• The Journal of Natural Sciences
• /
• v.1
• /
• pp.61-113
• /
• 1987

The increasing emphasis placed on the production of fine turf for lawns, golf courses, parks, and other recreational sites has led to many unsolved problems as to how such turf could be best established and mainteined. For this purpose, a series of experiments were conducted under con ditions of pot and field. The results obtained were as follows EXPERIMENT I. The effect of nitrogen fertilizer and clipping interval on Zoysia japonica. 1. Increasing the rate of nitrogen and frequent clipping increased tiller number of Zoysis japonica and the maximum number of tillers were obtained from 700 kg N application and freqnent clippings (10 days interval ) in October. Treatment of 350kg N with 10 days clipping interval increased tillers much more than those of 700 kgN with 20 and 30 days clipping intervals. 2. The average number of green leaves occurred during the growth period maximized by applying 700 kg N and clipping 10 days interval. 3. Increasing tiller numbers significantly decreased tops DM weight per tiller by clipping plants at interval of 10 and 20 days, irrespective of nitrogen applied, and with nil N, at the interval of 30 days. By applying 700 kg N, however, top DM weight per tiller increased as the number of tillers increased consistently. 4. The highest top DM weight was achieved from late August to early September by applying 350 and 700kgN. 5. During the growth period, differences in unders ( stolon + root ) DM weight occurred bynitrogen application were found between nil N and two applied nitrogen levels, whereas, at the same level of nitrogen applied, the increase in stolon DM weight enhanced by lengthening the clipping interval to 30 days. 6. Nitrogen efficiency to green leaves, stolon nodes and DM weight of root with high nitrogen was achieved as clipping interval was shortened. 7. By increasing fertilizer nitrogen rate applied, N content n the leaves and stems of Zoysiajaponica was increased. On the other hand, N content in root and stolon had little effect onfertilizer nitrogen, resulting in the lowest content among plant fractions. The largest content of N was recorded in leaves. Lengthening the clipping interval from 10 or 20 to 30 days tends to decrease the N content in the leaves and stems, whereas this trend did not appeared in stolon androot. 8. A positive correlations between N and K contents in tops and stolon were established andthus K content increased as N content in tops and stolon increased. Meanwhile, P content was not affected by N and clipping treatments. 9. Total soluble carbohydrate content in Zoysia japonica was largest in stolon and stem, and was reduced by increasing fertilizer nitrogen rate. Reduction in total soluble carbohydrate due to increased nitrogen rate was severer in the stolons and stems than in the leaves. 10. Increasing the rate of nitrogen applied increased the number of small and large vascular bundles in leaf blade, but shortened distance among the large vascular bundles. Shortening the clipping interval resulted in increase of the number of large vascular bundles but decrease ofdistance between large vascular bundles.EXPERIMENT II. Growth response of Zoysia japonica imposed by different plant densities. 1. Tiller numbers per unit area increased as plant density heightened. Differences in num ber between densities at higher densities than 120 D were of no significance. 2. Tiller numbers per clone attained by 110 days after transplanting were 126 at 40D,77 at 80D, 67 at 120D, 54 at 160D, and 41 at 200D. A decreasing trend of tiller numbers per clone with increasing density was noticable from 100 days after transplanting onwards. 3. During the growth period, the greatest number of green leaves per unit area were attainedin 90days after transplanting at 160D and 200D, and 100 days after transplanting at 40D, 80Dand 120D. Thus the period to reach the maximum green leaf number with the high plantdensity was likely to be earlier that with the low plant density. 4. Stolon growth up to 80 days after transplaning was relatively slow, but from 80 daysonwards, the growth quickened to range from 1.9 m/clone at 40D to 0.6m/clone at 200Din 200 days after transplanting, these followed by the stolon node produced. 5. Plant density did not affect stolon weight/clone and root weight/clone until 80 daysafter transplanting. 6. DM weight of root was heavier in the early period of growth than that of stolon, butthis trend was reversed in the late period of growth : DM weight of stolon was much higherthan that of root.EXPERIMENT Ill. Vegetative growth of Zoysia japonica and Zoysia matrella as affected by nitrogen and clipping height. 1. When no nitrogen was applied to Zoysia japonica, leaf blade which appeared during theAugust-early September period remained green for a perid of about 10 weeks and even leavesemerged in rate September lived for 42 days. However, leaf longevity did not exceed 8 weeks asnitrogen was applied. In contrast the leaf longevity of Zoysia matrella which emerged during the mid August-earlySeptember period was 11 weeks and, under the nitrogen applied, 9 weeks, indicating that thelife-spen of individual leaf of Zoysia matrella may be longer than that of Zoysia japorica. Clipping height had no effect on the leaf longevity in both grasses. 2. During the July-August period, tiller number, green leaf number and DM weightof Zoysia japonica were increased significantly with fertilizer nitrogen, but were not with twolevel of clipping height. This trend was reversed after late September ; no effect of nitrogen wasappeared. Instead, lax clipping increased tiller number, green leaf number and DM weight. Greenleaves stimulated by lax clipping resulted in the occurrance of more dead leaves in late October. 3. Among the stolons outgrown until early September, the primary stolon was not influencedby nitrogen and clipping treatments to produce only 2-3 stolons. However, 1st branch stoIon asaffected by nitrogen increased significantly, so most of stolons which occurred consisted of 1st branch stolons. 4. Until early September, stolon length obtained at nil nitrogen level was chiefly caused bythe primary stolons. By applying nitrogen, the primary stolons of Zoysia japonica waslonger than 1st branch stolons when severe clipping was involved and in turn, shorter than 1stbranch stolons when lax clipping was concerned. In Zoysia matrella, 1st branch stolons were muchlonger than the primary stolon when turf was clipped severely but in conditions of lax clippingthere was little difference in length between primary and 1st branch stolons. 5. Stolon nodes of both Zoysia japonica and Z. matrella were positively influenced by nit rogen, but no particular increase by imposing clipping height treatment was marked in Zoysiamatrella. Although the stolon of Zoysia japonica grew until late October, the growthstimulated by nitrogen was not so remarkable as to exceed that by nil N.

