• Electronics and Telecommunications Trends
• /
• v.36 no.2
• /
• pp.56-64
• /
• 2021

Many computer vision algorithms are computationally expensive and require a lot of computing resources. Recently, owing to machine learning technology and high-performance embedded systems, vision processing applications, such as object detection, face recognition, and visual inspection, are widely used. However, on-devices need to use their resources to handle powerful vision works with low power consumption in heterogeneous environments. Consequently, global manufacturers are trying to lock many developers into their ecosystem, providing integrated low-power chips and dedicated vision libraries. Khronos Group-an international standard organization-has released the OpenVX standard for high-performance/low-power vision processing in heterogeneous on-device systems. This paper describes vision libraries for the embedded systems and presents the OpenVX standard along with related trends for on-device vision system.

• KIPS Transactions on Computer and Communication Systems
• /
• v.6 no.6
• /
• pp.261-270
• /
• 2017

With the development of smart on-devices and communication technology, demand for non-stop services is increasing. Therefore, the high availability for continuously providing services in the event of system failure has been spotlighted. Meanwhile, because most internet-based services are provided by using TCP, an efficient TCP session recovery technique for providing non-stop services is required. However, the existing TCP session recovery techniques are inefficient because it has a high recovery cost or does not support failover operation, To solve these problems, in this paper, we propose a TCP session recovery technique for high availability in smart on-devices. For this, we first recover the TCP session without re-establish the TCP session by correcting a sequence number and a acknowledgment number. Second, we synchronize the TCP session recovery data between the master and the server, and then we operate the failover operation when master server fails. Finally, we provide the non-stop service to peer by using the virtual IP number and the transmission of GARP (Gratuitous ARP) packet.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.649-652
• /
• 1998

본 논문에서는 MICROTEC〔3,4〕시뮬레이터를 이용하여 소트키 다이오드를 형성하고 금속-반도체 쇼트키 접촉에서 턴 온 전압과 항복 전압을 관찰하였다. 또한 여러 가지 쇼트키 장벽 높이를 가지는 금속을 사용하여 동일한 디바이스에서 이들 금속-반도체 접촉에 전압을 인가했을 때, 순 방향에서 턴 온 특성을 관찰하여 턴 온 전압과 역 방향에서의 항복 현상을 관찰하여 항복 전압을 확인하였다. 사용된 금속은 Au(0.8V), Mo(0.68V), Pt(0.9V), Ti(0.5V) 이며 반도체는 실리콘 n/n 구조가 형성되었다. 쇼트키 다이오드는 대 전력용 보다는 높은 속도의 스위칭 디바이스에 주로 응용되고 있으며 장벽의 높이가 높을수록 뚜렷한 정류 특성을 나타내어 순 방향 바이어스에서 빠른 턴 온 특성이 예상되는데 시뮬레이션 결과 또한 잘 일치하였다. 그리고 다이오드의 I-V 특성을 관찰하기 위해 역 방향 바이어스에서의 항복 전압을 관찰하였는데 쇼트키 장벽이 높을수록 낮은 항복 전압이 나타났다. 또한 디바이스 공정에서 epitaxial과 열처리 공정 후의 2차원적인 농도 분포를 나타내었다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.3
• /
• pp.754-766
• /
• 2010

In the wireless communication protocols, a network device is responsible for the operation of lower-layers. The network device consists of hardware and software modules, so it can be designed using system-on-chip simulator. The simulator design needs the support of a network simulator as well as system-on-chip simulator, because the network device interact with various higher layer communication protocols. Therefore the co-simulator can become a development environment of the network device through the combining of them. In this paper we propose a co-simulator combining these two simulators. The proposed co-simulator does not degrade performance due to integrations. Also, it is easy to integrate them because the implementation of the kernel is independent.

