• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 1999.03a
• /
• pp.261-270
• /
• 1999

Sintering transforms powdered ore into lumped ore so that the latter can be used in a blast furnace. The powdered or combined with coke and other materials is loaded into a container and moved along by a pallet while the ignited coke burns. The speed by which the pallet moves determines how much sintering takes place. Since the process is complicated and lacks an accurate mathematical model, human operators manually control the speed by monitoring various factors in the plant. In this paper, we propose a neural network-based pallet speed controller which copies human operator knowledge. Actual process data were collected from a sintering plant for eight months and preprocessed to remove noisy and inconsistent data. A multilayer perceptron was trained using a back-propagation learning algorithm. In on-line testing at the sinter plant, the proposed model reliably controlled pallet speed during normal operation without the help of human operators. Moreover, the quality and productivity was as good as with human operators.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.410-414
• /
• 1996

A visual servoing algorithm is proposed for a robot with a camera in hand. Specifically, novel image features are suggested by employing a viewing model of perspective projection to estimate relative pitching and yawing angles between the object and the camera. To compensate dynamic characteristics of the robot, desired feature trajectories for the learning of visually guided line-of-sight robot motion are obtained by measuring features by the camera in hand not in the entire workspace, but on a single linear path along which the robot moves under the control of a, commercially provided function of linear motion. And then, control actions of the camera are approximately found by fuzzy-neural networks to follow such desired feature trajectories. To show the validity of proposed algorithm, some experimental results are illustrated, where a four axis SCARA robot with a B/W CCD camera is used.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.36C no.4
• /
• pp.87-94
• /
• 1999

This paper presents a study on the off-line handwritten Hangul (Korean) character recognition using the partially connected neural network (PCNN), which is based on partial connections between the input receptive fields and the hidden nodes. The hidden nodes of three PCNNs have ten receptive fields and different input feature sets. And we introduce modular partially connected neural network (MPCNN), The MPCNN combines three PCNNs with a merging network. The learning scheme of the proposed networks is composed of two steps: PCNN learning step and the merging step of combining three PCNN s. In the merging step, another merging PCNN network is introduced and trained by regarding the hidden output of each PCNN as a new input feature vector. The performance of the proposed classifier is verified on the recognition of 18 off-line handwritten Hangul characters widely used in business cards in Korea.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.2
• /
• pp.586-595
• /
• 1997

We proposea a more inproved alperithm which can reconstruct the berrer 3D terrains from cintour line data usong the fractals and the Neural Networks and which is an improvement based on that in[1, 2, 3]with the con-sideration on neighboring patch.We have learned the feature data in addition to reflecththe charateristics of complicated toprgraphy, and have implemented on mountainous and flatness topography using the proposed learning pattern by the reduced average error.The results of implements reprsented that the mountainous top-ography is better than that of fltness on the similarity and the visuality.

• KIEE International Transactions on Electrophysics and Applications
• /
• v.5C no.3
• /
• pp.138-142
• /
• 2005

In this paper, we compared recognition rates between NN(neural networks) and clustering method as a scheme of off-line PD(partial discharge) diagnosis which occurs at the stator coil of traction motor. To acquire PD data, three defective models are made. PD data for classification were acquired from PD detector. And then statistical distributions are calculated to classify model discharge sources. These statistical distributions were applied as input data of two classification tools, BP(Back propagation algorithm) and ANFIS(adaptive network based fuzzy inference system) pre-processed FCM(fuzzy c-means) clustering method. So, classification rate of BP were somewhat higher than ANFIS. But other items of ANFIS were better than BP; learning time, parameter number, simplicity of algorithm.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.8 no.1
• /
• pp.82-86
• /
• 2008

The article presents ARTMAP and Fuzzy BK-Product approach underwater obstacle avoidance for the Autonomous underwater Vehicles (AUV). The AUV moves an unstructured area of underwater and could be met with obstacles in its way. The AUVs are equipped with complex sensorial systems like camera, aquatic sonar system, and transducers. A Neural integrated Fuzzy BK-Product controller, which integrates Fuzzy logic representation of the human thinking procedure with the learning capabilities of neural-networks (ARTMAP), is developed for obstacle avoidance in the case of unstructured areas. In this paper, ARTMAP-Fuzzy BK-Product controller architecture comprises of two distinct elements, are 1) Fuzzy Logic Membership Function and 2) Feed-Forward ART component. Feed-Forward ART component is used to understanding the unstructured underwater environment and Fuzzy BK-Product interpolates the Fuzzy rule set and after the defuzzyfication, the output is used to take the decision for safety direction to go for avoiding the obstacle collision with the AUV. An on-line reinforcement learning method is introduced which adapts the performance of the fuzzy units continuously to any changes in the environment and make decision for the optimal path from source to destination.