• The Korean Journal of Vision Science
• /
• v.20 no.4
• /
• pp.579-588
• /
• 2018

Purpose : To analyze the eyeglasses supply system for ametropic soldiers in ROK military. Methods : We investigated and analyzed the supply system of eyeglasses for the ametropic soldiers provided by the Korean military. The refractive powers and corrected visual acuity were measured for 37 ametropic soldiers who wear insert glasses for ballistic protective and gas-masks supplied by the military based on their habitual prescriptions. Full correction of refractive error was prescribed for subjects having less than 1.0 of distance visual acuity, and comparison was held for inspecting the changes in corrected visual acuity. Suggestions were provided for solving the issues regarding current supplying system, and this study investigated the applicabilities for utilizing professional optometric manpower. Results : The new glasses supplied by army for ametropic soldiers were duplicated from the glasses they worn when entering the army. The spherical equivalent refractive powers of the conventional, ballistic protective and gas-mask insert glasses supplied for 37 ametropic soldiers were $-3.47{\pm}1.69D$, $-3.52{\pm}1.66D$ and $-3.55{\pm}1.63D$, respectively, and the spherical equivalent refractive power of full corrected glasses was $-3.79{\pm}1.66D$, which showed a significant difference(p<0.05). The distant corrected visual acuity measured at high and low contrast(logMAR) of conventional, ballistic protective and gas-mask insert glasses were $0.06{\pm}0.80$, $0.21{\pm}0.82$, $0.15{\pm}0.74$, $0.34{\pm}0.89$, $0.10{\pm}0.70$ and $0.22{\pm}0.27$, respectively, while the corrected visual acuity by full corrected glasses were increased to $0.02{\pm}1.05$, $0.10{\pm}0.07$, $0.09{\pm}0.92$, $0.26{\pm}0.10$, $0.04{\pm}1.00$ and $0.19{\pm}1.00$, respectively. There was a significant difference(p<0.05) except for the case of the low contrast corrected visual acuity of the conventional and gas-mask insert glasses. The procedure for ordering, dispensing, and supplying military glasses consists of 5 steps, and it was found that approximately two weeks or more are required to supply from the initial examination. Conclusion : The procedure of supplying the military glasses showed three issues: 1) a lack of refraction for prescription system, 2) relatively long length of time required for supplying the glasses, 3) an inaccurate power of supplied glasses. In order to solve those issues, in the short term, education is necessarily required for soldiers on the measurement of the refractive powers, and in the near future, further standard procedures for prescription of glasses as well as the securement of optometric manpower are expected.

