• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.4
• /
• pp.145-154
• /
• 1994

This paper proposes an efficient method for improving the training performance of the neural network by using a hybrid of a stochastic approximation and a backpropagation algorithm. The proposed method improves the performance of the training by appliying a global optimization method which is a hybrid of a stochastic approximation and a backpropagation algorithm. The approximate initial point for a stochastic approximation and a backpropagation algorihtm. The approximate initial point for fast global optimization is estimated first by applying the stochastic approximation, and then the backpropagation algorithm, which is the fast gradient descent method, is applied for a high speed global optimization. And further speed-up of training is made possible by adjusting the training parameters of each of the output and the hidden layer adaptively to the standard deviation of the neuron output of each layer. The proposed method has been applied to the parity checking and the pattern classification, and the simulation results show that the performance of the proposed method is superior to that of the backpropagation, the Baba's MROM, and the Sun's method with randomized initial point settings. The results of adaptive adjusting of the training parameters show that the proposed method further improves the convergence speed about 20% in training.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10a
• /
• pp.103-106
• /
• 1996

This paper presents an iterative fuzzy learning control scheme which is applicable to a broad class of nonlinear systems. The control scheme achieves system stability and boundedness by using the linear feedback plus adaptive fuzzy controller and achieves precise tracking by using the iterative learning rules. The switching mode control unit is added to the adaptive fuzzy controller in order to compensate for the error that has been inevitably introduced from the fuzzy approximation of the nonlinear part. It also obviates any supervisory control action in the adaptive fuzzy controller which normally requires high gain signal. The learning control algorithm obviates any output derivative terms which are vulnerable to noise.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.247-250
• /
• 2001

In this paper, we propose an growing and pruning algorithm to design the adaptive structure of modular wavelet neural network(MWNN) with F-projection and geometric growing criterion. Geometric growing criterion consists of estimated error criterion considering local error and angle criterion which attempts to assign wavelet function that is nearly orthogonal to all other existing wavelet functions. These criteria provide a methodology that a network designer can constructs wavelet neural network according to one's intention. The proposed growing algorithm grows the module and the size of modules. Also, the pruning algorithm eliminates unnecessary node of module or module from constructed MWNN to overcome the problem due to localized characteristic of wavelet neural network which is used to modules of MWNN. We apply the proposed constructing algorithm of the adaptive structure of MWNN to approximation problems of 1-D function and 2-D function, and evaluate the effectiveness of the proposed algorithm.

• Proceedings of the KIEE Conference
• /
• 2003.11b
• /
• pp.187-190
• /
• 2003

In this paper, we address the robust adaptive backstepping controller using fuzzy neural network (FHIN) for a class of uncertain output feedback nonlinear systems with disturbance. A new algorithm is proposed for estimation of unknown bounds and adaptive control of the uncertain nonlinear systems. The state estimation is solved using K-fillers. All unknown nonlinear functions are approximated by FNN. The FNN weight adaptation rule is derived from Lyapunov stability analysis and guarantees that the adapted weight error and tracking error are bounded. The compensated controller is designed to compensate the FNN approximation error and external disturbance. Finally, simulation results show that the proposed controller can achieve favorable tracking performance and robustness with regard to unknown function and external disturbance.

• International Journal of CAD/CAM
• /
• v.6 no.1
• /
• pp.73-79
• /
• 2006

In this paper, we propose a new triangular finite element mesh generation method based on simplification of high-density mesh and adaptive Level-of-Detail (LOD) methods for efficient CAE. In our method, mesh simplification is used to control the mesh properties required for FE mesh, such as the number of triangular elements, element shape quality and size while keeping the specified approximation tolerance. Adaptive LOD methods based on vertex hierarchy according to curvature and region of interest, and global LOD method preserving density distributions are also proposed in order to construct a mesh more appropriate for CAE purpose. These methods enable efficient generation of FE meshes with properties appropriate for analysis purpose from a high-density mesh. Finally, the effectiveness of our approach is shown through evaluations of the FE meshes for practical use.

• 제어로봇시스템학회:학술대회논문집
• /
• 1995.10a
• /
• pp.178-182
• /
• 1995

In this paper, the fuzzy approximator and nonlinear inversion control scheme are considered. An adaptive nonlinear control is proposed based on the speed gradient algorithms proposed by Fradkov. This proposed control scheme is that three types of adaptive law is utilized to approximate the unknown function f by fuzzy logic system in designing the nonlinear inversion controller for the nonlinear system. In order to reduce the approximation errors, the differences of nonlinear function and fuzzy approximator, another three types of adaptive law is also introduced and the stability of proposed control scheme are proven with SG algorithm.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.5
• /
• pp.852-857
• /
• 2008

A new adaptive neuro-control algorithm for a SISO strict-feedback nonlinear system is proposed. All the previous adaptive neural control algorithms for strict-feedback nonlinear systems are based on the backstepping scheme, which makes the control law and stability analysis very complicated. The main contribution of the proposed method is that it demonstrates that the state-feedback control of the strict-feedback system can be viewed as the output-feedback control problem of the system in the normal form. As a result, the proposed control algorithm is considerably simpler than the previous ones based on backstepping. Depending heavily on the universal approximation property of the neural network (NN), only one NN is employed to approximate the lumped uncertain system nonlinearity. The Lyapunov stability of the NN weights and filtered tracking error is guaranteed in the semi-global sense.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.54 no.11
• /
• pp.650-653
• /
• 2005

In this paper, adaptive high-order neural network controller(AHONNC) is adopted to control an induction servomotor. A algorithm is developed by combining compensation control and high-order neural networks. Moreover, an adaptive bound estimation algorithm was proposed to estimate the bound of approximation error. The weight of the high-order neural network can be online tuned in the sense of the Lyapunov stability theorem; thus, the stability of the closed-loop system can be guaranteed. Simulation results for induction servomotor drive system are shown to confirm the validity of the proposed controller.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.57 no.1
• /
• pp.22-29
• /
• 2008

In this paper, we propose an adaptive color enhancement algorithm. It is used for the flat panel displays (FPDs) such as LCD, PDP, and so on. The proposed algorithm consists of an adaptive linear approximation CDF(Cumulative Density Function) algorithm and an adaptive saturation enhancement algorithm. The one is for contrast enhancement which prevents an image from the distortion by luminance transient of an input image. The other is the algorithm which improves the saturation without the contour artifact and over-saturation, whose problems are generated during the enhancing saturation. In addition, it allows to achieve the high quality image using the saturation enhancement method for a preferred color of original image. Visual test and standard deviation of their histograms have been applied to evaluate the resultant output images of the proposed algorithm.

• Structural Engineering and Mechanics
• /
• v.29 no.2
• /
• pp.117-134
• /
• 2008

A new adaptive dual reciprocity boundary element method for dynamic analysis of 3-D structures is presented in this paper. It is based on finding the best approximation function of a radial basis function (RBF) group $f=r^n+c$ which minimize the error of displacement field expansion. Also, the effects of some parameters such as the existence of internal points, number of RBF functions and position of collocation nodes in discontinuous elements are investigated in this adaptive procedure. Three numerical examples show improvement in dynamic response of structures with adaptive RBF in dual reciprocity with respect to ordinary BEM.