Daegu Gyeongbuk Institute of Science and Technology
DGIST Library Institutional RepositoryNot a member yet
12664 research outputs found
Sort by
Selective Artificial Neural Network and Stimulation Magnetically Controlled Cell-based Cellbots
No full textWe report various magnetically actuated cell-based microrobots (cellbots) for selective neurite alignment, neuronal connections, and/or selective stimulation. A biodegradable spherical gelatin methacrylate (GelMA) microrobot was fabricated in a flow-focusing droplet generator by shearing a mixture of GelMA, photoinitiator, and superparamagnetic iron oxide nanoparticles (SPIONs) with a mixture of oil and surfactant. Human turbinate stem cells (hNTSCs) were loaded onto the GelMA microrobot, and the hNTSC-loaded microrobot exhibited a precise rolling motion in response to an external rotating magnetic field. The microrobot was enzymatically degraded by collagenase and released hNTSCs proliferated and differentiated into neuronal cells. The proposed cellbots show the potential for in vitrocell delivery and neural experiments to understand how neurons communicate in the neural network
3D-printed magnetic-based air pressure sensor for continuous respiration monitoring and breathing rehabilitation
Get PDFThe rapid development of point-of-care testing has made prompt diagnosis, monitoring and treatment possible for many patients suffering from chronic respiratory diseases. Currently, the biggest challenge is further optimizing testing devices to facilitate more functionalities with higher efficiency and performance, along with specificity toward patient needs. By understanding that patients with chronic respiratory diseases may have difficulty breathing within a normal range, a respiration sensor is developed focusing on sensitivities in the lower air pressure range. In contrast to the simpler airflow data, the sensor can provide respiratory air pressure as an output using a magnetic-based pressure sensor. This unconventional but highly reliable approach, combined with the rest of the simple 3D-printed design of the sensor, offers a wide range of tunability and functionalities. Due to the detachable components of the respiration sensor, the device can be easily transformed into other respiratory uses such as an inspiratory muscle training device or modified to cater for higher-ranged deep breathing. Therefore, not only does it reach very low air pressure measurement (0.1 cmH2O) for normal, tidal breathing, but the sensor can also be manipulated to detect high levels of air pressure (up to 35 cmH2O for exhalation and 45 cmH2O for inhalation). With its excellent sensitivities (0.0456 mV/cmH2O for inhalation,-0.0940 mV/cmH2O for exhalation), impressive distinction between inhalation and exhalation, and fully reproducible and convenient design, we believe that this respiration sensor will pave the way for developing multimodal and multifunctional respiration sensors within the biomedical field. © The Author(s) 2024.TRUEscopu
Non-Abelian fractional quantum anomalous Hall states and first Landau level physics of the second moiré band of twisted bilayer MoTe2
No full textUtilizing the realistic continuum description of twisted bilayer MoTe2 and many-body exact diagonalization calculation, we establish that the second moiré band of twisted bilayer MoTe2, at a small twist angle of approximately 2∘, serves as an optimal platform for achieving the long-sought non-Abelian fractional quantum anomalous Hall states without the need for external magnetic fields. Across a wide parameter range, our exact diagonalization calculations reveal that the half-filled second moiré band demonstrates the ground state degeneracy and spectral flows, which are consistent with the Pfaffian state in the first Landau level. We further elucidate that the emergence of the non-Abelian state is deeply connected to the remarkable similarity between the second moiré band and the first Landau level. Essentially, the band not only exhibits characteristics akin to the first Landau level, 12π∫BZd2ktrη(k)≈3 where ηab(k) is the Fubini-Study metric of the band, but also that its projected Coulomb interaction closely mirrors the Haldane pseudopotentials of the first Landau level. Motivated by this observation, we introduce a metric of first Landau level-ness of a band, which quantitatively measures the alignment of the projected Coulomb interaction with the Haldane pseudopotentials in Landau levels. This metric is then compared with the global phase diagram of the half-filled second moiré band, revealing its utility in predicting the parameter region of the non-Abelian state. In addition, we uncover that the first and third moiré bands closely resemble the lowest and second Landau levels, revealing a remarkable sequential equivalence between the moiré bands and Landau levels. We finally discuss the potential implications on experiments. © 2024 American Physical Society.FALSEsciescopu
