• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.3
• /
• pp.459-466
• /
• 2006

Numerical values of thermodynamic properties such as temperature, pressure, dryness, volume, enthalpy and entropy are required in numerical analysis on evaluating the thermal performance. But the steam table itself cannot be used without modelling. From this point of view the neural network with function approximation characteristics can be an alternative. the multi-layer neural networks were made for saturated vapor region and superheated vapor region separately. For saturated vapor region the neural network consists of one input layer with 1 node, two hidden layers with 10 and 20 nodes each and one output layer with 7 nodes. For superheated vapor region it consists of one input layer with 2 nodes, two hidden layers with 15 and 25 nodes each and one output layer with 3 nodes. The proposed model gives very successful results with ${\pm}0.005%$ of percentage error for temperature, enthalpy and entropy and ${\pm}0.025%$ for pressure and specific volume. From these successful results, it is confirmed that the neural networks could be powerful method in function approximation of the steam table.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.36C no.12
• /
• pp.78-89
• /
• 1999

Previous wavelet network structures are determined by considering the relationship between wavelet windows distribution of training patterns that are transformed into time-frequency space. Because it is separated two algorithms that determines wavelet network structure and that modifies parameters of network, learning process that minimizes output error of network is executed after the network structure is determined. But this method has some weakness that training patterns must be transformed into time-frequency space by additional preprocessing and the network structure should be fixed during learning process. In this paper, we propose a new constructing method for wavelet network structure by using differences between the output and the desired response without preprocessing. Because the algorithm perform network construction and error minimizing process simultaneously, it can determine the number of hidden nodes adaptively as with the complexity of problems. In addition, the network structure is optimized by inserting new hidden nodes in the area that has maximum error and extracting hidden nodes that has no effect to the output of network. This algorithm has no constraint condition that all training patterns must be known, because it removes preprocessing procedure for training patterns and it can be applied effectively to systems that has time varying outputs.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.12B
• /
• pp.1066-1074
• /
• 2008

In this paper, we proposed an efficient algorithm using pulse signal based on group-node-contention in LR-WPAN. The purpose of IEEE 802.15.4 is low speed, low cost and low power consumption. Recently, as applications of LR-WPAN have been extended, there is a strong probability of collision as well and almost collision occurs because of hidden node problem. Moreover, if the collision continuously occurs due to hidden node collision, network performance could be decreased. Nowadays, although several papers focus on the hidden node collision, algorithms waste the channel resource if continuous collisions frequently occur. In this paper, we assume that PAN has been already formed groups, and by using pulse signal, coordinator allocates channel and orders, and then, nodes in the allocated group can compete each other. Hence, contention nodes are reduced significantly, channel wastage caused by collision is decreased, and data transmission rate is improving. Finally, this algorithm can protect the network from disruption caused by frequent collisions. Simulation shows that this algorithm can improve the performance.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.160-162
• /
• 2006

BP 알고리즘은 지역 최소점이나 고원 문제와 같은 수렴 실패문제와 학습 속도가 느리다고 알려져 있다. 이제까지 알려진 BP 알고리즘의 대체 방법들은 수렴 속도와 인자에 따른 수렴의 안정성에 대한 불균형을 해소하는데 치중했다. 기존의 전통적인 BP 알고리즘에서 발생하는 위와 같은 문제를 해결하기 위하여, 본 논문에서는 적은 용량의 저장 공간만을 요구하며 수렴이 빠르고 상대적으로 안정성이 보장되는 알고리즘을 제안한다. 이 방법은 상위연결(upper connections), 은닉층-출력층(hidden to output), 하위연결(lower connections), 입력층-은닉층(input to hidden)에 대해 개별적으로 훈련을 시키는 분리 학습방법을 적용한다.

• Journal of Biomedical Engineering Research
• /
• v.19 no.5
• /
• pp.469-476
• /
• 1998

In the case of acquisition and displaying medical Images such as ultrasonography and endoscopy on VGA monitor of PC system, image degradation of tear-drop appears through scan conversion. In this study, we compare several methods which can solve this degradation and implement the hardware system that resolves this problem in real-time with PC. It is possible to represent high quality image display and real-time processing and acquisition with specific de-interlacing device and PCI bridge on our hardware system. Image quality is improved remarkably on our hardware system. It is implemented as PC-based system, so acquiring, saving images and describing text comment on those images and PACS networking can be easily implemented.metabolism. All images were spatially normalized to MNI standard PET template and smoothed with 16mm FWHM Gaussian kernel using SPM96. Mean count in cerebral region was normalized. The VOls for 34 cerebral regions were previously defined on the standard template and 17 different counts of mirrored regions to hemispheric midline were extracted from spatially normalized images. A three-layer feed-forward error back-propagation neural network classifier with 7 input nodes and 3 output nodes was used. The network was trained to interpret metabolic patterns and produce identical diagnoses with those of expert viewers. The performance of the neural network was optimized by testing with 5~40 nodes in hidden layer. Randomly selected 40 images from each group were used to train the network and the remainders were used to test the learned network. The optimized neural network gave a maximum agreement rate of 80.3% with expert viewers. It used 20 hidden nodes and was trained for 1508 epochs. Also, neural network gave agreement rates of 75~80% with 10 or 30 nodes in hidden layer. We conclude that artificial neural network performed as well as human experts and could be potentially useful as clinical decision support tool for the localization of epileptogenic zones.

