생활체육의 관리운영 시스템 탐색

하니 영남대학교 스포츠과학대학원 2009 국내석사

M.A. Thesis Research on the Administrative and Operational System of Sports for All Ny Ha Sports Science Major Graduate School of Sports Science Yeungnam University (Supervised by Professor Dong-Kyu Kim) ABSTRACT The improved standard of living and increased leisure time brought about the development of sports for all. According to the change of the era, the necessity of establishment of efficient and systematical administration and operation has been raised for national health and welfare improvement. This study defined the characteristics and form of the sports for all, reviewed the system and characteristics of the sports for all in advanced countries and considered the sports for all in Korea. The current administration and operation system into organization, finance, personnel, facility, program and event were also analyzed. The results of study showed followings; First, the establishment of infra like policy support or facility installation intensifies the capacity of the government and its major participant should be private section to stabilize the structure of sports for all and to operate advanced and autonomous sports for all. Second, as long as various events and participation in program are voluntary and self-regulating, the sport club activity should be administrated and operated spontaneously under the principle of self-pay. Third, multi-purpose sports activity space fit for the local feature and the public neighboring sports facility within the residential district should be developed or expanded, and the level of use and service of public sports facility should be raised. Fourth, it is required to cultivate and re-educate the coachers who lead the sports activity in the field of sports for all and trigger the participants' quality and motivation. Lastly, the structural problem in distributing local sports for all program should be improved, the programs that are fit for local circumstances and residents' desire should be distributed, and the sports for all activities in various and new form should be developed.

Machine Learning Techniques for Trajectory Extraction and Similarity Search Indexing

라마단, 하니 부산대학교 대학원 2022 국내박사

Performing analysis tasks on trajectory data can be challenging due to the different nature of indoor and outdoor trajectories. We can discover interesting common movement patterns from indoor and outdoor trajectory in a certain time interval, for example a common movement pattern of going from the art to history to literature areas when some people visit a museum; whereas common movement pattern of the vehicles from office area to market area. Some location-based services, such as COVID-19 people tracing and real-time next-to-visit recommendations, gain benefits from the common movement patterns. However, analyzing indoor trajectories is hard due to noisy indoor signal observations, as we cannot accurately acquire the signals of Global Positioning System (GPS) inside a building. On the other hand, the massive outdoor trajectory also hinders the efficiency of the trajectory analysis due to the huge number of the data source: nearly six billion smartphone users in the world. To solve the challenges from the two types of trajectory, machine learning techniques can handle the cases of noisy indoor trajectories and massive outdoor trajectories. First, we developed a machine learning-based indoor positioning technique with valid movement constraints to reduce the incorrect positioning results in extracting semantic trajectories from the noisy indoor signal observations. The constraints prevent the impossible movement from distant positions in a short time, thus improving the correctness of the extracted trajectories. Unlike the conventional indoor 2D positioning techniques, we directly work viii on semantic positions, thus we re-purpose the positioning use case as classification task in machine learning. The semantic trajectory serves as the base of mining the common patterns from several trajectories. We apply the movement constraints to three type of models: generic classification models, K-Nearest neighbor (KNN), and Hidden Markov Model (HMM) and compare their performance to several approaches in four different datasets of indoor trajectories. The experimental results demonstrate that the constrained machine learning model improved the correctness of the indoor semantic trajectory extraction by 25-80%. On the other hand, the results also show that the constraints improve the efficiency of the KNN and HMM techniques over the original version. Moreover, we also demonstrate that our proposed approach is applicable in real-world practice of real-time setting and various crowd setting. Second, we devised X-FIST, a framework of Flood learned index extensions for mining similar trajectories in a massive trajectory dataset. Unlike the original Flood, we extend the learned indices to represent Minimum Bounding Rectangles of the sub-trajectories instead of the points, which can lead to inefficient search. The framework improves the efficacy of the dissimilar trajectory pruning thanks to the effective indexing of the skewed distributed datasets. The skewed data distribution often occurs in trajectory data and can cause the conventional tree-based index to prune ineffectively. We evaluate the performance of X-FIST on three real-world datasets and compare it to a number of other learned and tree indices. The results on the index construction demonstrate that X-FIST has 30% smaller index storage size and 6× shorter building times. Additionally, the experiment on the similarity search task exhibited that X-FIST at least speed up the search time by 2 times from the other approaches. We also show that X-FIST does not need retraining when the measure of the similarity search changes. Hence, we can conclude that our machine learning techniques demonstrated the improvement of analysis tasks on the noisy and massive trajectory data.