• Korean Journal of Heritage: History & Science
• /
• v.51 no.2
• /
• pp.170-203
• /
• 2018

From the period of Japanese colonialism up to the present, the researchers of archaeology and ancient history in Korea and Japan have paid much attention to the Honam area. Recently the ruins and relics of Wae lineage are often discovered at archaeological excavation sites in this region. In particular, at least 17 keyhole-shaped mounds were confirmed. The fact that three mounds were built on one site was newly revealed. Haniwa, a representative earthenware object of the Kofun period in Japan, was discovered as well. Therefore, the study of a historical meaning of archaeological materials about Wae lineage in the Honam area thus far must be reexamined. The ruins and relics of Wae lineage in the Honam area have been studied by selecting only specific cases. I identified all the ruins and relics of Wae lineage in the Honam area and analyzed the process of their change in this paper. I reviewed the relationship between Wae and Mahan, and the purpose of their negotiation based on archaeological characteristics, changing processes, and historical records on a quarterly basis. The ruins and relics of Wae lineage have increased and widely spread since the early period of the 5th century in the Honam area. This tendency continued until the late period of the 6th century. Weapons of Wae lineage were introduced and tombs in the style of Wae were built from the late 4th century to the early period of the 5th century (TG232~TK216 period). Sueki was introduced from the middle to late period of the 5th century (TK208~TK23 period). Keyhole-shaped mounds and tombs in the style of Wae were built from the late period of the 5th century to the early period of the 6th century (TK47~MT15 period). Japanese weapons were introduced from the middle to late period of the 6th century (TK10~TK209 period). In other words, the archaeological appearance is different in each quarterly period. There was an intensive diplomatic relationship between Baekje and Wae in the TG232~TK216 period. The military might be included in a mission of Wae. The materials of Wae lineage of this period are likely to be related to this. Sueki spread to the inland part of the Honam area in the TK208~TK23 period. This Sueki tends to be excavated on advantageous sites for the accumulation and distribution of supplies. The main characters of the keyhole-shaped mounds are presumed to be a group of traders which were under the control of a certain influence in the north of Kyushu. The keyhole-shaped mounds were abruptly built at some distance from mounds of the leaders in the Honam area. Additionally, there was no special influence base to which the characters of the keyhole-shaped mounds belonged in the surroundings. However, it might have been impossible for the group of traders to build the keyhole-shaped mounds without the support from the residents at all because there was a big difference in the building technology of the keyhole-shaped mounds between Japan and the Honam area. The time of building some keyhole-shaped mounds is the same or almost the same with that of the mounds built for the regional leaders. This proves a close relationship with the residents as well. What do the archaeological materials of Wae lineage which have been used and buried mean over 200 years in the Honam area? Geumgwan Gaya, which had exported iron resources to Japan, perished in the early period of the 5th century. Instead of Gaya, the Honam area might have played an important role to supply the necessary resources to Japan. We assume that the Japanese (Wae) actively worked to acquire various resources focusing on the underground resources in the Honam area.

• Korean Journal of Heritage: History & Science
• /
• v.45 no.4
• /
• pp.54-73
• /
• 2012

