• 인터넷정보학회논문지
• /
• 제20권5호
• /
• pp.105-111
• /
• 2019

인공지능, 빅데이터, 자율주행 등 4차 산업혁명시대를 이끄는 핵심기술은 컴퓨팅 파워의 급속한 발전과 사물인터넷에 기반한 초연결 네트워크를 통해 구현되고 서비스된다. 본 논문에서는 자율주행을 위한 기본적인 기능으로 다양한 환경에서도 정확하게 주행가능한 영역을 인식하여 추출하는 인공지능 딥러닝 모델들을 구현하고, 그 결과를 비교, 분석한다. 주행가능한 영역을 추출하는 딥러닝 모델은 영상 분할 분야에서 성능이 우수하고 자율주행 연구에서 많이 사용하는 Deep Lab V3+와 Mask R-CNN을 활용하였다. 다양한 환경에서의 주행 정보를 위해 여러 가지 날씨 조건과 주 야간 환경에서의 주행 영상 및 이미지를 제공하는 BDD 데이터셋을 학습데이터로 사용하였다. 활용한 모델들의 실험 결과, DeepLab V3+는 48.97%의 IoU를 보였으며, Mask R-CNN은 68.33%의 IoU로 더 우수한 성능을 보였다. 또한, 구현한 모델로 추출된 주행가능 영역을 이미지에 표시하여 육안으로 검사한 결과, Mask R-CNN은 83%, Deep Lab V3+는 69% 정확도로 Mask R-CNN이 Deep Lab V3+ 보다 주행가능한 영역을 추출하는 분야에서는 더 성능이 높은 것으로 확인하였다.

• 대한기계학회논문집A
• /
• 제36권8호
• /
• pp.929-934
• /
• 2012

본 연구에서는 레이저 TWB로 용접된 이종재료의 동적 특성에 대한 해석이 수행되었다. 해석프로그램은 Hyper works 10.0으로 Solver는 LS-DYNA v.971, 모델링 요소는 2D-Shell, 요소 수는 35,641개, 노드 수는 36,561개이다. 충격속도는 10 km/h이다. 상 하부의 용접선 높이에 따르는 영향을 연구하기 위하여, 하부의 길이를 300 mm와 400 mm로 하였다. 얻어진 결론은 다음과 같다. 길이 300 mm인 하부 재료 SPFC980의 변형은 가장 작고, 충돌 흡수 에너지는 가장 크다. 하부 냉연강 기준으로 TWB의 위치가 짧을수록 성능이 우수하게 나타났다. 즉, 상대적으로 상부 고강도강인 SABC1470 재료가 차지하는 비율에 따라 성능이 좌우되었다. 하부 재료 SPFH590은 큰 변형이 나타났고, 충돌성능은 SPFC980보다 현저히 떨어졌다. 따라서 충돌 성능 해석 결과에서 하부 재료 SPFC980인 길이 300 mm가 가장 우수한 것으로 나타났다.

• Journal of Ginseng Research
• /
• 제46권3호
• /
• pp.337-347
• /
• 2022

Coronavirus disease 2019 (COVID-19) is currently a pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). COVID-19 are directly associated with hyper-activation of innate immune response that excessively produce pro-inflammatory cytokines and induce cytokine storm, leading to multi-organ-failure and significant morbidity/mortality. Currently, several antiviral drugs such as Paxlovid (nirmatrelvir and ritonavir) and molnupiravir are authorized to treat mild to moderate COVID-19, however, there are still no drugs that can specifically fight against challenges of SARS-CoV-2 variants. Panax ginseng, a medicinal plant widely used for treating various conditions, might be appropriate for this need due to its anti-inflammatory/cytokine/viral activities, fewer side effects, and cost efficiency. To review Panax ginseng and its pharmacologically active-ingredients as potential phytopharmaceuticals for treating cytokine storm of COVID-19, articles that reporting its positive effects on the cytokine production were searched from academic databases. Experimental/clinical evidences for the effectiveness of Panax ginseng and its active-ingredients in preventing or mitigating cytokine storm, especially for the cascade of cytokine storm, suggest that they might be beneficial as an adjunct treatment for cytokine storm of COVID-19. This review may provide a new approach to discover specific medications using Panax ginseng to control cytokine storm of COVID-19.

