• Journal of Advanced Navigation Technology
• /
• v.17 no.5
• /
• pp.516-523
• /
• 2013

In this paper, The path loss model of Air Traffic Control(ATC) telecommunication radio channel has been studied at the Incheon International Airport(IIA) with the terminal with two antlers. We measured two frequencies among VHF/UHF channel bands. The transmitting site radiated the Continuous Wave(CW). The propagation measurement was taken using the moving vehicle equipped with receiver and antenna. The transmitting power, frequency and antenna height are the same as the current operating condition. The path loss exponent and intercept parameters were extracted by the basic path loss model and hata model. The path loss exponents at passager terminal areas were 3.32 and 3.10 respectively in 128.2 MHz and 269.1 MHz. The deviation of prediction error is 9.69 and 9.65. The new path loss equation at the terminal area was also developed using the derived path loss parameters. The new path loss was compared with other models. This result will be helpful for the ATC site selection and service quality evaluation.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38C no.3
• /
• pp.263-270
• /
• 2013

Recently, the prediction and analysis technology of marine environment are actively being studied since the ocean resources in the world is taken notice. The prediction of marine disaster by automatic collecting marine environment data and analyzing the collected data can contribute to minimized the damages with respect to marine pollution of oil spill and fisheries damage by red tide blooms and marine environment upsets. However the studies of the marine environment monitoring and analysis system are limited in South Korea. In this paper, we study the marine disasters prediction system model to analyze collection marine information of out sea and near sea. This paper proposes the models for the marine disasters prediction system as communication system model, a marine environment data monitoring system model, prediction and analyzing system model, and situations propagation system model. The red tide prediction model and summarizing and analyzing model is proposed for prediction and analyzing system model.

• Journal of Advanced Navigation Technology
• /
• v.5 no.1
• /
• pp.62-73
• /
• 2001

In this paper, performance of DSRC systems is analyzed with considering the real roads and height of vehicles. The channels are modeled as 2-Ray and 4-Ray with a direct wave and reflected waves by a road and buildings in a physical layer because DSRC keeps LOS propagation characteristics, and the pass loss for each model is calculated respectively. Rician factor is obtained through the calculated path loss on two models for DSRC, and the performance of the systems is analyzed in AWGN and Rician fading channels, Impulsive noise and Rician fading channels respectively. As a result, in Rician fading channels with impulsive noise(A=0.2, ${\Gamma}^{\prime}=0.22$), BER is below $10^{-6}$ when the distance is farther than 80[m] and 40[m] in 2-Ray model and 4-Ray model respectively. For performance improvement, BCH coding scheme and MRC diversity scheme are adopted.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.8
• /
• pp.3-10
• /
• 2014

Millimeter wave (mmWave) has attracted great interest recently and the necessity of Millimeter Mobile Broadband (MMB) system has appeared based on the 4 Generation Long Term Evolution-Advanced (LTE-A) Specification. Currently, there are many studies about the mmWave communication channel. And it is subject of interest to analyze the performance in MMB channel environments. In this paper, we design the MMB system for 5th Generation mobile communication and propose channel models through the analysis of the mmWave propagation characteristics. Also, we have analyzed the performance of the MMB system of 28 GHz band in MMB channel environments.

• Bulletin of the Korean Chemical Society
• /
• v.26 no.12
• /
• pp.2007-2016
• /
• 2005

An artificial neural network (ANN) is successfully presented for prediction acidity constant (pKa) of various benzoic acids and phenols with diverse chemical structures using a nonlinear quantitative structure-property relationship. A three-layered feed forward ANN with back-propagation of error was generated using six molecular descriptors appearing in the multi-parameter linear regression (MLR) model. The polarizability term $(\pi_1)$, most positive charge of acidic hydrogen atom $(q^+)$, molecular weight (MW), most negative charge of the acidic oxygen atom $(q^-)$, the hydrogen-bond accepting ability $(\epsilon_B)$ and partial charge weighted topological electronic (PCWTE) descriptors are inputs and its output is pKa. It was found that properly selected and trained neural network with 205 compounds could fairly represent dependence of the acidity constant on molecular descriptors. For evaluation of the predictive power of the generated ANN, an optimized network was applied for prediction pKa values of 37 compounds in the prediction set, which were not used in the optimization procedure. Squared correlation coefficient $(R^2)$ and root mean square error (RMSE) of 0.9147 and 0.9388 for prediction set by the MLR model should be compared with the values of 0.9939 and 0.2575 by the ANN model. These improvements are due to the fact that acidity constant of benzoic acids and phenols in water shows nonlinear correlations with the molecular descriptors.