In this article, we examined the patterns of activities of the Sarmathians though in a humble measure, with a focus on the regions where the Sarmathian sheaths spreaded. One of the main weapons the mounted nomads like the Scythias, the Sarmathians, and the Alans used at war was a spear. Though complementary, a sword was the most convenient and appropriate weapon when fighting at a near distance, fallen from the horse to the ground. The Sarmathian swords continued the tradition of the Akinakes which the Scythias or the Persians used, but those of the Sarmathians showed some advances in terms of the easiness with which a sword was drawn out from a sheath, and the way the sheaths were worn to parts of a human body. It turns out that the Sarmathian sheaths, which were designed for the people to draw swords easily, having the sheaths attached to thighs through 4 bumps, spread extensively from Pazyryk, Altai, to South Siberia, Bactria, Parthia and Rome. The most noteworthy out of all the Sarmathian sheaths were the ones that were excavated from the 4th tomb in Tillatepe, Afghanistan which belonged to the region of Bactria. The owner of the fourth tomb of Tilla-tepe whose region was under the control of Kushan Dynasty at that time, was buried wearing Sarmathian swords, and regarded as a big shot in the region of Bactria which was also under the governance of Kushan Dynasty. The fact that the owner of the tomb wore two swords suggests that there had been active exchange between Bactria and Sarmathia. It seemed that the reason why the Sarmathians could play an important role in the exchange between the East and the West might have something to do with their role of supplying Chinese goods to Silk Road. That's why we are interested in how the copper mirrors of Han Dynasty, decoration beads like melon-type beads, crystal beads and goldring articulated beads, and the artifacts of South China which produced silks were excavated in the northern steppe route where the Sarmathians actively worked. Our study have established that the eye beads discovered in Sarmathian tomb estimated to have been built around the 1st century B.C. were reprocessed in China, and then imported to Sarmathia again. We should note the Huns as a medium between the Sarmathians and the South China which were far apart from each other. Thus gold-ring articulated beads which were spread out mainly across the South China has been discovered in the Huns' remains. On the other hand, between 2nd century B.C. and 2nd century A.D. which were main periods of the Sarmathians, it was considered that the traffic route connecting the steppe route and the South China might be West-South silk road which started from Yunnan, passed through Myanmar, Pakistan, and Afghanistan, and then went into the east of India. The West-south Silk road is presumed to have been used by nomadic tribes who wanted to get the goods from South China before the Oasis route was activated by the Han Dynasty's policy of managing the countries bordering on Western China.

• Korean Journal of Heritage: History & Science
• /
• v.43 no.1
• /
• pp.160-189
• /
• 2010

This study has been designed to grasp the present situation, shapes and meaning of the standing stones and rock pillars in the whole area of Noseong Mountain Fortress in Nonsan City which have never been academically reported yet. Accordingly, the research was carried out to grasp the spatial identity of Noseong Mt. and Noseong Mountain Fortress and the dispersion of standing stones scattered around inside and outside Noseong Mountain Fortress, while the shapes and structural characteristics of stones were investigated and analyzed focusing on Chongsuk Temple, which was considered to have the highest density of standing stones and greatest values for preservation as a cultural property. In consideration of the reference to the 'Top Sa' (tower temple) at the 'Bul Woo Jo' (Article about Buddhism Houses) of 'Shinjoong Dongguk Yeoji Seungram', theoretical existence of the temple according to surveying investigation, and the excavation records of roof tile pieces with the name of 'Gwan Eum Temple', it is presumed that there had been a Buddhist sanctum inside the fortress and it could be connected to the carved letters, 'Chongsuk Temple'. According the observation survey, the 6th place of standing stones among many other places inside the fortress shows that Chongsuk Temple appears to have the strong characteristics of artificially constructed space in consideration of the size of trees and stones, the composite trend of tree and stone composition, and trace of the adjacent well and strand and the construction of stairway leading to the stone gate. Along with the constellation of the Big Dipper carved on a rock at the same space, the stones, on which the letters of 'Shinseonam', 'Chilseongam' and 'Daejangam' were carved, including 'Chongsuksa', and the carved statue of Buddha, which was assumed to be Avalokitesvara Guan Yin, have offered clue which make it possible to infer that the space was a space for Chilseong and Mountain god(Folk Belief) that had originated from the combination of Buddhism, Taoism and folk religion. According to the actual measurement of standing stones at Chonsuk Temple, it was identified that there were big differences in height among 24 stones in total, ranging from 402~29cm and the averaged distance between each stone appeared to be 23.6cm. And the shape of stones appeared to be standing or flat, and various stones such as mountain-like stones and Buddha-like stones were placed in a special arrangement or assorted arrangement, but the direction of the stones had a consistency pointing to the west. And comparing to the trace of construction of ZEN Landscape Garden well known in the country, the three flat stones except for the standing and shaped stones appeared to have the shape of meditation statue, which is the typical formational factors of a ZEN Landscape Garden, on the basis of formational technique of stones. Among them, the flat stone facing the Buddhist saint statue, was formed by way of symbolization of three-mountain stone, which was assumed to be an offering stone for sacrificial food rather than carrying out ZEN Meditation. In consideration of the formation of standing stones at Chong-suk Temple, which was carried out in the composite stoning method based using the scalene triangle with ratio of 3:5:7 in order to seek the in-depth beauty based on the stone statues of three Buddhas where the three factors such as heaven, earth and humans are embodied in the elevated or flat formation, the stones at Chongsuk Temple and the space seemed to the trace of contracted garden construction that was formed with stones for a temple, so that could be used for ZEN meditation.