• 천문학회보
• /
• 제40권1호
• /
• pp.49.2-49.2
• /
• 2015

IntraCluster Medium (ICM) located at the galaxy cluster is in the state of very hot, tenuous, magnetized, and highly ionized X-ray emitting plasmas. High temperature and low density make ICM very viscous and conductive. In addition to the high conductivity, fluctuating random plasma motions in ICM, occurring at all evolution stages, generate and amplify the magnetic fields in such viscous ionized gas. The amplified magnetic fields in reverse drive and constrain the plasma motions beyond the viscous scale through the magnetic tension. Moreover, without the influence of resistivity viscous damping effect gets balanced only with the magnetic tension in the extended viscous scale leading to peculiar ICM energy spectra. This overall collisionless magnetohydrodynamic (MHD) turbulence in ICM was simulated using a hyper diffusivity method. The results show the plasma motions and frozen magnetic fields have power law of $E_V^k{\sim}k^{-3}$, $E_M^k{\sim}k^{-1}$. To explain these abnormal power spectra we set up two simultaneous differential equations for the kinetic and magnetic energy using an Eddy Damped Quasi Normal Markovianized (EDQNM) approximation. The solutions and dimensions of leading terms in the coupled equations derive the power spectra and tell us how the spectra are formed. We also derived the same results with a more intuitive balance relation and stationary energy transport rate.

• Journal of electromagnetic engineering and science
• /
• 제10권3호
• /
• pp.132-137
• /
• 2010

We present results of room-temperature characterization of lithographically manufactured antenna-coupled NbN micro-bolometers. The bolometers are assembled together with a hyper-hemispherical Si lens to couple the incident radiation to the bolometer from the back-side of the substrate. The bolometers are designed to operate at 300~1,000 GHz and they are characterized at 321~782 GHz. Radiation patterns are measured at 321 GHz, 400 GHz, 654 GHz, and at 782 GHz. The frequency dependency of the beamwidth is studied with several azimuthal beam profile measurements at 321~500 GHz.

• Journal of Microbiology and Biotechnology
• /
• 제20권4호
• /
• pp.718-726
• /
• 2010

A selected fungal strain, for production of the raw-starchdigesting enzyme by solid-state fermentation, was improved by two repeated sequential exposures to ${\gamma}$-irradiation of $Co^{60}$, ultraviolet, and four repeated treatments with Nmethyl-N'-nitrosoguanidine. The mutant strain Aspergillus sp. XN15 was chosen after a rigorous screening process, with its production of the raw-starch-digesting enzyme being twice that of usual wild varieties cultured under preoptimized conditions and in an unsupplemented medium. After 17 successive subculturings, the enzyme production of the mutant was stable. Optimal conditions for the production of the enzyme by solid-state fermentation, using wheat bran as the substrate, were accomplished for the mutant Aspergillus sp. XN15. With the optimal fermentation conditions, and a solid medium supplemented with nitrogen sources of 1% urea and 1% $NH_4NO_3$, 2.5 mM $CoSO_4$, 0.05% (v/w) Tween 80, and 1% glucose, the mutant Aspergillus sp. XN15 produced the raw-starch-digesting enzyme in quantities 19.4 times greater than a typical wild variety. Finally, XN15, through simultaneous saccharification and fermentation of a raw rice corn starch slurry, produced a high level of ethanol with $Y_{p/s}$ of 0.47 g/g.

