• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.7
• /
• pp.399-412
• /
• 2006

A clustering based wireless internet proxy server with cooperative caching has a problem of minimizing overall performance because some servers become overloaded if client request pattern is Hot-Spot or uneven. We propose a self-learning based adaptive clustering scheme to solve the poor performance problems of the existing clustering in case of Hot-Spot or uneven client request pattern. In the proposed scheme, requests are dynamically redistributed to the other servers if some servers supposed to handle the requests become overloaded. This is done by a self-learning based method based dynamic weight adjustment algorithm so that it can be applied to a situation with even various request pattern or a cluster of hosts with different performance. We performed experiments in a clustering environment with 16 PCs and a load balancer. Experimental results show the 54.62% performance improvement of the proposed schemes compared to the existing schemes.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.37 no.12
• /
• pp.879-883
• /
• 1988

In this paper, self tuning control of interconnected systems are dealt in view point of large scale system control. The plant model is given in MIMO ARMA procss. This process is simlified as independent SISO ARMA processes having offset terma, which are considered as effects of interconnections. In each decentralized system, self tuning controller with instrumental variable method is adopted. As a result, this algorithm enables the paramter estimation to be unbiased and non-drift. This controller contains a new implicit offset rejection technique. Simulation results consider well with the analysis in case of linear interconnection.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.1
• /
• pp.287-297
• /
• 2018

To improve the performance of sensorless induction motor (IM) drives, a novel speed estimation method based on the real-time adaptive extended Kalman filter (RAEKF) is proposed in this paper. In this algorithm, the fuzzy factor is introduced to tune the measurement covariance matrix online by the degree of mismatch between the actual innovation and the theoretical. Simultaneously, the fuzzy factor can be continuously self-tuned tuned by the fuzzy logic reasoning system based on Takagi-Sugeno (T-S) model. Therefore, the proposed method improves the model adaptability to the actual systems and the environmental variations, and reduces the speed estimation error. Furthermore, a simple exponential function based on the fuzzy theory is used to reduce the computational burden, and the real-time performance of the system is improved. The correctness and the effectiveness of the proposed method are verified by the simulation and experimental results.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.4
• /
• pp.21-28
• /
• 2004

This paper presents a direct multivariable self-tuning controller using neural network which adapts to the changing parameters of the higher order multivariable nonlinear system with nonminimum phase behavior, mutual interactions and time delays. The nonlinearities are assumed to be globally bounded, and a multivariable nonlinear system is divided linear part and nonlinear part. The neural network is used to estimate the controller parameters, and the control output is obtained through estimated controller parameter. In order to demonstrate the effectiveness of the proposed algorithm the computer simulation is done to adapt the multivariable nonlinear nonminimm phase system with time delays and changed system parameter after a constant time. The proposed method compared with direct multivariable adaptive controller using neural network.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.39 no.6
• /
• pp.1-8
• /
• 2002

This paper presents a generalized minimum-variance self-tuning controller with a PID structure using neural network which adapts to the changing parameters of the nonlinear system with nonminimum phase behavior and time delays. The neural network is used to estimate the controller parameters, and the control output is obtained through estimated controller parameter. In order to demonstrate the effectiveness of the proposed algorithm, the computer simulation is done to adapt the nonlinear nonminimum phase system with time delays and changed system parameter after a constant time. The proposed method compared with direct adaptive controller using neural network.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.39 no.4
• /
• pp.364-370
• /
• 1990

In this paper, we propose a self-tuning PID controller for unknown systems with time-varying delay. Using pole placement equations, we derive the controller that can be extended to the multi-step time delay case. The time-varying delays are estimated by a prediction error delay method using multiple predictors. Since the order of the estimation vector is not increased, the persistant exciting condition of control input is alleviated. Since the least square method gives biased parameter estimates for colored noise cases, the recursive instrumental variable method is used to estimate system parameters. The computational burden of the proposed method is less than the conventional adaptive methods. Computer simulations are performed to illustrate the efficiency of the proposed method.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.11
• /
• pp.1110-1116
• /
• 2010

This paper present an effective complementary filtering method using encoder and gyro sensors for the self-localization(including heading and velocity) of indoor mobile robot. The main idea of the proposed approach is to find the pros and cons of each sensor through a various maneuvering tests and to design of an adaptive complementary filter that works for the entire maneuvering phases. The proposed method is applied to an indoor mobile robot and the performances are verified through extensive experiments.

• Proceedings of the KIEE Conference
• /
• 2006.07d
• /
• pp.1848-1849
• /
• 2006

The new robust controller design method is proposed for the flight control systems with model uncertainties. The proposed control system is a combination of the adaptive dynamic surface control (DSC) technique and the self recurrent wavelet neural network (SRWNN). The adaptive DSC technique provides us with the ability to overcome the "explosion of complexity" problem of the backstepping controller. The SRWNNs are used to observe the arbitrary model uncertainties of flight systems and all their weights are trained on-line. From the Lyapunov stability analysis, their adaptation laws are induced and the uniformly ultimately boundedness of all signals in a closed-loop adaptive system is proved. Finally, simulation results for a high performance aircraft (F-16) are utilized to validate the good tracking performance and robustness of the proposed control system.

• International Journal of Control, Automation, and Systems
• /
• v.5 no.3
• /
• pp.269-282
• /
• 2007

This paper presents a robust control method via the intelligent adaptive backstepping design technique for stable walking of nine-link biped robots with unknown model uncertainties and external disturbances. In our control structure, the self recurrent wavelet neural network(SRWNN) which has the information storage ability is used to observe the uncertainties of the biped robots. The adaptation laws for all weights of the SRWNN are induced from the Lyapunov stability theorem, which are used for on-line controlling biped robots. Also, we prove that all signals in the closed-loop adaptive system are uniformly ultimately bounded. Through computer simulations of a nine-link biped robot with model uncertainties and external disturbances, we illustrate the effectiveness of the proposed control system.

• Journal of Communications and Networks
• /
• v.5 no.3
• /
• pp.249-257
• /
• 2003

Ring networks are very commonly exploited among local area and metropolitan area networks (LAN/MAN), whereas cells or small fixed-size packets are widely used in synchronized ring networks. In this paper, we present an analytical method for evaluating the delay-throughput performance of a MAC protocol with multiple self-tokens in a slotted ring network under uniform traffic. In our analysis, we introduce the stationary probability, which indicates the number of packets in a node. Also, it is assumed that each node has a sufficiently large amount of self-tokens, and a slotted ring has the symmetry. The analytical results with respect to delay-throughput performance have similar values to computer simulated ones. Furthermore, in order to achieve fair access under non-uniform traffic, we propose an adaptive MAC protocol, where the number of self-tokens in a node dynamically varies, based on the number of packets transmitted within a specified period. In the proposed protocol, when the number of packets transmitted by a node within a specified period is larger than a specified threshold, the node decreases the number of self-tokens in a per-node distributed method. That results in creating free slots in the ring, thus all nodes can obtain an equal opportunity to transmit into the ring. Performance results obtained by computer simulation show that our proposed protocol can maintain throughput fairness under non-uniform traffic.