• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.3
• /
• pp.89-100
• /
• 2023

Recently, on-device artificial intelligence (AI) solutions using mobile devices and embedded edge devices have emerged in various fields, such as computer vision, to address network traffic burdens, low-energy operations, and security problems. Although vision transformer deep learning models have outperformed conventional convolutional neural network (CNN) models in computer vision, they require more computations and parameters than CNN models. Thus, they are not directly applicable to embedded edge devices with limited hardware resources. Many researchers have proposed various model compression methods or lightweight architectures for vision transformers; however, there are only a few studies evaluating the effects of model compression techniques of vision transformers on performance. Regarding this problem, this paper presents a performance evaluation of vision transformers on embedded platforms. We investigated the behaviors of three vision transformers: DeiT, LeViT, and MobileViT. Each model performance was evaluated by accuracy and inference time on edge devices using the ImageNet dataset. We assessed the effects of the quantization method applied to the models on latency enhancement and accuracy degradation by profiling the proportion of response time occupied by major operations. In addition, we evaluated the performance of each model on GPU and EdgeTPU-based edge devices. In our experimental results, LeViT showed the best performance in CPU-based edge devices, and DeiT-small showed the highest performance improvement in GPU-based edge devices. In addition, only MobileViT models showed performance improvement on EdgeTPU. Summarizing the analysis results through profiling, the degree of performance improvement of each vision transformer model was highly dependent on the proportion of parts that could be optimized in the target edge device. In summary, to apply vision transformers to on-device AI solutions, either proper operation composition and optimizations specific to target edge devices must be considered.

• Nuclear Engineering and Technology
• /
• v.55 no.9
• /
• pp.3150-3163
• /
• 2023

To improve the performance of blanket: maximizing the tritium breeding rate (TBR) for tritium self-sufficiency, and minimizing the Dose of backplate for radiation protection, most previous studies are based on manual corrections to adjust the blanket structure to achieve optimization design, but it is difficult to find an optimal structure and tends to be trapped by local optimizations as it involves multiphysics field design, which is also inefficient and time-consuming process. The artificial intelligence (AI) maybe is a potential method for the optimization design of the blanket. So, this paper aims to develop an intelligent optimization method based on an improved multi-objective NSGA-III algorithm and an adaptive BP neural network to solve these problems mentioned above. This method has been applied on optimizing the radial arrangement of a conceptual design of CFETR HCSB blanket. Finally, a series of optimal radial arrangements are obtained under the constraints that the temperature of each component of the blanket does not exceed the limit and the radial length remains unchanged, the efficiency of the blanket optimization design is significantly improved. This study will provide a clue and inspiration for the application of artificial intelligence technology in the optimization design of blanket.

• Nuclear Engineering and Technology
• /
• v.55 no.4
• /
• pp.1382-1399
• /
• 2023

Pressure relief valve (PRV) is one of the important control valves used in nuclear power plants, and its sealing performance is crucial to ensure the safety and function of the entire pressure system. For the sealing performance improving purpose, an explicit function that accounts for all design parameters and can accurately describe the relationship between the multi-design parameters and the seal performance is essential, which is also the challenge of the valve seal design and/or optimization work. On this basis, a surrogate model-based design optimization is carried out in this paper. To obtain the basic data required by the surrogate model, both the Finite Element Model (FEM) and the Computational Fluid Dynamics (CFD) based numerical models were successively established, and thereby both the contact stresses of valve static sealing and dynamic impact (between valve disk and nozzle) could be predicted. With these basic data, the polynomial chaos expansion (PCE) surrogate model which can not only be used for inputs-outputs relationship construction, but also produce the sensitivity of different design parameters were developed. Based on the PCE surrogate model, a new design scheme was obtained after optimization, in which the valve sealing stress is increased by 24.42% while keeping the maximum impact stress lower than 90% of the material allowable stress. The result confirms the ability and feasibility of the method proposed in this paper, and should also be suitable for performance design optimizations of control valves with similar structures.