• Journal of electromagnetic engineering and science
• /
• v.18 no.1
• /
• pp.63-69
• /
• 2018

This work demonstrates attenuation effects of plasma on waves propagating in the 26.5-40 GHz range. The effect is investigated via experiments measuring the transmission between two Ka-band horn antennas set 30 cm apart. A dielectric-barrier-discharge (DBD) plasma generator with a size of $200mm{\times}100mm{\times}70mm$ and consisting of 20 layers of electrodes is placed between the two antennas. The DBD generator is placed in a $400mm{\times}300mm{\times}400mm$ acrylic chamber so that the experiments can be performed for plasma generated under various conditions of gas and pressure, for instance, in air, Ar, and He environments at 0.001, 0.05, and 1 atm of pressure. Attenuation is calculated by the difference in the transmission level, with and without plasma, which is generated with a bias voltage of 20 kV in the 0.1-1.4 kHz range. Results show that the attenuation varies from 0.05 dB/m to 9.0 dB/m depending on the environment. Noble gas environments show higher levels of attenuation than air, and He is lossier than Ar. In all gas environments, attenuation increases as pressure increases. Finally, electromagnetic models of plasmas generated in various conditions are provided.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.63 no.7
• /
• pp.929-933
• /
• 2014

The measurement of joint angle is important to evaluate the patient's disability. The modified fiber tip and light propagation of the developed fiber sensor were investigated to increase the range of angle detection. Different shapes of fiber tips were manufactured with a polishing machine to deliver light signal in various patterns. Output signals were analyzed to obtain joint angle change with inverse polynomial models. The measured joint angles were displayed with LabVIEW program and the reliability was tested by comparing with a commercial angle sensor. This method can be used in rehabilitation field to determine patient's progress.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.35 no.7
• /
• pp.677-683
• /
• 2011

Until now, several empirical models for assessing the damage due to TNT explosions have been proposed. A set of analytical solutions for the time-dependent radius of an expanding fire-ball after detonation of TNT was obtained by solving the continuity, Euler (momentum), and energy equations with a "polytrope" assumption at the fire-ball center. The shock waves developed from the rapid expansion of a fire-ball under water were obtained by using the KirkwoodBBethe hypothesis. The calculated period of bubble oscillation and the maximum radius of the bubble resulting from the fire-ball due to a violent underwater TNT explosion were in good agreement with the experimental data.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.3 no.4 s.11
• /
• pp.119-129
• /
• 2003

Concrete and steel are essential structural materials in the construction. But, concrete, different from steel, consists of many materials and is affected by many factors such as properties of materials, site environmental situations, and skill of constructors. Concrete have two kinds of properties, immediately knowing properties such as slump, air contents and time dependent one like strength. Therefore, concrete mixes depend on experiences of experts. However, at point of time using High Performance Concrete, new method is wanted because of more ingredients like mineral and chemical admixtures and lack of data. Artificial Neural Networks(ANN) are a mimic models of human brain to solve a complex nonlinear problem. They are powerful pattern recognizers and classifiers, also their computing abilities have been proven in the fields of prediction, estimation and pattern recognition. Here, among them, the back propagation network and radial basis function network ate used. Compositions of high-performance concrete mixes are eight components(water, cement, fine aggregate, coarse aggregate, fly ash, silica fume, superplasticizer and air-entrainer). Compressive strength, slump, and air contents are measured. The results show that neural networks are proper tools to minimize the uncertainties of the design of concrete mixtures.