made by TRNSYS ver. 15 with the actual data. The parametric study on proportional factor, control time interval and outdoor air temperatures changes were done to compare control characteristics and energy performance, respectively. As a result, the simulation results with various parameter changes show good heating performance and energy saving.

• Nuclear Engineering and Technology
• /
• 제26권3호
• /
• pp.420-424
• /
• 1994

가압경수로형 원자력발전소의 저출력 및 과도상태에서의 개선된 증기발생기 수위 제어 방식을 제시하였다. 수축 및 팽창 현상에 의한 수위의 요동을 줄이기 위해 기존의 비례·적분 제어기에 증기발생기 압력 및 급수온도를 고려한 앞먹임 보상부를 첨가하였다. 원전 훈련용 시뮬레이터를 이용하여 시뮬레이션을 수행한 결과 기존방식에 비해 적은 수위오차, 훨씬 빠른 진정시간을 얻을 수 있었다. 제시된 알고리즘은 구현이 용이하고 실제 적용도 가능하리라 판단된다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1993년도 한국자동제어학술회의논문집(국내학술편); Seoul National University, Seoul; 20-22 Oct. 1993
• /
• pp.552-558
• /
• 1993

In this paper, control of a planar two-link structurally flexible robotic manipulator executing unconstrained and constrained maneuvers is considered. The dynamic model, which is obtained by using the extended Hamilton's principle and the Galerkin criterion, includes the impact force generated during the transition from unconstrained to constrained segment of the robotic task. A method is presented to obtain the linearized equations of motion in Cartesian space for use in designing the control system. The linear quadratic Gaussian with loop transfer recovery (LQG/LTR) design methodology is exploited to design a robust feedback control system that can handle modeling errors and sensor noise, and operate on Cartesian space trajectory errors. The LQG/LTR compensator together with a feedforward loop is used to control the flexible manipulator. Simulated results are presented for a numerical example.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1990년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 26-27 Oct. 1990
• /
• pp.30-35
• /
• 1990

Learning control of a robot manipulator is proposed using the backpropagation neural network. The learning controller is composed of both a linear feedback controller and a neural network-based feedforward controller. The stability analysis of the learning controller is presented. Three energy functions are selected in teaching the neural network controller : 1/2.SIGMA.vertical bar torque error vertical bar $^{2}$, 1/2.SIGMA..alpha. vertical bar position error vertical bar $^{2}$ + .betha. vertical bar velocity error vertical bar $^{2}$ + .gamma. vertical bar acceleration error vertical bar $^{2}$ and learning methods are presented. Simulation results show that the learning controller which is learned to minimize the third energy function performs better than the others in tracking problems. Some properties of the learning controller are discussed with simulation results.

• 한국정밀공학회지
• /
• 제19권9호
• /
• pp.106-110
• /
• 2002

A need fer stable and comfortable cabins in the high-speed passenger ships has increased. For active control of the motion of the ship cabin, a few control algorithms have been applied to the three dimensional real models in the vibration basin. Experimental results show that the feedforward neural network with a linear feedback controller is one of the promising control algorithms for this active control.

• 동력기계공학회지
• /
• 제11권1호
• /
• pp.134-139
• /
• 2007

A delivery ship is used to handle the cargo with the crane to/from the ships. The ship is inclined in the direction of a cargo which is hung on a crane. In this case, a arc shaped rail should be in the equilibrium state to get good anti-rolling performance. In this study, a device and control algorithm are developed to take accurate and quick equilibrium of the rail. The device is composed of a hinged immovable support, screw jack and damper. And the control system is based on I-PD control law to consider of control input saturation and overshoot. The controller is composed of integral controller of feedforward path and proportional-derivative controller of feedback path. The parameters of controller is designed to follow the reference signal and to remove overshoot. The simulation results show that the desirable control performance is achieved.