• Smart Structures and Systems
• /
• v.31 no.3
• /
• pp.283-299
• /
• 2023

Structural stability of floating platforms has long since been a crucial issue in the field of marine engineering. Excessive motions would not only deteriorate the operating conditions but also seriously impact the safety, service life, and production efficiency. In recent decades, several control devices have been proposed to reduce unwanted motions, and an attractive one is the tuned heave plate (THP). However, the THP system may reduce or even lose its effectiveness when it is mistuned due to the shift of dominant wave frequency. In the present study, a novel adaptive tuned heave plate (ATHP) is proposed based on inerter by adjusting its inertance, which allows to overcome the limitation of the conventional THP and realize adaptations to the dominant wave frequencies in real time. Specifically, the analytical model of a representative semisubmersible platform (SSP) equipped with an ATHP is created, and the equations of motion are formulated accordingly. Two optimization strategies (i.e., J₁ and J₂ optimizations) are developed to determine the optimum design parameters of ATHP. The control effectiveness of the optimized ATHP is then examined in the frequency domain by comparing to those without control and controlled by the conventional THP. Moreover, parametric analyses are systematically performed to evaluate the influences of the pre-specified frequency ratio, damping ratio, heave plate sizes, peak periods and wave heights on the performance of ATHP. Furthermore, a Simulink model is also developed to examine the control performance of ATHP in the time domain. It is demonstrated that the proposed ATHP could adaptively adjust the optimum inertance-to-mass ratio by tracking the dominant wave frequencies in real time, and the proposed system shows better control performance than the conventional THP.

• Wind and Structures
• /
• v.37 no.4
• /
• pp.303-314
• /
• 2023

In this study, two steady RANS turbulence models (SST k-ω and Realizable k-ε) and four unsteady turbulence models (URANS SST k-ω and Realizable k-ε, SST-SAS, and SST-IDDES) are evaluated with respect to their capacity to predict crosswind characteristics on high-speed trains (HSTs). All of the numerical simulations are compared with the wind tunnel values and LES results to ensure the accuracy of each turbulence model. Specifically, the surface pressure distributions, time-averaged aerodynamic coefficients, flow fields, and computational cost are studied to determine the suitability of different models. Results suggest that the predictions of the pressure distributions and aerodynamic forces obtained from the steady and transient RANS models are almost the same. In particular, both SAS and IDDES exhibits similar predictions with wind tunnel test and LES, therefore, the SAS model is considered an attractive alternative for IDDES or LES in the crosswind study of trains. In addition, if the computational cost needs to be significantly reduced, the RANS SST k-ω model is shown to provide relatively reasonable results for the surface pressures and aerodynamic forces. As a result, the RANS SST k-ω model might be the most appropriate option for the expensive aerodynamic optimizations of trains using machine learning (ML) techniques because it balances solution accuracy and resource consumption.

• Molecules and Cells
• /
• v.46 no.12
• /
• pp.764-777
• /
• 2023

Recombinant immunotoxins (RITs) are fusion proteins consisting of a targeting domain linked to a toxin, offering a highly specific therapeutic strategy for cancer treatment. In this study, we engineered and characterized RITs aimed at mesothelin, a cell surface glycoprotein overexpressed in various malignancies. Through an extensive screening of a large nanobody library, four mesothelin-specific nanobodies were selected and genetically fused to a truncated Pseudomonas exotoxin (PE24B). Various optimizations, including the incorporation of furin cleavage sites, maltose-binding protein tags, and tobacco etch virus protease cleavage sites, were implemented to improve protein expression, solubility, and purification. The RITs were successfully overexpressed in Escherichia coli, achieving high solubility and purity post-purification. In vitro cytotoxicity assays on gastric carcinoma cell lines NCI-N87 and AGS revealed that Meso(Nb2)-PE24B demonstrated the highest cytotoxic efficacy, warranting further characterization. This RIT also displayed selective binding to human and monkey mesothelins but not to mouse mesothelin. The competitive binding assays between different RIT constructs revealed significant alterations in IC₅₀ values, emphasizing the importance of nanobody specificity. Finally, a modification in the endoplasmic reticulum retention signal at the C-terminus further augmented its cytotoxic activity. Our findings offer valuable insights into the design and optimization of RITs, showcasing the potential of Meso(Nb2)-PE24B as a promising therapeutic candidate for targeted cancer treatment.

• International Journal of Internet, Broadcasting and Communication
• /
• v.16 no.2
• /
• pp.237-246
• /
• 2024

In recent years, the frequent occurrence of fire accidents in university libraries has posed significant threats to the safety of students' lives and property, alongside negative social impacts. Accurately analyzing the factors affecting evacuation during library fires and proposing optimized measures for safe evacuation is thus crucial. This paper utilizes a specific university library as a case study, simulating fire evacuation scenarios using the Pathfinder software, to assess and validate evacuation strategies and propose relevant optimizations. Pathfinder, developed by Thunderhead Engineering in the United States, is an intuitive and straightforward personnel emergency evacuation assessment system, offering advanced visualization interfaces and 3D animation effects. This study aims to construct evacuation models and perform simulation analysis for the selected university library using Pathfinder. The library's structural layout, people flow characteristics, and the nature of fire and smoke spread are considered in the analysis. Additionally, evacuation scenarios involving different fire outbreak locations and the status of emergency exits are examined. The findings underscore the importance of effective evacuation in fire situations, highlighting how environmental conditions, individual characteristics, and behavioral patterns significantly influence evacuation efficiency. Through these investigations, the study enhances understanding and optimization of evacuation strategies in fire scenarios, thereby improving safety and efficiency. The research not only provides concrete and practical guidelines for building design, management, and emergency response planning in libraries but also offers valuable insights for the design and management of effective evacuation systems in buildings, crucial for ensuring occupant safety and minimizing loss of life in potential hazard situations