• The Journal of Korean Society for Radiation Therapy
• /
• v.31 no.1
• /
• pp.57-65
• /
• 2019

Purpose: The purpose is to clarify the effect of additional scattering ratio on the edge of the block according to the increasing block thickness with low melting point lead alloy and pure lead in electron beam therapy. Methods and materials: $10{\times}10cm^2$ Shielding blocks made of low melting point lead alloy and pure lead were fabricated to shield mold frame half of applicator. Block thickness was 3, 5, 10, 15, 20 (mm) for each material. The common irradiation conditions were set at 6 MeV energy, 300 MU / Min dose rate, gantry angle of $0^{\circ}$, and dose of 100 MU. The relative scattering ratio with increasing block thickness was measured with a parallel plate type ion chamber(Exradin P11) and phantom(RW3) by varying the position of the shielding block(cone and on the phantom), the position of the measuring point(surface ans depth of $D_{max}$), and the block material(lead alloy and pure lead). Results : When (depth of measurement / block position / block material) was (surface / applicator / pure lead), the relative value(scattering ratio) was 15.33 nC(+0.33 %), 15.28 nC(0 %), 15.08 nC(-1.31 %), 15.05 nC(-1.51 %), 15.07 nC(-1.37 %) as the block thickness increased in order of 3, 5, 10, 15, 20 (mm) respectively. When it was (surface / applicator / alloy lead), the relative value(scattering ratio) was 15.19 nC(-0.59 %), 15.25 nC(-0.20 %), 15.15 nC(-0.85 %), 14.96 nC(-2.09 %), 15.15 nC(-0.85 %) respectively. When it was (surface / phantom / pure lead), the relative value(scattering ratio) was 15.62 nC(+2.23 %), 15.59 nC(+2.03 %), 15.53 nC(+1.67 %), 15.48 nC(+1.31 %), 15.34 nC(+0.39 %) respectively. When it was (surface / phantom / alloy lead), the relative value(scattering ratio) was 15.56 nC(+1.83 %), 15.55 nC(+1.77 %), 15.51 nC(+1.51 %), 15.42 nC(+0.92 %), 15.39 nC(+0.72 %) respectively. When it was (depth of $D_{max}$ / applicator / pure lead), the relative value(scattering ratio) was 16.70 nC(-10.87 %), 16.84 nC(-10.12 %), 16.72 nC(-10.78 %), 16.88 nC(-9.93 %), 16.90 nC(-9.82 %) respectively. When it was (depth of $D_{max}$ / applicator / alloy lead), the relative value(scattering ratio) was 16.83 nC(-10.19 %), 17.12 nC(-8.64 %), 16.89 nC(-9.87 %), 16.77 nC(-10.51 %), 16.52 nC(-11.85 %) respectively. When it was (depth of $D_{max}$ / phantom / pure lead), the relative value(scattering ratio) was 17.41 nC(-7.10 %), 17.45 nC(-6.88 %), 17.34 nC(-7.47 %), 17.42 nC(-7.04 %), 17.25 nC(-7.95 %) respectively. When it was (depth of $D_{max}$ / phantom / alloy lead), the relative value(scattering ratio) was 17.45 nC(-6.88 %), 17.44 nC(-6.94 %), 17.47 nC(-6.78 %), 17.43 nC(-6.99 %), 17.35 nC(-7.42 %) respectively. Conclusions: When performing electron therapy using a shielding block, the block position should be inserted applicator rather than the patient's body surface. The block thickness should be made to the minimum appropriate shielding thickness of each corresponding using energy. Also it is useful that the treatment should be performed considering the influence of scattering dose varying with distance from the edge of block.

• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.24 no.2
• /
• pp.267-281
• /
• 2019

Energetic mesoscale eddies in the East Sea (ES) associated with strong mesoscale variability impacting circulation and environments were statistically characterized by analyzing satellite altimeter data collected during 1993-2017 and in-situ data obtained from four cruises conducted between 2015 and 2017. A total of 1,008 mesoscale eddies were detected, tracked, and identified and then classified into 27 groups characterized by mean lifetime (L, day), amplitude (H, m), radius (R, km), intensity per unit area (EI, $cm^2/s^2/km^2$), ellipticity (e), eddy kinetic energy (EKE, TJ), available potential energy (APE, TJ), and direction of movement. The center, boundary, and amplitude of mesoscale eddies identified from satellite altimeter data were compared to those from the in-situ observational data for the four cases, yielding uncertainties in the center position of 2-10 km, boundary position of 10-20 km, and amplitude of 0.6-5.9 cm. The mean L, H, R, EI, e, EKE, and APE of the ES mesoscale eddies during the total period are $95{\pm}104$ days, $3.5{\pm}1.5cm$, $39{\pm}6km$, $0.023{\pm}0.017cm^2/s^2/km^2$, $0.72{\pm}0.07$, $23{\pm}21TJ$, and $588{\pm}250TJ$, respectively. The ES mesoscale eddies tend to move following the mean surface current rather than propagating westward. The southern groups (south of the subpolar front) have a longer L, larger H, R, and higher EKE, APE; and stronger EI than those of the northern groups and tend to move a longer distance following surface currents. There are exceptions to the average characteristics, such as the quasi-stationary groups (the Wonsan Warm, Wonsan Cold, Western Japan Basin Warm, and Northern Subpolar Frontal Cold Eddy groups) and short-lived groups with a relatively larger H, higher EKE, and APE and stronger EI (the Yamato Coastal Warm, Central Yamato Warm, and Eastern Japan Basin Coastal Warm eddy groups). Small eddies in the northern ES hardly resolved using the satellite altimetry data only, were not identified here and discussed with potential over-estimations of the mean L, H, R, EI, EKE, and APE. This study suggests that the ES mesoscale eddies 1) include newly identified groups such as the Hokkaido and the Yamato Rise Warm Eddies in addition to relatively well-known groups (e.g., the Ulleung Warm and the Dok Cold Eddies); 2) have a shorter L; smaller H, R, and lower EKE; and stronger EI and higher APE than those of the global ocean, and move following surface currents rather than propagating westward; and 3) show large spatial inhomogeneity among groups.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.57-73
• /
• 2021

Maintenance and prevention of failure through anomaly detection of ICT infrastructure is becoming important. System monitoring data is multidimensional time series data. When we deal with multidimensional time series data, we have difficulty in considering both characteristics of multidimensional data and characteristics of time series data. When dealing with multidimensional data, correlation between variables should be considered. Existing methods such as probability and linear base, distance base, etc. are degraded due to limitations called the curse of dimensions. In addition, time series data is preprocessed by applying sliding window technique and time series decomposition for self-correlation analysis. These techniques are the cause of increasing the dimension of data, so it is necessary to supplement them. The anomaly detection field is an old research field, and statistical methods and regression analysis were used in the early days. Currently, there are active studies to apply machine learning and artificial neural network technology to this field. Statistically based methods are difficult to apply when data is non-homogeneous, and do not detect local outliers well. The regression analysis method compares the predictive value and the actual value after learning the regression formula based on the parametric statistics and it detects abnormality. Anomaly detection using regression analysis has the disadvantage that the performance is lowered when the model is not solid and the noise or outliers of the data are included. There is a restriction that learning data with noise or outliers should be used. The autoencoder using artificial neural networks is learned to output as similar as possible to input data. It has many advantages compared to existing probability and linear model, cluster analysis, and map learning. It can be applied to data that does not satisfy probability distribution or linear assumption. In addition, it is possible to learn non-mapping without label data for teaching. However, there is a limitation of local outlier identification of multidimensional data in anomaly detection, and there is a problem that the dimension of data is greatly increased due to the characteristics of time series data. In this study, we propose a CMAE (Conditional Multimodal Autoencoder) that enhances the performance of anomaly detection by considering local outliers and time series characteristics. First, we applied Multimodal Autoencoder (MAE) to improve the limitations of local outlier identification of multidimensional data. Multimodals are commonly used to learn different types of inputs, such as voice and image. The different modal shares the bottleneck effect of Autoencoder and it learns correlation. In addition, CAE (Conditional Autoencoder) was used to learn the characteristics of time series data effectively without increasing the dimension of data. In general, conditional input mainly uses category variables, but in this study, time was used as a condition to learn periodicity. The CMAE model proposed in this paper was verified by comparing with the Unimodal Autoencoder (UAE) and Multi-modal Autoencoder (MAE). The restoration performance of Autoencoder for 41 variables was confirmed in the proposed model and the comparison model. The restoration performance is different by variables, and the restoration is normally well operated because the loss value is small for Memory, Disk, and Network modals in all three Autoencoder models. The process modal did not show a significant difference in all three models, and the CPU modal showed excellent performance in CMAE. ROC curve was prepared for the evaluation of anomaly detection performance in the proposed model and the comparison model, and AUC, accuracy, precision, recall, and F1-score were compared. In all indicators, the performance was shown in the order of CMAE, MAE, and AE. Especially, the reproduction rate was 0.9828 for CMAE, which can be confirmed to detect almost most of the abnormalities. The accuracy of the model was also improved and 87.12%, and the F1-score was 0.8883, which is considered to be suitable for anomaly detection. In practical aspect, the proposed model has an additional advantage in addition to performance improvement. The use of techniques such as time series decomposition and sliding windows has the disadvantage of managing unnecessary procedures; and their dimensional increase can cause a decrease in the computational speed in inference.The proposed model has characteristics that are easy to apply to practical tasks such as inference speed and model management.

