• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.107-119
• /
• 2023

The technology that recognizes a soldier's motion and movement status has recently attracted large attention as a combination of wearable technology and artificial intelligence, which is expected to upend the paradigm of troop management. The accuracy of state determination should be maintained at a high-end level to make sure of the expected vital functions both in a training situation; an evaluation and solution provision for each individual's motion, and in a combat situation; overall enhancement in managing troops. However, when input data is given as a timer series or sequence, existing feedforward networks would show overt limitations in maximizing classification performance. Since human behavior data (3-axis accelerations and 3-axis angular velocities) handled for military motion recognition requires the process of analyzing its time-dependent characteristics, this study proposes a high-performance data-driven classifier which utilizes the long-short term memory to identify the order dependence of acquired data, learning to classify eight representative military operations (Sitting, Standing, Walking, Running, Ascending, Descending, Low Crawl, and High Crawl). Since the accuracy is highly dependent on a network's learning conditions and variables, manual adjustment may neither be cost-effective nor guarantee optimal results during learning. Therefore, in this study, we optimized hyperparameters using Bayesian optimization for maximized generalization performance. As a result, the final architecture could reduce the error rate by 62.56% compared to the existing network with a similar number of learnable parameters, with the final accuracy of 98.39% for various military operations.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.4
• /
• pp.273-282
• /
• 2023

Probabilistic safety assessment (PSA) has been widely used to evaluate the seismic risk of nuclear power plants (NPPs). However, studies on seismic PSA for process plants, such as gas plants, oil refineries, and chemical plants, have been scarce. This is because the major disasters to which these process plants are vulnerable include explosions, fires, and release (or dispersion) of toxic chemicals. However, seismic PSA is essential for the plants located in regions with significant earthquake risks. Seismic PSA entails probabilistic seismic hazard analysis (PSHA), event tree analysis (ETA), fault tree analysis (FTA), and fragility analysis for the structures and essential equipment items. Among those analyses, ETA can depict the accident sequence for core damage, which is the worst disaster and top event concerning NPPs. However, there is no general top event with regard to process plants. Therefore, PSA cannot be directly applied to process plants. Moreover, there is a paucity of studies on developing fragility curves for various equipment. This paper introduces PSA for gas plants based on FTA, which is then transformed into Bayesian network, that is, a probabilistic graph model that can aid risk-informed decision-making. Finally, the proposed method is applied to a gas plant, and several decision-making cases are demonstrated.

• Information Systems Review
• /
• v.25 no.1
• /
• pp.129-143
• /
• 2023

In the era of big data, the field of artificial intelligence is showing remarkable growth, and in particular, the image classification learning methods by deep learning are becoming an important area. Various studies have been actively conducted to further improve the performance of CNNs, which have been widely used in image classification, among which a representative method is the Convolutional Recurrent Neural Network (CRNN) algorithm. The CRNN algorithm consists of a combination of CNN for image classification and RNNs for recognizing time series elements. However, since the inputs used in the RNN area of CRNN are the flatten values extracted by applying the convolution and pooling technique to the image, pixel values in the same phase in the image appear in different order. And this makes it difficult to properly learn the sequence of arrangements in the image intended by the RNN. Therefore, this study aims to improve image classification performance by proposing a novel hybrid method of CNN and RNN applying the concepts of encoder and decoder. In this study, the effectiveness of the new hybrid method was verified through various experiments. This study has academic implications in that it broadens the applicability of encoder and decoder concepts, and the proposed method has advantages in terms of model learning time and infrastructure construction costs as it does not significantly increase complexity compared to conventional hybrid methods. In addition, this study has practical implications in that it presents the possibility of improving the quality of services provided in various fields that require accurate image classification.