• 한국소음진동공학회논문집
• /
• 제19권11호
• /
• pp.1151-1157
• /
• 2009

Ensuring both product quality and reducing material cost are important issue for the design of the piping system of an air conditioner outdoor unit. This paper describes a shape optimization that achieves mass reduction of an air conditioner piping system while satisfying two design constraints on resonance avoidance and the maximum stress in the pipes. In order to obtain optimized design results with various analysis fields considered simultaneously, an automated multidisciplinary analysis system was constructed using PIAnO v.2.4, a commercial process integration and design optimization(PIDO) tool. As the first step of the automated analysis system, a finite element model is automatically generated corresponding to the specified shape of the pipes using a morphing technique included in HyperMesh. Then, the performance indices representing various design requirements (e.g. natural frequency, maximum stress and pipe mass) are obtained from the finite element analyses using appropriate computer-aided engineering(CAE) tools. A sequential approximate optimization(SAO) method was employed to effectively obtain the optimum design. As a result, the pipe mass was reduced by 18 % compared with that of an initial design while all the constraints were satisfied.

• 한국항행학회논문지
• /
• 제15권2호
• /
• pp.313-318
• /
• 2011

최근 들어서 가상화 기술이 유망기술로 인정받고 있지만, 이 기술은 이미 1960년대부터 시작된 기술이었다. 이때의 기술은 시스템을 효율적으로 사용하기 보다는 서비스에 목적을 두어 가상화가 되었고, 이후 기술이 지속적으로 발전하면서 x86 기반의 서버에서 가상화 기술이 실제적으로 적용되고 있다. 가상화 기술을 이용하는 가장 큰 이유는 경제적인 이득일 것이다. 서버들을 따로 따로 운영해야 할 경우, 전기세와 서버 구입비용 및 유지보수 비용 등이 많이 들지만, 이것을 가상화 기술로 통합하게 되면 서비스는 똑같이 할 수 있으면서 운영비(인건비 포함)가 상대적으로 절감되기 때문이다. 본 논문에서는 이러한 가상화 기술의 트렌드를 분석하고 서버통합 방안에 대한 로드맵을 제안하였다.

• 소성∙가공
• /
• 제29권4호
• /
• pp.218-228
• /
• 2020

The prediction of final mass and optimized process conditions of injection molded products using Artificial Neural Network (ANN) were demonstrated. The ANN was modeled with 10 input parameters and one output parameter (mass). The input parameters, i.e.; melt temperature, mold temperature, injection speed, packing pressure, packing time, cooling time, back pressure, plastification speed, V/P switchover, and suck back were selected. To generate training data for the ANN model, 77 experiments based on the combination of orthogonal sampling and random sampling were performed. The collected training data were normalized to eliminate scale differences between factors to improve the prediction performance of the ANN model. Grid search and random search method were used to find the optimized hyper-parameter of the ANN model. After the training of ANN model, optimized process conditions that satisfied the target mass of 41.14 g were predicted. The predicted process conditions were verified through actual injection molding experiments. Through the verification, it was found that the average deviation in the optimized conditions was 0.15±0.07 g. This value confirms that our proposed procedure can successfully predict the optimized process conditions for the target mass of injection molded products.

• Structural Monitoring and Maintenance
• /
• 제9권3호
• /
• pp.289-303
• /
• 2022

In this study, the impact of assigned pixel labels on the accuracy of crack image identification of steel structures is examined by using an atrous separable convolution neural network (ASCNN). Firstly, images containing fatigue cracks collected from steel structures are classified into four datasets by assigning different pixel labels based on image features. Secondly, the DeepLab v3+ algorithm is used to determine optimal parameters of the ASCNN model by maximizing the average mean-intersection-over-union (mIoU) metric of the datasets. Thirdly, the ASCNN model is trained for various image sizes and hyper-parameters, such as the learning rule, learning rate, and epoch. The optimal parameters of the ASCNN model are determined based on the average mIoU metric. Finally, the trained ASCNN model is evaluated by using 10% untrained images. The result shows that the ASCNN model can segment cracks and other objects in the captured images with an average mIoU of 0.716.