• Journal of Intelligence and Information Systems
• /
• v.27 no.1
• /
• pp.191-207
• /
• 2021

Up to this day, mobile communications have evolved rapidly over the decades, mainly focusing on speed-up to meet the growing data demands of 2G to 5G. And with the start of the 5G era, efforts are being made to provide such various services to customers, as IoT, V2X, robots, artificial intelligence, augmented virtual reality, and smart cities, which are expected to change the environment of our lives and industries as a whole. In a bid to provide those services, on top of high speed data, reduced latency and reliability are critical for real-time services. Thus, 5G has paved the way for service delivery through maximum speed of 20Gbps, a delay of 1ms, and a connecting device of 106/㎢ In particular, in intelligent traffic control systems and services using various vehicle-based Vehicle to X (V2X), such as traffic control, in addition to high-speed data speed, reduction of delay and reliability for real-time services are very important. 5G communication uses high frequencies of 3.5Ghz and 28Ghz. These high-frequency waves can go with high-speed thanks to their straightness while their short wavelength and small diffraction angle limit their reach to distance and prevent them from penetrating walls, causing restrictions on their use indoors. Therefore, under existing networks it's difficult to overcome these constraints. The underlying centralized SDN also has a limited capability in offering delay-sensitive services because communication with many nodes creates overload in its processing. Basically, SDN, which means a structure that separates signals from the control plane from packets in the data plane, requires control of the delay-related tree structure available in the event of an emergency during autonomous driving. In these scenarios, the network architecture that handles in-vehicle information is a major variable of delay. Since SDNs in general centralized structures are difficult to meet the desired delay level, studies on the optimal size of SDNs for information processing should be conducted. Thus, SDNs need to be separated on a certain scale and construct a new type of network, which can efficiently respond to dynamically changing traffic and provide high-quality, flexible services. Moreover, the structure of these networks is closely related to ultra-low latency, high confidence, and hyper-connectivity and should be based on a new form of split SDN rather than an existing centralized SDN structure, even in the case of the worst condition. And in these SDN structural networks, where automobiles pass through small 5G cells very quickly, the information change cycle, round trip delay (RTD), and the data processing time of SDN are highly correlated with the delay. Of these, RDT is not a significant factor because it has sufficient speed and less than 1 ms of delay, but the information change cycle and data processing time of SDN are factors that greatly affect the delay. Especially, in an emergency of self-driving environment linked to an ITS(Intelligent Traffic System) that requires low latency and high reliability, information should be transmitted and processed very quickly. That is a case in point where delay plays a very sensitive role. In this paper, we study the SDN architecture in emergencies during autonomous driving and conduct analysis through simulation of the correlation with the cell layer in which the vehicle should request relevant information according to the information flow. For simulation: As the Data Rate of 5G is high enough, we can assume the information for neighbor vehicle support to the car without errors. Furthermore, we assumed 5G small cells within 50 ~ 250 m in cell radius, and the maximum speed of the vehicle was considered as a 30km ~ 200 km/hour in order to examine the network architecture to minimize the delay.