• Journal of the Korean Institute of Telematics and Electronics
• /
• v.26 no.6
• /
• pp.38-47
• /
• 1989

A purely analytical model for HEMT's based on a two dimensional charge control simul-ation[4] is proposed. In this model proper treatment of diffusion effect of electron transport along a 2-DEG (two dimensional electron gas) channel is perfoemed. This diffusion effect is shown to effectively increase the bulk mibility and threshold voltage of the I-V curves compared to the existing models. The channel thickness and gate capacitance are expressed as functions of gate voltages covering subthreshold characteristics of HEMT's analytically. By introducing the finite channel opening and an effiective channel-length modulation, the solpe of the saturation region of the I-V curves ws modeled. The smooth transition of the I-V curves at linear-to-saturation regions of the I-V curves was possible using the continuous Troffimenkoff-type of field dependent mobility. Furthermore, a correction factor f was introduced to account for the finite transition section forming between a GCA and a saturated section. This factor removes large discrepancies in the saturation region of the I-V curve predicted by existing l-dimensional models.

• Electrical & Electronic Materials
• /
• v.6 no.2
• /
• pp.147-151
• /
• 1993

PECVD에 의해 Burried gate 비정질 실리콘 박막트랜지스터를 제작하여 포화 전압 대 전류 특성에 대하여 새로운 해석을 하였고 해석 결과는 실험적으로 증명되었다. 본 연구의 결과 실험된 전달특성과 출력특성을 모델화 하였는데 이 모델식은 I$_{D}$와 V$_{G}$의 실험결과에서 얻어지는 3가지 함수를 기본으로 모델화 되었다. 포화 드레인 전류는 V$_{G}$가 증가할수록 증가되었고 디바이스의 포화는 드레인 전압이 커질수록 증가되었으며 문턱전압은 감소됨을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.576-578
• /
• 2022

딥러닝 기술의 발전으로 객체 인색, 영상 분석에 관한 성능이 비약적으로 발전하였다. 하지만 고성능 GPU 를 사용하는 컴퓨팅 환경이 아닌 제한적인 엣지 디바이스 환경에서의 영상 처리 및 딥러닝 모델의 적용을 위해서는 엣지 디바이스에서 딥러닝 모델 실행 환경 과 이에 대한 분석이 필요하다. 본 논문에서는 RISC-V ISA 를 구현한 RISC-V 가상 플랫폼에 yolov3-tiny 모델 기반 객체 인식 시스템을 소프트웨어 레벨에서 포팅하여 구현하고, 샘플 이미지에 대한 네트워크 딥러닝 연산 및 객체 인식 알고리즘을 적용하여 그 결과를 도출하여 보았다. 본 적용을 바탕으로 RISC-V 기반 임베디드 엣지 디바이스 플랫폼에서 딥러닝 네트워크 연산과 객체 인식 알고리즘의 수행에 대한 분석과 딥러닝 연산 최적화를 위한 알고리즘 연구에 활용할 수 있다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.38 no.11
• /
• pp.1201-1206
• /
• 2014

This paper presents the development of an event-based operator model of a wheel loader for virtual V-pattern working. The objective of this study is to analyze the performance and dynamic behavior of the wheel loader for a typical V-pattern. The proposed typical V-pattern working is divided into four stages. The developed operator model is based on eight events, and the operator's inputs are occurred sequentially by event. A 3D dynamic simulation model of the wheel loader is developed and used to analyze the dynamic behavior during working, and the simulation results are compared with the experimental data of V-pattern working. The proposed 3D dynamic simulation model and operator model are constructed using MATLAB/Simulink. The proposed operator model for V-pattern working is expected to enable evaluation of the working performance and dynamic behavior of the wheel loader.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.69-71
• /
• 2023

본 논문은 YOLO 기반 모델의 철도 시스템 내 선로 고정장치 탐지 성능을 비교하고 분석한다. 여기서 철도 시스템은 열차가 주행하기 위한 선로, 침목, 패스너 등의 구성요소를 포함한다. 침목은 지반과 직접적으로 연결되며, 선로를 지반 위에 안정적으로 지지하고 궤간을 정확하게 유지하는 역할을 한다. 또한, 패스너는 선로를 침목에 단단히 고정시키는 역할을 한다. 이러한 선로 고정장치의 부재는 인명 사고로 이어질 수 있어 지속적인 관리와 유지 보수가 필수적이다. 본 논문에서는 철도 시스템의 선로 고정장치 탐지를 위해 YOLO V5 및 V8 딥러닝 모델의 적용 가능성을 실험적으로 접근하며, 두 모델의 탐지 성능을 비교한다. 실험 결과, YOLO V8 및 V5 모델은 모두 뛰어난 성능을 보이는데, 특히 YOLO V8 모델이 더욱 우수한 성능을 보인다. 이로써 YOLO 알고리즘은 선로 고정장치 탐지에 적합하다는 것을 증명한다. 그러나 일부 False Positive Sample이 관측되었음을 확인하고, 이로부터 모델 성능의 개선이 필요하다는 결론을 도출하였다.

