• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.5
• /
• pp.552-559
• /
• 2005

Recurrent neural networks(RNNs) trained with gradient based such as real time recurrent learning(RTRL) has a drawback of slor convergence rate. This algorithm also needs the derivative calculation which is not trivialized in error back propagation process. In this paper a derivative free Kalman filter, so called the unscented Kalman filter(UKF), for training a fully connected RNN is presented in a state space formulation of the system. A derivative free Kalman filler learning algorithm makes the RNN have fast convergence speed and good tracking performance without the derivative computation. Through experiments of nonlinear channel equalization, performance of the RNNs with a derivative free Kalman filter teaming algorithm is evaluated.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2000.04b
• /
• pp.371-378
• /
• 2000

In this study an adaptive node generation procedure in the Element-free Galerkin (EFG) method using bubble-meshing technique is proposed. Based on the error function that obtained by projected error estimation method, the initial node arrangement is defined along the background cell that is used in the numerical integration. To obtain the smooth nodal configuration, the nodal configuration are regenerated by bubble-meshing technique. This bubble meshing technique was originally developed to generate a set of well-shaped triangles and tetrahedra. Its basic idea is packing circles or spheres, called bubble, into the specified area or space naturally using some dynamic equations with attracting and repelling force. To demonstrate the performance of proposed scheme, the convergence behaviors are investigated for several problems.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.2
• /
• pp.43.3-43.3
• /
• 2019

A coronal mass ejection (CME) mass is generally estimated by the total brightness measured from white-light coronagraph observations. The total brightness are determined from the integration of the Thomson scattering by free electrons of solar corona along the line of sight. It is difficult to estimate the masses of halo CMEs due to the projection effect. To solve this issue, we construct a synthetic halo CME with a power-law density distribution (ρ = ρ₀r^-3) based on a full ice-cream cone model using SOHO/LASCO C3 observations. Then we compute a conversion factor from observed CME mass to CME mass for each CME. The final CME mass is determined as their average value of several CME masses above 10 solar radii. Our preliminary analysis for six CMEs show that their CME mass are well determined within the mean absolute relative error in the range of 4 to 15 %.

• Journal of Korea Multimedia Society
• /
• v.19 no.6
• /
• pp.1003-1013
• /
• 2016

This paper proposes a framework of superpixel-based vehicle detection method using plane normal vector in disparity space. We utilize two common factors for detecting vehicles: Hypothesis Generation (HG) and Hypothesis Verification (HV). At the stage of HG, we set the regions of interest (ROI) by estimating the lane, and track them to reduce computational cost of the overall processes. The image is then divided into compact superpixels, each of which is viewed as a plane composed of the normal vector in disparity space. After that, the representative normal vector is computed at a superpixel-level, which alleviates the well-known problems of conventional color-based and depth-based approaches. Based on the assumption that the central-bottom of the input image is always on the navigable region, the road and obstacle candidates are simultaneously extracted by the plane normal vectors obtained from K-means algorithm. At the stage of HV, the separated obstacle candidates are verified by employing HOG and SVM as for a feature and classifying function, respectively. To achieve this, we trained SVM classifier by HOG features of KITTI training dataset. The experimental results demonstrate that the proposed vehicle detection system outperforms the conventional HOG-based methods qualitatively and quantitatively.

• Korean Journal of Soil Science and Fertilizer
• /
• v.29 no.2
• /
• pp.92-106
• /
• 1996

Imaginary grass field was assumed suitable as the representative one for simplified estimation of local drought, and a moisture balance booking model computing drought was developed with the limited numbers of its determining factors, such as crop coefficient of the field, reservoir capacity of the soil, and the beginning point of drought as defined by soil moisture status. The maximum effective rainfall was assumed to be the same as the available free space of soil reservoir capacity. The model is similar to a definite depth evaporation pan, which stores rainfall as much as the available free space on the water in it and consumes the water by evaporation. When the pan keeps water less than a certain defined level, it is droughty. The model simulates soil moisture deficit on the assumed grass field for the drought estimation. The model can assess the water requirement, drought intensity, and the index of yield decrement due to drought. The influencing intensity indices of the selected factors were 100, 21, and 16 respectively for crop coefficient, reservoir capacity, and drought beginning point, determined by the annual water requirements as influenced by them in the model. The optimum values of the selected factors for the model were respectively 58% for crop coefficient defined on the energy indicator scale of the small copper pan evaporation, 50 mm for reservoir capacity on the basis of the average of experimentally determined values for sandy loam, loam, clay loam, and clay soils, and 65% of the reservoir capacity for the beginning point of drought.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.35 no.10
• /
• pp.859-867
• /
• 2007

The objective of this study is to develop and validate a numerical method which can handle the multi-rotor aerodynamic characteristics. For the purpose of power estimation, table look-up method is implemented to the existing unsteady panel code that is coupled with a time-marching free wake model. Also, the Reynolds number scaling is implemented for the application to various regions of Reynolds number. The computed results are validated against the available experimental data for coaxial and tandem rotors. In the validation case for the coaxial rotor, more accurate result is acquired when the thickness effect is considered. The wake instability problem occurs at a particular separation distance between the rotors for tandem rotors. The wake instability is avoided by setting the single-rotor wake geometry as the initial wake geometry for the multi-rotor analysis. The estimated result for rotor separation effect is compared with the result of the momentum theory.