Sustainable Li-metal Protection in Li-S Batteries with Lean Electrolyte
No full text높은 이론용량을 갖는 황 (1675 mAh g-1)과 리튬금속 (3860 mAh g-1)을 포함하는 리튬-황 (Li-S) 전지의 실질적인 고에너지밀도를 구현하기 위해서는 전해액 양을 최소화하는 것이 핵심이다. 희박 전해질 (Lean electrolyte) 사용은 전지 저항의 급격한 증가와 리튬과의 부반응으로 기인한 전해액 변질 및 소모를 유발해 급격한 수명 퇴화로 이어져 상용화의 제약이 되고 있다. 리튬 음극에 보호막을 도입해 극복 가능이 예상되나 초기 리튬 용출 과정으로 부터 기인한 리튬-황 전지 내 보호막의 불규칙한 구조적 진화 및 퇴화로 인해 적용의 한계가 있다. 본 연구에서는 리튬 음극에 고분자계 보호막 도입에 따른 열화 메커니즘과 희박 전해액을 포함한 리튬-황 전지 성능에 미치는 영향을 규명한다. 특히, 상용 리튬의 불균일한 표면 산화층에 기인한 국부적인 리튬 용출의 심각성과 이에 따른 고분자 보호막 종류에 따른 구조 진화과정을 관찰하였다. 리튬-황 전지의 초기 방전/충전에 따라 보호막의 탈리 및 천공, 변형 등 퇴화모드가 고분자/리튬 간 계면 에너지 및 유연 특성에 대한 의존성을 다루고자 한다. 또한, 퇴화모드에 따른 전해액 침투 및 리튬-전해액간 부반응 정도 차이와 리튬 표면 열화도 (전극 부피팽창 및 쿨롱 효율)와의 상관관계를 규명하고자 한다. 앞선 결과를 바탕으로 고분자와 계면 에너지를 낮추면서 기계적 유연성이 충분한 고분자가 결합된 이중층 (Dual-layer) 보호막을 도입하여 희박 전해질 환경에서 기존 리튬 음극 대비 2배 향상된 성능을 확인했다 (80% 용량유지율 기준). 본 연구를 통해 향후 Dual-layer보호막을 설계에 있어 추가로 고려할 수 있는 기술들과 초기 리튬 용출 거동 제어를 통한 성능 향상 가능성을 제안하고자 한다
Estimation of Multi-state Dependent Disturbance by Using Multi-dimensional Gaussian Process
No full textHigh-precision linear motor stages have been widely used for their excellent positioning accuracy and speed. However, core-type linear motor stages have performance limitations because of various nonlinear factors including cogging force, friction, and geometrical imbalance. This paper analyzes disturbances in velocity and position domains and trains a Two-Input-Single-Output (TISO) nonlinear model using the Gaussian process for the disturbance. With this, two state-dependent disturbances are removed effectively. As a result, the control performance with a proposed controller is enhanced. Ultimately, this paper introduces three contribution points: 1) analysis of disturbances based on position/velocity, 2) design of TISO Gaussian process model, and 3) validation of estimation performance of proposed algorithm through simulation. © 2024 IEEE
Enhancing the oxidation of polystyrene through a homogeneous liquid degradation system for effective microbial degradation
No full textPlastics play a crucial role in modern industries; however, their resistance to natural degradation contributes to environmental pollution, and microplastics pose a health threat. The hydrophobic nature of microplastics poses a considerable challenge, rendering them resistant to dissolving in water. In this study, we conducted a comparative analysis of the microbial biodegradation capabilities of polystyrene in solid and liquid states. Polystyrene in its solid foam form, along with polystyrene converted into a liquid state using ethyl-ester oil, was biodegraded by microorganisms. Subsequently, the liquid plastic was re-extracted into its solid form, and the degree of degradation was assessed using weight loss measurement, XPS, FT-IR, GPC, and TGA. Liquid-state polystyrene exhibited a higher degradation rate than that reported previously. Furthermore, liquid polystyrene undergoes more pronounced oxidation than its solid counterpart, leading to an increased oxygen atom ratio. Chemical structure analysis highlighted the distinct formation of –OH and C=O functional groups in the liquid state compared to those in the solid state. Additionally, notable changes in the molecular weight and thermal stability of polystyrene were observed during biodegradation in the liquid state. This study suggests that a heterogeneous reaction (solid plastic-liquid medium) might impede plastic biodegradation, while indicating the potential to enhance the degradation efficiency through a homogeneous reaction (liquid plastic-liquid medium). The follow-up study identifies appropriate solvents and optimizes cultivation conditions, offering potential to enhance the efficiency of biological plastic degradation. Copyright © 2024 Kim, Koh, Shin, Suh, Lee and Choi.TRUEsciescopu
Generation of synthetic PET/MR fusion images from MR images using a combination of generative adversarial networks and conditional denoising diffusion probabilistic models based on simultaneous 18F-FDG PET/MR image data of pyogenic spondylodiscitis