• Korean Journal of Crystallography
• /
• v.2 no.2
• /
• pp.13-21
• /
• 1991

X-ray methods have been used to determine the chain conformation and packing of the thermotropic liquid crystalline copolyester prepared from 50% tarephthaloyl chloride(TPA) and 50% (1-phenylethyl) hydroquinone(PEHQ). The x-ray patterns of annealed melt-spun fibers contain a series of annealed melt-Pointing to a well ordered crystalline structure, despite the random sense(2 or 3-) of the 1-phenylethyl substiuttion on the TPA-hydroquinone backbone. The crystalline fiber is monoclinic with space group P2l and the unit cell has dimensions 11=12.77 A, b=10.17 A (upique axis), c=12.58 h (fiber axis). and β=90.1° and contains TPA-PEHO units of to or chains. The random substitution of 1-phenylethyl groups was modelled by placing these groups at both the 2and 3 positions and giving each a weight of one-hal(. T he structure has been refined by linked a rom least square methods(LALS) against 16 observed and 21 unobserved reflections. and had a final R value of 0.20. Packing of the side chains is effected by staggering adjacent chains along the b axis by approximately c/2, so that the side chains are interleaved. The phenyl-COO and COO-phenyl torsion angles are -6.1 and 65.6, respectively, such that the main chain phenyls are mutually inclined at 59.5 (the ester groups are assumed to be planar). These torsion angles compare very well with those for the model compounds, notably phenylbenzoate, and can be used in future analyses of the structures of more complex random sequence copolyesters.

• Molecules and Cells
• /
• v.28 no.6
• /
• pp.529-536
• /
• 2009

Genome sequencing of the pig is being accelerated because of its importance as an evolutionary and biomedical model animal as well as a major livestock animal. However, information on expressed porcine genes is insufficient to allow annotation and use of the genomic information. A series of expressed sequence tags of 5' ends of five full-length enriched cDNA libraries (SUSFLECKs) were functionally characterized. SUSFLECKs were constructed from porcine abdominal fat, induced fat cells, loin muscle, liver, and pituitary gland, and were composed of non-normalized and normalized libraries. A total of 55,658 ESTs that were sequenced once from the 5′ ends of clones were produced and assembled into 17,684 unique sequences with 7,736 contigs and 9,948 singletons. In Gene Ontology analysis, two significant biological process leaf nodes were found: gluconeogenesis and translation elongation. In functional domain analysis based on the Pfam database, the beta transducin repeat domain of WD40 protein was the most frequently occurring domain. Twelve genes, including SLC25A6, EEF1G, EEF1A1, COX1, ACTA1, SLA, and ANXA2, were significantly more abundant in fat tissues than in loin muscle, liver, and pituitary gland in the SUSFLECKs. These characteristics of SUSFLECKs determined by EST analysis can provide important insight to discover the functional pathways in gene networks and to expand our understanding of energy metabolism in the pig.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.5
• /
• pp.331-339
• /
• 2016

This paper suggests an estimation method of the source signal and the channel impulse response (CIR) using ray-based blind deconvolution (RBD) in the underwater acoustic channel environment where Doppler effect exists by the relative motion between source and receiver. It is difficult to estimate the CIR on Doppler effect by the matched filter with a highly Doppler-sensitive waveform such as the m-sequence signal because Doppler shift can severely degrade the correlation between the received signal corrupted by Doppler effect and the original source signal. In this study, the Doppler-shifted source-signal's phase is estimated using the RBD, and the received signal is compensated by it to obtain the Doppler-corrected CIR. It is verified that using the matched filter with the received signal from the experimental data fails to estimate the CIR while the obtained CIR by the suggested method has the similarity to the propagation path of the ray model. Also, the results show that the reconstructed source signal using the RBD has the better Doppler shift compensation than the Doppler-shifted source signal derived from scattering function.

