• Proceedings of the KIEE Conference
• /
• 1990.11a
• /
• pp.402-406
• /
• 1990

A new approach of a time-varying switching hyperplane based on the theory of variable structure system (VSS) is proposed for the control of multivariable systems. While the conventional switching surface can net achieve the robust performance against parameter variations and disturbances before the sliding mode occurs, the proposed switching hyperplane, which is obtained from the eigen-structure assignment theory powerfully used in the linear multivariable systems, ensures the sliding mode from the initial state. And new continuous control input which guarantees the sliding mode is proposed. This new control input does not arise chattering problem which arises with the conventional control input of variable structure control systems. Through numerical examples, the expellant performances of the proposed controller are verified.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.300-303
• /
• 2003

An Algorithm based on direct implementation of variable structure systems theory approach is proposed for on-line training of multilayer perceptrons. Network structures which have multiple inputs, single output and one hidden layer are considered and the weights are assumed to have capabilities for continuous time adaptation. The zero level set of the network learning error is regarded as a sliding surface in the learning parameters space. A sliding mode trajectory can be brought on and reached in finite time on such a sliding manifold. Results from simulated on-line identification task for a two-link planar manipulator dynamics are also presented.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.153-156
• /
• 2001

This paper discusses a stochastic stabilization of Takagi-Sugeno (75) fuzzy system with Markovian input delay. The finite Markovian process is adopted to model the input delay of the overall control system. It is assumed that the zero and hold devices are used for control input. The continuous-time 75 fuzzy system with the Markovian input delay is discretized for easy handling delay, accordingly, the discretized 75 fuzzy system is represented by a discrete-time 75 fuzzy system with jumping parameters. The stochastic stabilizibility of the jump 75 fuzzy system is derived and formulated in terms of linear matrix inequalities (LMls).

• Proceedings of the KIEE Conference
• /
• 2000.11d
• /
• pp.621-623
• /
• 2000

In this paper, the effective design method of the predictive controller using fuzzy neural networks(FNNs) is presented for the Intelligent control of chaotic nonlinear systems. In our design method of controller, predictor parameters are tuned by the error value between the actual output of a chaotic nonlinear system and that of a fuzzy neural network model. And the parameters of predictive controller using fuzzy neural network are tuned by the gradient descent method which uses control error value between the actual output of a chaotic nonlinear system and the reference signal. In order to evaluate the performance of our controller, it is applied to the Duffing system which are the representative continuous-time chaotic nonlinear systems and the Henon system which are representative discrete-time chaotic nonlinear systems.

• 제어로봇시스템학회:학술대회논문집
• /
• 1995.10a
• /
• pp.432-435
• /
• 1995

In this paper a new approach to obtain the solution of the linear-quadratic Gaussian control problem for singularly perturbed discrete-time stochastic systems is proposed. The alogorithm proposed is based on exploring the previous results that the exact solution of the global discrete algebraic Riccati equations is found in terms of the reduced-order pure-slow and pure-fast nonsymmetric continuous-time algebraic Riccati equations and, in addition, the optimal global Kalman filter is decomposed into pure-slow and pure-fast local optimal filters both driven by the system measurements and the system optimal control input. It is shown that the optimal linear-quadratic Gaussian control problem for singularly perturbed linear discrete systems takes the complete decomposition and parallelism between pure-slow and pure-fast filters and controllers.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.53 no.2
• /
• pp.65-75
• /
• 2004

This paper presents a generalized predictive control method based on a fuzzy neural network(FNN) model, which uses the on-line multi-step prediction, fur the intelligent control of chaotic nonlinear systems whose mathematical models are unknown. In our design method, the parameters of both predictor and controller are tuned by a simple gradient descent scheme, and the weight parameters of FNN are determined adaptively during the operation of the system. In order to design a generalized predictive controller effectively, this paper describes computing procedure for each of the two important parameters. Also, we introduce a projection matrix to determine the control input, which deceases the control performance function very rapidly. Finally, in order to evaluate the performance of our controller, the proposed method is applied to the Doffing and Henon systems, which are two representative continuous-time and discrete-time chaotic nonlinear systems, res reactively.

• Proceedings of the KIEE Conference
• /
• 1993.07a
• /
• pp.369-372
• /
• 1993

In this thesis, we consider an iterative learning controller to control the continuous trajectory of 2 links direct drive robot manipulator and process computer simulation and real-time experiment. To improve control performance, we adapt an iterative learning control algorithm, drive a sufficient condition for convergence from which is drived extended conventional control algorithm and get better performance by extended learning control algorithm than that by conventional algorithm from simulation results. Also, experimental results show that better performance is taken by extended learning algorithm.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.2
• /
• pp.138-143
• /
• 2006

This paper presents intelligent digital redesign method for hybrid state space fuzzy-model-based controllers. For effectiveness and stabilization of continuous-time uncertain nonlinear systems under discrete-time controller, Takagi-Sugeno(TS) fuzzy model is used to represent the complex system. And global approach design problems viewed as a convex optimization problem that we minimize the error of the norm bounds between nonlinearly interpolated linear operators to be matched. Also, by using the bilinear and inverse bilinear approximation method, we analyzed nonlinear system's uncertain parts more precisely. When a sampling period is sufficiently small, the conversion of a continuous-time structured uncertain nonlinear system to an equivalent discrete-time system have proper reason. Sufficiently conditions for the global state-matching of the digitally controlled system are formulated in terms of linear matrix inequalities (LMIs). Finally, a TS fuzzy model for the chaotic Lorentz system is used as an . example to guarantee the stability and effectiveness of the proposed method.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.16 no.4
• /
• pp.270-275
• /
• 2016

We present an approach for acceleration actor-critic algorithm for reinforcement learning with continuous action space. Actor-critic algorithm has already proved its robustness to the infinitely large action spaces in various high dimensional environments. Despite that success, the main problem of the actor-critic algorithm remains the same-speed of convergence to the optimal policy. In high dimensional state and action space, a searching for the correct action in each state takes enormously long time. Therefore, in this paper we suggest a search accelerating function that allows to leverage speed of algorithm convergence and reach optimal policy faster. In our method, we assume that actions may have their own distribution of preference, that independent on the state. Since in the beginning of learning agent act randomly in the environment, it would be more efficient if actions were taken according to the some heuristic function. We demonstrate that heuristically-accelerated actor-critic algorithm learns optimal policy faster, using Educational Process Mining dataset with records of students' course learning process and their grades.

• 제어로봇시스템학회:학술대회논문집
• /
• 1988.10b
• /
• pp.1045-1050
• /
• 1988

The problem of applying the linear integral filter in analysis and parameter estimation of linear continuous systems is discussed. A discrete-time model, which is equivalent to that obtained using the bilinear z transformation, is derived and employed to predict system output. It is shown that the output error can be controlled through the sampling interval. In order to obtain unbiased estimates, an instrumental variable (IV) method is proposed, where the instrumental variables are constituted using adaptive filtering. Some problems on implementation of the recursive IV algorithm are discussed. Both theoretical analysis and simulation study are given to illustrate the proposed methods.