• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.10a
• /
• pp.560-567
• /
• 1998

This paper concerned about a imputed color printed material image in camera to decrease noise and distortion by processing median filtering with input image to identical condition. Also this paper proposed the way of compares a normal printed material with an abnormal printed material color tone with trained a learning of the error back-propagation to block classification by extracting five place from identical block(3${\times}$3) of color printed material R, G, B value. As a representative algorithm of multi-layer perceptron the error Back-propagation technique used to solve complex problems. However, the Error Back-propagation is algorithm which basically used a gradient descent method which can be converged to local minimum and the Back Propagation train include problems, and that may converge in a local minimum rather than get a global minimum. The network structure appropriate for a given problem. In this paper, a good result is obtained by improve initial condition and adjust th number of hidden layer to solve the problem of real time process, learning and train.

• Structural Engineering and Mechanics
• /
• v.87 no.1
• /
• pp.47-56
• /
• 2023

Any pre-existed delamination defect present during manufacturing or induce during service loading conditions in the wingskin adherend invariably shows a greater loss of structural integrity of the spar wingskin joint (SWJ). In the present study, inter-laminar delamination propagation at the critical location of the SWJ has been carried out using contact and multi-point constraint finite elements available with commercial FE software (ANSYS APDL). Strain energy release rates (SERR) based on virtual crack closure technique have been computed for evaluation of the opening (Mode-I), sliding (Mode-II) and cross sliding (Mode-III) modes of delamination by sequential release of multi point constraint elements. The variations of different modes of SERR are observed to be significant by considering varied delamination lengths, material properties of adherends and radius of curvature of the SWJ panel. The SERR rates are seen to be much different at the two pre-embedded delamination ends. This shows dissimilar delamination propagation rates. The maximum is seen to occur in the delamination front in the unstiffened region of the wingskin. The curvature geometry and material anisotropy of SWJ adherends significantly influences the SERR values. Increase in the SERR values are observed with decrease in the radius of curvature of wingskin panel, keeping its width unchanged. SWJs made with flat FRP composite adherends have superior resistance to delamination damage propagation than curved composite laminated SWJ panels. SWJ made with Boron/Epoxy (B/E) material shows greater resistance to the delamination propagation.

• 제어로봇시스템학회:학술대회논문집
• /
• 2002.10a
• /
• pp.102.2-102
• /
• 2002

In this paper, the improved Multi-FNN (Fuzzy-Neural Networks) model is identified and optimized using HCM (Hard C-Means) clustering method and optimization algorithms. The proposed Multi-FNN is based on FNN and use simplified and linear inference as fuzzy inference method and error back propagation algorithm as learning rules. We use a HCM clustering and genetic algorithms (GAs) to identify both the structure and the parameters of a Multi-FNN model. Here, HCM clustering method, which is carried out for the process data preprocessing of system modeling, is utilized to determine the structure of Multi-FNN according to the divisions of input-output space using I/O process data. Also, the parame...

• Journal of Korean Society of Water and Wastewater
• /
• v.27 no.4
• /
• pp.479-487
• /
• 2013

This study was to investigate characteristics for the pressure wave propagation and the maximum pressure near a rapid closure valve which was installed the end of multi piping network. The multi piping network consists of one inlet and three outlet with straight pipes. The diameter of the pipes including the valve was 100 mm, 80 mm, 80 mm respectively. The valve was rapidly closed with the instantaneous time which was 0.023s in the level for the water hammer. For the simulation, the influence of the pipe thickness and deformation due to pressure-wave-propagation was not considered. CFD was conducted under the following condition : the initial pressure was 1bar in the inlet and the mass flow rate was 7.83 kg/s in the outlet(the velocity in the pipe with 100 mm diameter was 1 m/s). As the valve have conditions that were status with and without fluid flow in the pipe after valve closing, the maximum pressure change and the frequency analysis were examined. As the results, the case that was status with fluid flow appeared the higher maximum pressure than another's, the maximum frequency band was about 10 ~ 11 Hz.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.35C no.3
• /
• pp.79-89
• /
• 1998

The Error Back-Propagation(EBP) algorithm is widely applied to train a multi-layer perceptron, which is a neural network model frequently used to solve complex problems such as pattern recognition, adaptive control, and global optimization. However, the EBP is basically a gradient descent method, which may get stuck in a local minimum, leading to failure in finding the globally optimal solution. Moreover, a multi-layer perceptron suffers from locking a systematic determination of the network structure appropriate for a given problem. It is usually the case to determine the number of hidden nodes by trial and error. In this paper, we propose a new algorithm to efficiently train a multi-layer perceptron. OUr algorithm uses stochastic perturbation in the weight space to effectively escape from local minima in multi-layer perceptron learning. Stochastic perturbation probabilistically re-initializes weights associated with hidden nodes to escape a local minimum if the probabilistically re-initializes weights associated with hidden nodes to escape a local minimum if the EGP learning gets stuck to it. Addition of new hidden nodes also can be viewed asa special case of stochastic perturbation. Using stochastic perturbation we can solve the local minima problem and the network structure design in a unified way. The results of our experiments with several benchmark test problems including theparity problem, the two-spirals problem, andthe credit-screening data show that our algorithm is very efficient.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.42 no.3
• /
• pp.51-56
• /
• 2005

To improve the BER of OFDM on a fading channel, a low-density parity check coded OFDM system is proposed in this paper. LDPC codes are decoded with Sum-Product or Belief Propagation Algorithm known by probability propagation algorithm. When LDPC codes are applied to OFDM system, the BER performance is dependant on the iteration number of decoding. To improve the spectral efficiency, multi-level modulations are used in mobile communication system. But, It is not clear how to decode LDPC code used in OFDM with multi-level modulations. In the paper, a decoding algorithm is described for LDPC coded OFDM with MPSK. When use the proposed decoding algorithm, we get the good BER for AWGN and a Fading Channel. Simulation results show that the proposed decoding algorithm is confirmed LDPC coded OFDM with MPSK.