• Progress in Medical Physics
• /
• v.10 no.1
• /
• pp.23-32
• /
• 1999

Combined MRI and Relaxogram approach was introduced as a very useful tool for fat study. The phantoms simulating homogeneous mixture of fat and non-fat environments were measured with spin echo pulse sequence on a 0.15 T whole body imager. From 45 scans, the Tl values were obtained by fitting the data to continuous distribution (CONTIN) of relaxation time. This relaxogram gives broad distributions of relaxation time, which are characterized by a number of peaks with characteristic T1 values. Two distinct peaks in relaxogram were observed and identified as signals from com oil and gelatin gel. This model system can be served as simulating the distribution of fat in muscle. Also the relative ratio of two components, which is proportional to the area under the peak, is estimated and compared to nominal values. Based on the good agreement between two predictions, the values from our proposed method agreed with nominal values within $\pm$7 % error. The effects of different concentration of contrast agent and different region of interest are presented. To optimize total scan times, the minimum required data points and so further reduction in total scan times are discussed.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.48 no.8
• /
• pp.555-563
• /
• 2020

Foward-swept wings are known to possess superior aerodynamic performance compared to the conventional straight wings. However major concerns regarding forward-swept wings include divergence at lower airspeeds which require careful consideration at the design stage. As an endeavor to overcome such drawbacks, aeroelastic tailoring is attempted. In order to find an optimal ply sequence, recursive aeroelastic analyses is conducted and one-dimensional beam analysis coupled with simple aerodynamics is used for the improved computational efficiency and modelling convenience. The analysis used in this paper, DYMORE and analytic formula, both use one-dimensional beam model for the structure. Cross-sectional analysis for multi-cell NACA0015 airfoil section is conducted using VABS and oblique function is used for the sweep angle. Throughout the present aeroelastic tailoring, the maximum divergence speed of 290.2m/s is achieved which is increased by approximately 43% than that for the conventional ply configuration.

• Investigative Magnetic Resonance Imaging
• /
• v.2 no.1
• /
• pp.73-82
• /
• 1998

Purpose : A precise NMR technique for measuring the rate of water exchange and cell membrane permeability across the hepatocyte membrane using liver-specific MR contrast agent is described. Materials and Methods : The rat hepatocytes isolated by perfusion of the livers were used for the NMR measurements. All experiments were performed on an IBM field cycling relaxometer operating from 0.02MHz to 60 MHz proton Larmor frequency. spin-echo pulse sequence was empolyed to measure spin-lattice relaxation time, T1. The continuous distribution analysis of water proton T1 data from rat hepatocytes containing low concentrations of the liver specific contrast agent, Gd-EOB-DTPA, modeled by a general two compartment exchange model. Results : The mean residence time of water molecule inside the hepatocyte was approximately 250 msec. The lower limit for the permeability of the hepatocyte membrane was $(1.3{\pm}0.1){\;}{\times}{\;}10^{-3}cm/sec$. The CONTIN analysis, which seeks the natural distribution of relaxation times, reveals direct evidence of the effect of diffusive exchange. the diffusive water exchange is not small in the intracellular space in the case of hepatocytes. Conclusions : Gd-EOB-DTPA, when combined with continuous distribution analysis, provides a robust method to study water exchange and membrane permeability in hepatocytes. Water exchange in hepatocyte is much slower thatn that in red blood cells. Therefore, tissue-specific contrast agent may be used as a functional agent to give physiological information such as cell membrane permeability.

• Journal of the Korean Institute of Gas
• /
• v.24 no.1
• /
• pp.17-22
• /
• 2020

Light Naphtha is distillated from crude oil unit and separated into the methane, ethylene and propylene by boiling point difference in sequence. This separation is conducted using a series of binary-like columns. This separation method is known that the energy consumed in the reboiler is used to separate the heaviest components and most of this energy is discarded as vapor condensation in the overhead cooler. In this study, the first two columns of the separation process are replaced with the Petlyuk column. A structural design was exercised by the stage computation with ideal tray efficiency in equilibrium condition. Compared with the performance of a conventional system of 3-column model, The design outcome shows that the procedure is simple and efficient because the composition of the liquid component in the column tray was designed to be similar to the equilibrium distillation curve. The performance of the new process indicates that an energy saving of 12.1% is obtained and the cost savings of 44 million won per day based on gross domestic product is reduced under same total number of trays and the initial investment cost is saved.