• Journal of Institute of Control, Robotics and Systems
• /
• v.5 no.5
• /
• pp.550-557
• /
• 1999

This paper describes a road-following controller using the proposed neural network for autonomous vehicle. Road-following with visual sensor like camera requires intelligent control algorithm because analysis of relation from road image to steering control is complex. The proposed neural network, relative similarity modular network(RSMN), is composed of some learning networks and a partitioniing network. The partitioning network divides input space into multiple sections by similarity of input data. Because divided section has simlar input patterns, RSMN can learn nonlinear relation such as road-following with visual control easily. Visual control uses two criteria on road image from camera; one is position of vanishing point of road, the other is slope of vanishing line of road. The controller using neural network has input of two criteria and output of steering angle. To confirm performance of the proposed neural network controller, a software is developed to simulate vehicle dynamics, camera image generation, visual control, and road-following. Also, prototype autonomous electric vehicle is developed, and usefulness of the controller is verified by physical driving test.

• Journal of IKEEE
• /
• v.2 no.1 s.2
• /
• pp.150-159
• /
• 1998

This paper described the recognition of the Off-line handwritten Hangul based on neural network using a feature extraction method. Features of Hangul can be extracted by a $5{\times}5$ window method which is the modified $3{\times}3$ mask method. These features are coded to binary patterns in order to use neural network's inputs efficiently. Hangul character is recognized by the consonant, the vertical vowel, and the horizontal vowel, separately. In order to verify the recognition rate, three different coding methods were used for neural networks. Three methods were the fixed-code method, the learned-code I method, and the learned-code II method. The result was shown that the learned-code II method was the best among three methods. The result of the learned-code II method was shown 100% recognition rate for the vertical vowel, 100% for the horizontal vowel, and 98.33% for the learned consonants and 93.75% for the new consonants.

• The Journal of the Korea Contents Association
• /
• v.10 no.3
• /
• pp.72-79
• /
• 2010

A EAARS(Electronic Attendance-Absence Recording System) is the important LSS(Learning Support System) for blending a on-line learning in the face-to-face classroom. However, the EAARS based on the smart card can not identify a real owner of the checked card. Therefore, we develop the CS(Client-Sever) system that manages the attendance-absence checking automatically, which is used the self-organizing neural network for the face recognition. A client system creates the ID file by extracting the face feature, a server system analyzes the ID file sent from client system, and performs a student identification by using the Recognized weight file saved in Database. As a result, The proposed CS EAARS shows the 92% efficiency in the CS environment that includes the various face image database of the real classroom.

• Nuclear Engineering and Technology
• /
• v.54 no.8
• /
• pp.3027-3033
• /
• 2022

Background: In this study, various types of deep-learning models for predicting in vitro radiosensitivity from gene-expression profiling were compared. Methods: The clonogenic surviving fractions at 2 Gy from previous publications and microarray gene-expression data from the National Cancer Institute-60 cell lines were used to measure the radiosensitivity. Seven different prediction models including three distinct multi-layered perceptrons (MLP), four different convolutional neural networks (CNN) were compared. Folded cross-validation was applied to train and evaluate model performance. The criteria for correct prediction were absolute error < 0.02 or relative error < 10%. The models were compared in terms of prediction accuracy, training time per epoch, training fluctuations, and required calculation resources. Results: The strength of MLP-based models was their fast initial convergence and short training time per epoch. They represented significantly different prediction accuracy depending on the model configuration. The CNN-based models showed relatively high prediction accuracy, low training fluctuations, and a relatively small increase in the memory requirement as the model deepens. Conclusion: Our findings suggest that a CNN-based model with moderate depth would be appropriate when the prediction accuracy is important, and a shallow MLP-based model can be recommended when either the training resources or time are limited.