• Journal of Environmental Science International
• /
• v.16 no.7
• /
• pp.771-778
• /
• 2007

This study has been conducted to achieve the following objectives: First, in order to understand the horizontal propagation and attenuation characteristics for the railroad traffic noise, we selected areas within 100 meters away from the railroad and then selected Saemaul-ho and Mugoongwha-ho as the subjects for our experiment. In this way, we analyzed the horizontal propagation and attenuation characteristics for the traffic noise occurring in diversified areas. Second, in order to understand the vertical propagation and attenuation characteristics for the railroad traffic noise, we measured and analyzed the distributional characteristics of vertical sound pressure levels on each floor of multi-storied apartment buildings according to changes of traffic load and types, and the existence or nonexistence of soundproof walls. For the case of the railroad traffic noise, we also selected Samaul-ho and Mugoongwha-ho as the subjects for our experiment, and we measured and analyzed the different noise levels on each floor of multi-storied apartment buildings from the soundproof wall. The results of Horizontal propagation and attenuation characteristics for the railroad traffic noise are as follows: In cases of the flat land, cutting land, and bridge area, as distance increases, the sound pressure level steadily decreases. The sound pressure level for the bridge area is higher than that of the flat land with a measurement of $5.5{\sim}10.2\;dB(A)$. Vertical propagation and attenuation characteristics for the railroad traffic noise are as follows: The amount of sound pressure level decrease is $14.2{\sim}14.8\;dB(A)$ for Samaul-ho and $13.5{\sim}14.3\;dB(A)$ for Mugoongwha-ho when measuring the vertical sound pressure levels at heights lower than 4.5 m, which indicates a fairly large decrease. At 6 m, the amount of decrease is 8.6 dB(A) for Samaul-ho and 8.2 dB(A) for Mugoongwha-ho, which indicates a small decrease.

• Proceedings of the IEEK Conference
• /
• 2000.07b
• /
• pp.1041-1044
• /
• 2000

The enhancement to the back-propagation algorithm presented in this paper has resulted from the need to extract sparsely connected networks from networks employing product terms. The enhancement works in conjunction with the back-propagation weight update process, so that the actions of weight zeroing and weight stimulation enhance each other. It is shown that the error measure, can also be interpreted as rate of weight change (as opposed to ${\Delta}W_{ij}$), and consequently used to determine when weights have reached a stable state. Weights judged to be stable are then compared to a zero weight threshold. Should they fall below this threshold, then the weight in question is zeroed. Simulation of such a system is shown to return improved learning rates and reduce network connection requirements, with respect to the optimal network solution, trained using the normal back-propagation algorithm for Multi-Layer Perceptron (MLP), Higher Order Neural Network (HONN) and Sigma-Pi networks.

• International Journal of Aeronautical and Space Sciences
• /
• v.17 no.4
• /
• pp.467-475
• /
• 2016

Ultrasonic propagation imaging (UPI) has shown great potential for detection of impairments in complex structures and can be used in wide range of non-destructive evaluation and structural health monitoring applications. The software implementation of such algorithms showed a tendency in time-consumption with increment in scan area because the processor shares its resources with a number of programs running at the same time. This issue was addressed by using field programmable gate arrays (FPGA) that is a dedicated processing solution and used for high speed signal processing algorithms. For this purpose, we need an independent and flexible block of logic which can be used with continuously evolvable hardware based on FPGA. In this paper, we developed an FPGA-based ultrasonic propagation imaging system, where FPGA functions for both data acquisition system and real-time ultrasonic signal processing. The developed UPI system using FPGA board provides better cost-effectiveness and resolution than digitizers, and much faster signal processing time than CPU which was tested using basic ultrasonic propagation algorithms such as ultrasonic wave propagation imaging and multi-directional adjacent wave subtraction. Finally, a comparison of results for processing time between a CPU-based UPI system and the novel FPGA-based system were presented to justify the objective of this research.

• Journal of Environmental Impact Assessment
• /
• v.12 no.4
• /
• pp.271-279
• /
• 2003

Temperature and humidity forecasting have been performed using artificial neural networks model(ANN). We composed ANN with multi-layer perceptron which is 2 input layers, 2 hidden layers and 1 output layer. Back propagation algorithm was used to train the ANN. 6 nodes and 12 nodes in the middle layers were appropriate to the temperature model for training. And 9 nodes and 6 nodes were also appropriate to the humidity model respectively. 90% of the all data was used learning set, and the extra 10% was used to model verification. In the case of temperature, average temperature before 15 minute and humidity at present constituted input layer, and temperature at present constituted out-layer and humidity model was vice versa. The sensitivity analysis revealed that previous value data contributed to forecasting target value than the other variable. Temperature was pseudo-linearly related to the previous 15 minute average value. We confirmed that ANN with multi-layer perceptron could support pollutant dispersion model by computing meterological data at real time.