• Journal of the Korean Institute of Landscape Architecture
• /
• v.48 no.5
• /
• pp.16-27
• /
• 2020

This study presented a plan for the establishment of conservation areas and the ecological management of those areas in the stream based on the Marxan with Zones Program for a Jungrangcheon Stream in downtown Seoul. The application of the Marxan with Zones Program included the stage of planning unit setting, application of mapping indices, numerical correction for repetitive analysis, creation of scenario-specific optimizations through analysis, analysis of sensitivity by scenario, review, and the selection of optimal plans among the scenarios considered. As a result of the establishment of a conservation area near Jungrangcheon Stream, which has several watershed areas, including an upper-middle-class wildlife protection zone, which was previously designated and managed as a conservation area, and the migratory protection zone downstream of Jungrangcheon Stream were designated as key conservation areas. A number of wild birds were observed in the upper reaches of Jungrangcheon Stream, adjacent to the forests of Suraksan Mountain and Dobongsan Mountain. The downstream area is a habitat for migratory birds that travel along the stream and the adjacent river ecosystem, including the Hangang River confluence and Cheonggyecheon Stream confluence. Therefore, the upper and lower reaches of Jungrangcheon Stream are connected to forest ecosystems such as Dobongsan Mountain, Suraksan Mountain, and Eungbongsan Mountain, as well as urban green area and river ecosystems in the basin area, which influence the establishment of conservation areas. This study verified the establishment and evaluation of existing conservation areas through the Marxan with Zones Program during the verification of the conservation areas and was presented as in-stream management and basin management method to manage the basin areas derived from core conservation areas determined through the program.

• Journal of the Korean Applied Science and Technology
• /
• v.29 no.2
• /
• pp.199-204
• /
• 2012

Organic thin-film transistors (OTFTs) have received considerable attention because their potential applications for nano-scale thin-film structures have been widely researched for large-scale integration industries, such as semiconductors and displays. However, research in developing n-type materials and devices has been relatively shortage than developing p-type materials. Therefore, we report on the fabrication of top-contact [6,6]-phenyl-C61-butyricacidmethylester (PCBM) TFTs by using three different solvent, o-dichlorobenzene, toluene and chloroform. An appropriate choice of solvent shows that the electrical characteristics of PCBM TFTs can be improved. Moreover, our PCBM TFTs with the cross-linked Poly(4-vinylphenol) dielectric layer exhibits the most pronounced improvements in terms of the field-effect mobility (${\sim}0.034cm^2/Vs$) and the on/off current ratio (${\sim}1.3{\times}10^5$) for our results. From these results, it can be concluded that solvent-modification of an organic semiconductor in PCBM TFTs is useful and can be extended to further investigations on the PCBM TFTs having polymeric gate dielectrics. It is expected that process optimizations using solution-processing of organic semiconductor materials will allow the development of the n-type organic TFTs for low-cost electronics and various electronic applications.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.11
• /
• pp.635-639
• /
• 2016

In the devices used in the microwave frequency band, the gain decreases as the frequency increases due to the parasitic component. To compensate for these characteristics, a linear gain equalizer with an opposite slope is needed in wideband systems, such as those used for electronic warfare. In this study, a linear gain equalizer that can be used in the 18 ~ 40GHz band is designed and fabricated. Circuit design and momentum design (optimizations) were carried out to reduce the errors between design and manufacturing. A thin film process is used to minimize the parasitic components within the implementation frequency band. A sheet resistance of 100 ohm/square was employed to minimize the wavelength variation due to the length of the thin film resistor. This linear gain equalizer is a structure that combines a quarter wavelength-resonator on a series microstrip line with a resistor. All three 1/4 wavelength short resonators were used. The fabricated linear gain equalizer has a loss of more than -5dB at 40GHz and a 6dB slope in the 18 ~ 40GHz band. By using the manufactured gain equalizer in a multi-stage connected device such as an electronic warfare receiver, the gain flatness degradation with increasing frequency can be reduced.