• Proceedings of the Korean Statistical Society Conference
• /
• 2004.11a
• /
• pp.67-72
• /
• 2004

Support vector machine (SVM) has been very successful in pattern recognition and function estimation problems for crisp data. This paper proposes a new method to evaluate interval linear and nonlinear regression models combining the possibility and necessity estimation formulation with the principle of SVM. For data sets with crisp inputs and interval outputs, the possibility and necessity models have been recently utilized, which are based on quadratic programming approach giving more diverse spread coefficients than a linear programming one. SVM also uses quadratic programming approach whose another advantage in interval regression analysis is to be able to integrate both the property of central tendency in least squares and the possibilistic property In fuzzy regression. However this is not a computationally expensive way. SVM allows us to perform interval nonlinear regression analysis by constructing an interval linear regression function in a high dimensional feature space. In particular, SVM is a very attractive approach to model nonlinear interval data. The proposed algorithm here is model-free method in the sense that we do not have to assume the underlying model function for interval nonlinear regression model with crisp inputs and interval output. Experimental results are then presented which indicate the performance of this algorithm.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.7
• /
• pp.935-941
• /
• 2020

People estimation is important to provide IoT services. Most people counting technologies use camera or sensor data. However, the conventional technologies have the disadvantages of invasion of privacy and the need to install extra infrastructure. This paper proposes a method for estimating the number of people using a Wi-Fi AP. We use channel state information of Wi-Fi and analyze that using deep learning technology. It can be achieved by pre-installed Wi-Fi infrastructure that reduce cost for people estimation and privacy infringement. The proposed algorithm uses a k-binding data for pre-processing process and a 1D-CNN learning model. Two APs were installed to analyze the estimation results of six people. The result of the accurate number estimation was 64.8%, but the result of classifying the number of people into classes showed a high result of 84.5%. This algorithm is expected to be applicable to estimate the density of people in a small space.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.142-145
• /
• 2005

Since the successful launch of FORMOSAT -2 satellite by National Space Organization of Taiwan in May 2004, the Remote Sensing Instrument (RSI) on- board the FORMOSAT -2 has continuously acquired images at one panchromatic and four multi-spectral bands (http://www.nspo.org.tw). In general, the RSI performs well and receives high quality images which proved to be very useful for various applications. However, some RSI panchromatic products exhibit obvious trembling artifact that must be removed. Preliminary study reveals that the trembling artifact is caused by the instability of the spacecraft attitude. Though the magnitude of this artifact is actually less than half of a pixel, it affects the applicability of panchromatic products. A procedure removing this artifact is therefore needed for providing image products of consistent quality. Due to the nature of trembling artifact, it is impossible to describe the trembling amount by employing an analytic model. Relied only on image itself, an algorithm determining trembling amount and removing accordingly the trembling artifact is proposed. The algorithm consists of 3 stages. First, a cross-correlation based scheme is used to measure the relative shift between adjacent scan lines. Follows, the trembling amount is estimated from the measured value. For this purpose, the Fourier transform is utilized to characterize random shifts in frequency domain. An adaptive estimation method is then applied to deduce the approximate trembling amount. In the subsequent stage, image re-sampling operation is applied to restore the trembling-free product. Experimental results show that by applying the proposed algorithm, the unpleasant trembling artifact is no longer evident.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39A no.8
• /
• pp.437-444
• /
• 2014

In this paper, we propose a path loss model with the multiple antennas and doppler shift for high speed railroad communication. Path loss model is very important in order to design consider diverse characteristic in high-speed train communication. Currently wireless communication systems use the multiple antennas in order to improve the channel capacity or diversity gain. However, until recently, many researches on path loss model only consider geographical environment between the transmitter and the receiver. There is no study about path loss model considering diversity effect and doppler shift. In order to make average residuals considering doppler shift we use tuned free space path loss model which is utilized for measurement results at high speed railroad. The environment of high speed rail is mostly at viaduct and flatland over than 50 percent. And in order to make average residuals considering multiple antenna we use theoretical estimation of diversity gain with MRC scheme. proposed model predict loss of received signal by estimating average residuals between diversity effect and doppler shift.