No full textBACKGROUND CONTEXT: Cross-modality image generation from magnetic resonance (MR) to positron emission tomography (PET) using the generative model can be expected to have complementary effects by addressing the limitations and maximizing the advantages inherent in each modality. PURPOSE: This study aims to generate synthetic PET/MR fusion images from MR images using a combination of generative adversarial networks (GANs) and conditional denoising diffusion probabilistic models (cDDPMs) based on simultaneous 18F-fluorodeoxyglucose (18F-FDG) PET/MR image data. STUDY DESIGN: Retrospective study with prospectively collected clinical and radiological data. PATIENT SAMPLE: This study included 94 patients (60 men and 34 women) with thoraco-lumbar pyogenic spondylodiscitis (PSD) from February 2017 to January 2020 in a single tertiary institution. OUTCOME MEASURES: Quantitative and qualitative image similarity were analyzed between the real and synthetic PET/ T2-weighted fat saturation MR (T2FS) fusion images on the test data set. METHODS: We used paired spinal sagittal T2FS and PET/T2FS fusion images of simultaneous 18F-FDG PET/MR imaging examination in patients with PSD, which were employed to generate synthetic PET/T2FS fusion images from T2FS images using a combination of Pix2Pix (U-Net generator + Least Squares GANs discriminator) and cDDPMs algorithms. In the analyses of image similarity between the real and synthetic PET/T2FS fusion images, we adopted the values of mean peak signal to noise ratio (PSNR), mean structural similarity measurement (SSIM), mean absolute error (MAE), and mean squared error (MSE) for quantitative analysis, while the discrimination accuracy by three spine surgeons was applied for qualitative analysis. RESULTS: Total of 2,082 pairs of T2FS and PET/T2FS fusion images were obtained from 172 examinations on 94 patients, which were randomly assigned to training, validation, and test data sets in 8:1:1 ratio (1664, 209, and 209 pairs). The quantitative analysis revealed PSNR of 30.634 ± 3.437, SSIM of 0.910 ± 0.067, MAE of 0.017 ± 0.008, and MSE of 0.001 ± 0.001, respectively. The values of PSNR, MAE, and MSE significantly decreased as FDG uptake increased in real PET/T2FS fusion image, with no significant correlation on SSIM. In the qualitative analysis, the overall discrimination accuracy between real and synthetic PET/T2FS fusion images was 47.4%. CONCLUSIONS: The combination of Pix2Pix and cDDPMs demonstrated the potential for cross-modal image generation from MR to PET images, with reliable quantitative and qualitative image similarities. © 2024 Elsevier Inc.FALSEsciescopu
sEMG signal classification model with simultaneous consideration of localpatterns and global characteristics
No full text본 논문은 EMG pattern recognition(PR)을 위해 gesture에 따른 표면 근전도 신호를 높은 정확도로 분류하는 CNN-Transformer 결합 모델을 제안한다. 제안하는 모델은 EMG PR에서 로컬 패턴뿐만이 아니라 신호의 전역적인 특성을 고려하여 여러 gesture에 대해 높은 정확도를 유지 할 수 있다.제안하는 모델의 성능을 평가하기 위해 ninapro의 DB2 dataset으로 사용한다. 기존의 대부분의 딥러닝 기반의 PR 모델은 ninapro의 DB2 dataset의일부 gesture에 대해서만 실험을 진행하여 모델의 robustness를 가늠하기 어려웠다. 본 논문에서 모든 gesture 대해 일관적으로 높은 분류 정확도를보인다. 제안하는 모델은 각 gesture에서 평균 86.2%의 분류 정확도를 달성하였으며,기존 모델보다 4% 더 높은 성능을 보였다
BEE-SLAM: A 65nm 17.96 TOPS/W 97.55%-Sparse-Activity Hybrid Mixed-Signal/Digital Multi-Agent Neuromorphic SLAM Accelerator for Swarm Robotics
No full textMulti-agent (MA) AI holds great promise for enhancing edge devices with limited computing resources [1]. In particular, MA simultaneous localization and mapping (SLAM) is actively under investigation to improve map accuracy in swarm robotics. Conventional keyframe-based SLAM, leveraging landmarks [2]-[4], provides the appropriate map accuracy but is unsuitable for MA SLAM on decentralized edge devices due to computational complexity. The neuromorphic SLAM is a candidate for MA SLAM owing to its low complexity in singleagent operation [5]. However, this method is still infeasible in MA SLAM due to the drastic increase of complexity in MA map correction. As such, several challenges need to be addressed via circuit-algorithm co-design in deploying MA SLAM to edge devices. In this paper, we present the BEE-SLAM accelerator, inspired by bee communication, featuring hybrid mixed-signal/digital biomimetic circuits and MA map error correction (MAEC), achieving the energy efficiency of 17.96 TOPS/W in outdoor MA SLAM operation. © 2024 IEEE
입력 시간 지연 극복을 위한 강화학습 기반 지하철 열차 추적 제어
No full textReinforcement learning, Deep deterministic policy gradient, Input time delay, Automatic train operation, Metro train1 Introduction 1
1.1 Motivation 1
1.2 Related Work 2
1.3 Contribution 3
1.4 Thesis Outline 3
2 ATO Simulator and RL 5
2.1 ATO simulator 5
2.2 RL algorithms 8
3 Prediction-based input time delay compensation 11
3.1 Step 1 11
3.2 Step 2 13
4 Training 17
4.1 Training Episode 17
4.2 Neural Networks and HyperParameter 19
5 Results 21
5.1 Impact of Prediction 21
5.2 Comparison with Existing Controller in Selected Sections 24
5.3 Comparison with Existing Controller in Every Section 30
6 Conclusion 49
bibliography 54
국문초록 56MasterdCollectio