Effects of visible light on microbiological water quality and biofilm formation in plastic rainwater storage tanks

보니하니 서울대학교 대학원 2021 국내석사

불충분한 상수도 시스템과 인구 증가로 인하여 물부족이 발생하는데, 빗물 수집이 잠재적 해결책으로 각광을 받고 있다. 하지만 저장시설의 비용이 비싸기 때문에 개발도상국에서는 임시방편으로 간단한 빗물 수집 시스템을 사용한다. 각 가정에서는 불투명하고 두꺼운 빗물저장조 대신에 재활용된 반투명의 작은 플라스틱 재질을 사용한다. 플라스틱 저장조는 태양빛을 받으면 가시광선을 반사, 전달 및 흡수한다. 플라스틱 저장조의 이러한 특성은 저장된 물에서의 세균의 생장과 미생물에 영향을 미칠 수 있다. 검은색 플라스틱 저장조는 가시광선을 흡수하여 열로 변환되어 수온을 높이며, 결국 부유성 세균(Suspended Bacteria) 증식을 촉진한다. 얇은 플라스틱 저장조는 색상에 따라 가시광선 파장의 영역을 투과시킨다. 적색의 저장조는 630nm 파장을 전송할 수 있으며, 이 파장은 강한 살균효과를 가진다. 파란색 저장조는 reactive oxygen species (ROS)를 생성시켜 세포에 해를 주어 부유성 세균의 증식을 억제하거나 사멸시킨다. 반면 흰색의 저장조는 가시광선 전체 파장의 영역을 투과시킨다. 본 연구에서는 일반적인 가정용 플라스틱 탱크에서 가시광선이 미생물학적 수질에 미치는 영향을 연구하고, 저장조의 색갈에 따른 미생물학적 수질과 미생물막의 형성에 미치는 영향을 연구하였다. 첫째, 작은 아크릴 저장조를 이용하여 가시광선의 유무에 따른 부유성세군과 부착성 세균의 성장에 미치는 영향을 알아보기 위하여 빗물을 넣고 3개월간 실험을 하였다. 하나는 태양에 노출시키고 (TES), 또 하나는 태양에 노출시키지 않았다 (TNES). 부유성 세균의 숫자는 노출되었을 경우보다노출되지 않았을 때 더빨리 감소하였다. 이것은 햣빛에 노출시키지 않았을 때 미생물학적으로 수질이 더 좋아지는 것을 나타낸다. 부착성 미생물은 가시광선에 노출되었을 때 안정화 되었지만, 어두운 조건에서는 감소하였다. 둘째, 저장조의 색갈이 미생물학적 수질과 바이오 필름 형성에 미치는 영향을 평가하기 위해 다섯 개의 1리터짜리 원통형의 저장조에 흰색, 빨간색, 노란색, 파란색, 검은색으로 색칠을 하였다. 각 저장고에 빗물을 채워 6주 동안 파란색 소형 형광등 밑에 설치하였다. 부유 세균과 부착 세균의 농도는 HPC (Heterotrophic Plate Count) 방법으로 측정하였으며 표면에 부착한 세균 구조의 총 바이오매스 양은 크리스탈 바이올렛 분석으로 측정했다. 빨간색 빛은 세균 성장을 억제하는 것으로 나타났다. 하지만 흰색, 노란색, 파란색, 그리고 검은색 빛은 부유성 세균의 성장률을 자극했다. 3주 후에 세균이 감소했고, 실험 종료 시점에서 빨간 저장고에서 세균 수가 가장 적었다. 게다가, 투과된 빛의 스펙트럼은 또한 표면에 바이오 필름의 개발에 영향을 미쳤다. 빨간색과 파란색 빛은 세포 외 기질에서 (extracellular polymeric substances) 있는 스트레스 인한 단백질을 유발했다. 단백질이 풍부한 EPS는 보호 계층이므로 표면에 부착한 세균의 수는 표면에 더 오래 남았다. 흰색과 노란색의 저장고에서는 EPS에 통합된 종속영양 생물은 감소했지만, 광합성 생물은 증가하였다. 검정 저장고에서는 바이오 필름의 형성이 낮았다. 결론적으로 플라스틱 저장고의 색의 우선순위는 빨강, 검정, 파랑, 노랑, 흰색 순이며 흰색이 가장 나쁘다. 본 연구에서 알아낸 내용은 플라스틱 빗물 저장조안에서 가시광선의 존재 여부는 미생물학적인 수질과 생물막의 형성에 영향을 미친다는 것이다. 또한 저장조의 색상은 저장된 빗물의 질을 유지하는 데 커다란 영향을 주기 때문에 색상 선택에 있어서 신중하게 고려되어야 할 것이다. Insufficient water supply systems and population growth lead to water shortages, and rainwater collection is in the spotlight as a potential solution. However, because storage facilities are expensive, developing countries use a simple makeshift rainwater collection system. Each household uses a small recycled translucent plastic material instead of an opaque thick rainwater reservoir. The plastic reservoir reflects, transmits, and absorbs visible light under sunlight. These properties of plastic reservoirs can affect bacterial growth and microorganisms in stored water. The black plastic reservoir absorbs visible light and converts it into heat, increasing the water temperature, which in turn promotes the proliferation of Suspended Bacteria (SB). The thin plastic reservoir lets the visible light wavelength to pass through according to color. The