• Journal of Information Technology Services
• /
• v.11 no.sup
• /
• pp.51-60
• /
• 2012

In this paper, we provide the analysis of device model and method between device models for compatible M&S V&V of the NetSPIN. First of all, we analysis features, structure, and classification of the NetSPIN. The second, as a part of reliable V&V process, we analysis network system modeling process, correlation between device modeling process for M&S of the NetSPIN. The third, we suggest making a kind of pool of reference model and module of devices for the increase factor of reuse between device model. We also, at the point view of M&S V&V, conclude that there is the validity of the fidelity in device modeling process. Through the analysis of the NetSPIN device model and suggestion of the method for higher compatibility between device modes, the development process of device model be clearly understood. Also we present the effective method of the development for reliable device mode as the point of V&V.

• Korean Journal of Remote Sensing
• /
• v.39 no.3
• /
• pp.269-282
• /
• 2023

The purpose of this study was to construct land cover maps using a deep learning model and to select the optimal deep learning model for land cover classification by adjusting the dataset such as input image size and Stride application. Two types of deep learning models, the U-net model and the DeeplabV3+ model with an Encoder-Decoder network, were utilized. Also, the combination of the two deep learning models, which is an Ensemble model, was used in this study. The dataset utilized RapidEye satellite images as input images and the label images used Raster images based on the six categories of the land use of Intergovernmental Panel on Climate Change as true value. This study focused on the problem of the quality improvement of the dataset to enhance the accuracy of deep learning model and constructed twelve land cover maps using the combination of three deep learning models (U-net, DeeplabV3+, and Ensemble), two input image sizes (64 × 64 pixel and 256 × 256 pixel), and two Stride application rates (50% and 100%). The evaluation of the accuracy of the label images and the deep learning-based land cover maps showed that the U-net and DeeplabV3+ models had high accuracy, with overall accuracy values of approximately 87.9% and 89.8%, and kappa coefficients of over 72%. In addition, applying the Ensemble and Stride to the deep learning models resulted in a maximum increase of approximately 3% in accuracy and an improvement in the issue of boundary inconsistency, which is a problem associated with Semantic Segmentation based deep learning models.

• Journal of the Korea Society for Simulation
• /
• v.19 no.4
• /
• pp.219-233
• /
• 2010

The objective of the interoperation of L-V-C Simulation is to enable practical integration training by taking advantages and compensating disadvantages of simulation models, such as Live, Virtual and Constructive models. As a study on the interoperation of L-V-C simulation, this paper suggests effective interoperation method between Virtual and Constructive simulation models and demonstrates small-size intagrated combat training model through V-C Test-Bed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.451-452
• /
• 2023

최근 CNN 이 다양한 산업에 확산되고 있으며, IoT 기기 및 엣지 컴퓨팅에 적합한 경량 모델에 대한 연구가 급증하고 있다. 본 논문에서는 CNN 모델의 파라미터 비트 연산을 위한 자동화 프레임워크를 제안하고, 파라미터 비트와 모델 정확도 사이의 관계를 실험 및 연구한다. 제안된 프레임워크는 하위 n- bit 를 0 으로 설정하여 정보 손실 발생시킴으로써 ImageNet 데이터셋으로 사전 학습된 CNN 모델의 파라미터와 정확도의 강인성을 비트 단위로 체계적으로 실험할 수 있다. 우리는 비트 연산을 수행한 파라미터로 InceptionV3, InceptionResnetV2, ResNet50, Xception, DenseNet121, MobileNetV1, MobileNetV2 모델의 정확도를 평가한다. 실험 결과는 성능이 낮은 모델일수록 파라미터와 정확도 간의 강인성이 높아 성능이 좋은 모델보다 정확도를 유지하는 비트 수가 적다는 것을 보여준다.

• Journal of Advanced Marine Engineering and Technology
• /
• v.35 no.5
• /
• pp.616-624
• /
• 2011

The flow characteristics and heat transfer augment on the periodically arranged semi-circular ribs in a rectangular channel for turbulent flow has been investigated numerically. The aspect ratio of the rectangular channel was AR=5, the rib height to hydraulic diameter ratio were 0.07 and rib height to channel height ratio was set as e/H=0.117 for various PR(rib pitch-to-rib height rate) between 8~14, respectively. The SST k-${\omega}$ turbulence model and v2-f turbulence model were used to find out the heat transfer and the flow characteristics of near the wall which are suited to obtain realistic phenomena. The numerical analysis results show turbulent flow characteristics, heat transfer enhancement and friction factor as observed experimentally. The results predict that turbulent kinetic energy(k) is closely relative to the diffusion of recirculation flow. and v2-f turbulence model simulation results have a good agreement with experimental values.