• Journal of Broadcast Engineering
• /
• v.27 no.3
• /
• pp.341-350
• /
• 2022

In this paper, we propose a deep learning based on-device augmented reality (AR) system in which multiple input images are used to implement the correct occlusion in a real environment. The proposed system is composed of three technical steps; camera pose estimation, depth estimation, and object augmentation. Each step employs various mobile frameworks to optimize the processing on the on-device environment. Firstly, in the camera pose estimation stage, the massive computation involved in feature extraction is parallelized using OpenCL which is the GPU parallelization framework. Next, in depth estimation, monocular and multiple image-based depth image inference is accelerated using the mobile deep learning framework, i.e. TensorFlow Lite. Finally, object augmentation and occlusion handling are performed on the OpenGL ES mobile graphics framework. The proposed augmented reality system is implemented as an application in the Android environment. We evaluate the performance of the proposed system in terms of augmentation accuracy and the processing time in the mobile as well as PC environments.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.1-8
• /
• 2024

On-device AI technology, which can operate AI models at the edge devices to support real-time processing and privacy enhancement, is attracting attention. As intelligent IoT is applied to various industries, services utilizing the on-device AI technology are increasing significantly. However, general deep learning models require a lot of computational resources for inference and learning. Therefore, various lightweighting methods such as quantization and pruning have been suggested to operate deep learning models in embedded edge devices. Among the lightweighting methods, we analyze how to lightweight and apply deep learning models to edge computing devices, focusing on pruning technology in this paper. In particular, we utilize dynamic and static pruning techniques to evaluate the inference speed, accuracy, and memory usage of a lightweight AI vision model. The content analyzed in this paper can be used for intelligent video control systems or video security systems in autonomous vehicles, where real-time processing are highly required. In addition, it is expected that the content can be used more effectively in various IoT services and industries.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.597-599
• /
• 2022

When IoT device performs on-device AI, the device is required to use various AI models selectively according to target service and surrounding environment. Also, AI model can be updated by additional training such as federated learning or adapting the improved technique. Hence, for successful on-device AI, IoT device should acquire various AI models selectively or update previous AI model to new one. In this paper, we propose AI model repository to tackle this issue. The repository supports AI model registration, searching, management, and deployment along with dashboard for practical usage. We implemented it using Node.js and Vue.js to verify it works well.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2002.11a
• /
• pp.719-722
• /
• 2002

임베디드 시스템을 이용하여 TCP/IP 상의 인터넷 제어시스템을 구현하였다. AT90S8535 마이크로 컨트롤러 구조 및 각종명령 레지스터 동작 원리, avr-gcc 하드웨어 프로그래밍 및 명령 레지스터의 구현원리, JAVA 애플릿 프로그래밍, 전반적인 하드웨어 기초 이론을 바탕으로 Mellow Device 1300 임베디드 시스템과 AVR90S853S 마이크로 컨트롤러 상호간의 시리얼 통신을 이용하여 원격지의 온/습도 검침 및 각종 하드웨어 디바이스의 on/off를 구현하도록 설계하였고 실시간으로 검출된 온/습도 데이터를 JAVA 애플릿을 이용하여 그래픽 챠트로 보기 쉽게 표현하였으며 A/D 변환된 온도 및 습도 데이터와 각종 H/W 디바이스 on/off 상태 데이터를 RS232 인터페이스를 이용하여 Mellow Device 1300로 송/수신하도록 구현하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.11a
• /
• pp.305-306
• /
• 2009

본 논문에서는 홈네트워크 서비스 환경에서의 사용자 맞춤형 디바이스 제어 솔루션을 제공하기 위해 디바이스 제어와 관련된 사용자 선호도 스키마를 정의하고, 이를 기반으로 한 프로파일 관리 기법을 제시하였다. 나이, 성별, 직업 등과 같은 사용자 개인정보에 대한 스키마와 향기, 온도, 조명, 바람 등에 대한 선호도 스키마를 포함하는 가전 디바이스 제어와 관련된 다양한 요소들에 대한 스키마를 XML로 정의하였으며, 사용자의 인터랙션이나 프로파일 정보 입력을 통해 사용자 선호도 정보를 획득하여, 단방향의 수동적인 단일 디바이스 제어가 아닌 여러 개의 디바이스를 최적으로 제어할 수 있는 사용자 맞춤형 홈서비스 환경을 제시하였다.

• Journal of Korean Electronics
• /
• v.3 no.12
• /
• pp.40-48
• /
• 1983