• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.431-431
• /
• 2000

The modeling of 5-bar linkage robot manipulator dynamics by means of a mathematical and neural architecture is presented. Such a model is applicable to the design of a feedforward controller or adjustment of controller parameters. The inverse model consists of two parts: a mathematical part and a compensation part. In the mathematical part, the subsystems of a 5-bar linkage robot manipulator are constructed by applying Kawato's Feedback-Error-Learning method, and trained by given training data. In the compensation part, MLP backpropagation algorithm is used to compensate the unmodeled dynamics. The forward model is realized from the inverse model using the inverse of inertia matrix and the compensation torque is decoupled in the input torque of the forward model. This scheme can use tile mathematical knowledge of the robot manipulator and analogize the robot characteristics. It is shown that the model is reasonable to be used for design and initial gain tuning of a controller.

• Proceedings of the KIEE Conference
• /
• 2000.07d
• /
• pp.2380-2382
• /
• 2000

Fuzzy system which are based on membership functions and rules, can control nonlinear, uncertian, complex system well. However, Fuzzy controller has problems: It is difficult to design a stable for amateur. To update the then-part membership functions of the fuzzy controller can be designed using the error back-propagation algorithm to be minimized error. Then we could be optimized the system choosing a good performance index. The proposed fuzzy controller based on neural network is applied to control an inverted pendulum for demonstration of the robustness of proposed methodology.

• Proceedings of the KIEE Conference
• /
• 1996.11a
• /
• pp.407-410
• /
• 1996

In this Paper, we propose the feedback method having neural network to control the chaotic signals to periodic signals. This controller has very simple structure, it is immune to small parameter variations, the precise access to system parameters is not required and it is possible to follow ones of its inherent periodic orbits or the desired orbits without error, The controller consist of linear feedback gain and neural network. The learning of neural network is achieved by error-backpropagation algorithm. To prove and analyze the proposed method, we construct a software tool using c-language.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.2
• /
• pp.552-557
• /
• 2000

In the paper, we proposed a face recognition method that uses GA-BP(Genetic Algorithm-Back propagation Network) that optimizes initial parameters such as bias values or weights. Each pixel in the picture is used for input of the neuralnetwork. The initial weights of neural network is consist of fixed-point real values and converted to bit string on purpose of using the individuals that arte expressed in the Genetic Algorithm. For the fitness value, we defined the value that shows the lowest error of neural network, which is evaluated using newly defined adaptive re-learning operator and built the optimized and most advanced neural network. Then we made experiments on the face recognition. In comparison with learning convergence speed, the proposed algorithm shows faster convergence speed than solo executed back propagation algorithm and provides better performance, about 2.9% in proposed method than solo executed back propagation algorithm.

• The Journal of the Acoustical Society of Korea
• /
• v.37 no.4
• /
• pp.196-205
• /
• 2018

This paper presents an algorithm for passive localization of a ship by applying the ray back-propagation technique to the ship radiation noise data. The previous method [S. H. Abadi, D. Rouseff and D. R. Dowling, J. Acoust. Soc. Am. 131, 2599-2610 (2012)] estimates the position of a sound source in the near-field environment with no array tilt by using the RBD (Ray-based Blind Deconvolution) and ray back-propagation techniques. However, when there exists an array tilt, the above method leads to a large position estimation error. In order to overcome the problem, this study proposes an algorithm that estimates the position of a sound source by correcting the array tilt using the RBD and ray back-propagation techniques. The proposed algorithm was verified by using the ship noise of SAVEX15 (Shallow-water Acoustic Variability EXperiment in 2015) experimental data.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.7
• /
• pp.211-218
• /
• 1994

Stochastic computation employs random pulse streams to represent numbers. In this paper, we study a new method to implement the number system which uses the ratio of the numbers of ones and zeros in the pulse streams. In this number system. if P is the probability that a pulse is one in a pulse stream then the number X represented by the pulse stream is defined as P/(1-P). We propose circuits to implement the basic operations such as addition multiplication and sigmoid function with this number system and examine the error characteristics of such operations in stochastic computation. We also propose a neuron model and derive a learning algorithm based on backpropagation for the 3-layered feedforward neural networks. We apply this learning algorithm to a digit recognition problem. To analyze the results, we discuss the errors due to the variance of the random pulse streams and the quantization noise of finite length register.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.04a
• /
• pp.390-393
• /
• 2007

In this paper, applications of self recurrent neuro-fuzzy controller to stabilization of nonlinear system are considered. The architecture of self recurrent neuro-fuzzy controller is fix layer, and the hidden layer is comprised of self recurrent architecture. Also, generalized dynamic error-backpropagation algorithm is used for the learning of the self recurrent neuro-fuzzy controller. To demonstrate the efficiency of the self recurrent neuro-fuzzy control algorithm presented in this study, a self recurrent neuro-fuzzy controller was designed and then a comparative analysis was made with LQR controller through an simulation.

• Korean Journal of Hydrosciences
• /
• v.4
• /
• pp.1-9
• /
• 1993

It is a major objective for the management and operation of water resources system to forecast streamflows. The applicability of artificial neural network model to hydrologic system is analyzed and the performance is compared by statistical method with observed. Multi-layered perception was used to model rainfall-runoff process at Pyung Chang River Basin in Korea. The neural network model has the function of learning the process which can be trained with the error backpropagation (EBP) algorithm in two phases; (1) learning phase permits to find the best parameters(weight matrix) between input and output. (2) adaptive phase use the EBP algorithm in order to learn from the provided data. The generalization results have been obtained on forecasting the daily and hourly streamflows by assuming them with the structure of ARMA model. The results show validities in applying to hydrologic forecasting system.

• 제어로봇시스템학회:학술대회논문집
• /
• 1994.10a
• /
• pp.385-389
• /
• 1994

So far many researches have studied to control a cart system with a pole on the top of itself (forwards we call it simply a cart system) which is movable only to the directions to which a cart moves, using neural networks and genetic algorithms. Especially which it wag solved by genetic algorithms, it was possible to control a cart system more robustly than ordinary methods using neural networks but it had problems too, i.e., the control time to be achieved was short and the processing time for it was long. However we could control a cart system using standard genetic algorithm longer than ordinary neural network methods (for example error backpropagation) and could see that robust control was possible. Computer simulation was performed through the personal computer and the results showed the possibility of real time control because the cpu time which was occupied by processes was relatively short.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.36C no.3
• /
• pp.58-64
• /
• 1999

Evolutionary programming is mainly characterized by two factors; one is the selection strategy and the other the mutation rule. In this paper, a new mutation rule that is the same form of well-known backpropagation learning rule of neural networks has been presented. The proposed mutation rule adapts the best individual's value as the target value at the generation. The temporal error improves the exploration through guiding the direction of evolution and the momentum speeds up convergence. The efficiency and robustness of the proposed algorithm have been verified through benchmark test functions.