• Journal of Computing Science and Engineering
• /
• v.5 no.4
• /
• pp.324-330
• /
• 2011

To avoid collisions, IEEE 802.11 medium access control (MAC) uses predetermined inter-frame spaces and the random back-off process. However, the retransmission strategy of IEEE 802.11 MAC results in considerable time wastage. The hidden node problem is well known in wireless networks; it aggravates the consequences of time wastage for retransmission. Many collision prevention and recovery approaches have been proposed to solve the hidden node problem, but all of them have complex control overhead. In this paper, we propose a fast retransmission scheme as a recovery approach. The proposed scheme identifies collisions caused by hidden nodes and then allows retransmission without collision. Analysis and simulations show that the proposed scheme has greater throughput than request-to-send and clear-to-send (RTS/CTS) and a shorter average waiting time.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39C no.9
• /
• pp.820-827
• /
• 2014

One of the basic problems in Wireless Sensor Networks (WSNs) is the localization of the sensor nodes based on the known location of numerous anchor nodes. WSNs generally consist of a large number of sensor nodes and recording the location of each sensor nodes becomes a difficult task. On the other hand, based on the application environment, the nodes may be subject to mobility and their location changes with time. Therefore, a scheme that will autonomously estimate or calculate the position of the sensor nodes is desirable. This paper presents an intelligent localization scheme, which is an artificial neural network (ANN) based localization scheme used to estimate the position of the unknown nodes. In the proposed method, three anchors nodes are used. The mobile or deployed sensor nodes request a beacon from the anchor nodes and utilizes the received signal strength indicator (RSSI) of the beacons received. The RSSI values vary depending on the distance between the mobile and the anchor nodes. The three RSSI values are used as the input to the ANN in order to estimate the location of the sensor nodes. A feed-forward artificial neural network with back propagation method for training has been employed. An average Euclidian distance error of 0.70 m has been achieved using a ANN having 3 inputs, two hidden layers, and two outputs (x and y coordinates of the position).

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.2C
• /
• pp.31-40
• /
• 2005

we propose an enhanced self-generation supervised algorithm that by combining an ART algorithm and the delta-bar-delta method. Form the input layer to the hidden layer, ART-1 and ART-2 are used to produce nodes, respectively. A winner-take-all method is adopted to the connection weight adaption so that a stored pattern for some pattern is updated. we test the recognition of student identification, a certificate of residence, and an identifier from container that require nodes of hidden layers in neural network. In simulation results, the proposed self-generation supervised learning algorithm reduces the possibility of local minima and improves learning speed and paralysis than conventional neural networks.

• Journal of KIISE:Software and Applications
• /
• v.29 no.5
• /
• pp.289-294
• /
• 2002

It is known that restructuring feed-forward neural network affects generalization capability and efficiency of the network. In this paper, we introduce a new approach to restructure a neural network using abstraction of the hidden knowledge that the network has teamed. This method involves extracting local rules from non-input nodes and aggregation of the rules into global rule base. The extracted local rules are used for pruning unnecessary connections of local nodes and the aggregation eliminates any possible redundancies arid inconsistencies among local rule-based structures. Final network is generated by the global rule-based structure. Complexity of the final network is much reduced, compared to a fully-connected neural network and generalization capability is improved. Empirical results are also shown.

• Structural Engineering and Mechanics
• /
• v.26 no.3
• /
• pp.251-262
• /
• 2007

This paper examines the application of artificial neural networks (ANN) to the response prediction of geometrically nonlinear truss structures. Two types of analysis (deterministic and probabilistic analyses) are considered. A three-layer feed-forward backpropagation network with three input nodes, five hidden layer nodes and two output nodes is firstly developed for the deterministic response analysis. Then a back propagation training algorithm with Bayesian regularization is used to train the network. The trained network is then successfully combined with a direct Monte Carlo Simulation (MCS) to perform a probabilistic response analysis of geometrically nonlinear truss structures. Finally, the proposed ANN is applied to predict the response of a geometrically nonlinear truss structure. It is found that the proposed ANN is very efficient and reasonable in predicting the response of geometrically nonlinear truss structures.