red reservoir can transmit a wavelength of 630 nm, which has a strong sterilization effect. The blue reservoir (425nm and 470 nm) produces reactive oxygen specifications (ROS) which harms cells, inhibiting or killing the growth of floating bacteria. On the other hand, the white reservoir penetrates the entire wavelength of visible light. In this study, the effects of visible light on microbiological water quality in typical household plastic tanks were studied, and the microbiological water quality and biofilm formation due to the coloration of reservoirs were studied. First, a small acrylic reservoir was used to test the effect of visible light on the growth of suspended and surface-attached bacteria (SAB), and experiments were conducted for three months one exposed to the sun (TES) and the other not to the sun (TNES). The number of SB decreased faster when not exposed than when exposed. This represents a better microbiological water quality without exposure to white light. SAB stabilized when exposed to visible light, but decreased in dark conditions. Second, five cylindrical reservoirs (1L) were painted white, red, yellow, blue, and black to assess the effect of the color of the storage on microbiological water quality and biofilm formation. Each reservoir was filled with rainwater and installed under a small blue fluorescent lamp for six weeks. The concentration of suspended and attached bacteria was measured by HPC (Heterotrophic Plate Count) method, and the total amount of biomass in the bacterial structure attached to the surface was measured by crystal violet analysis. Red light has been shown to inhibit bacterial growth. However, white, yellow, blue, and black light stimulated the growth rate of SB. The bacteria decreased after three weeks, and at the end of the experiment, the number of bacteria in the red storage was the lowest. Furthermore, the spectrum of light transmitted also affected the development of biofilms on the surface. Red and blue light caused stress-induced proteins in extracellular polymer substances. Since protein-rich EPS is a protective layer, the number of bacteria attached to the surface remains longer. In the white and yellow reservoirs, the EPS-integrated heterotrophic bacteria decreased, but photosynthetic organisms increased. Biofilm formation was low in black storage. In conclusion, the color priorities of the plastic reservoir are red, black, blue, yellow, and white, with the worst color being white. The study found that the presence of visible light in plastic rainwater reservoirs affects microbiological water quality and biofilm formation. Furthermore, the color of the reservoir has a significant impact on maintaining the quality of stored rainwater, which should be considered carefully in color selection.