• Journal of Korea Multimedia Society
• /
• v.25 no.8
• /
• pp.1049-1058
• /
• 2022

HAR(Human Action Recognition) such as anomaly and object detection has become a trend in research field(s) that focus on utilizing Artificial Intelligence (AI) methods to analyze patterns of human action in crime-ridden area(s), media services, and industrial facilities. Especially, in real-time system(s) using video streaming data, HAR has become a more important AI-based research field in application development and many different research fields using HAR have currently been developed and improved. In this paper, we propose and analyze a deep-learning-based HAR that provides more efficient scheme(s) using an intelligent AI models, such system can be applied to media services using RGB video streaming data usage without feature extraction pre-processing. For the method, we adopt Slowfast based on the Deep Neural Network(DNN) model under an open dataset(HMDB-51 or UCF101) for improvement in prediction accuracy.

• Journal of Power Electronics
• /
• v.1 no.2
• /
• pp.65-70
• /
• 2001

Transmission System Operators are governed by operational security standards that are applied in real time. During system disturbances, the System Operators must rely on the installed protection and control equipment, prior to human intervention. New power electronic solutions bring rapid and repeatable responses to disturbances, which will help System Operators to guarantee a stable system. Last year, Alstom completed the world's first competitively bid STATCOM to support the voltage on National Grid's 400kV network that supplies London and the Southeast from the north of the UK. It is rated ${\pm}75MVAr$ and forms part of a Static Var System (SVS) with a total rating of 0 to 225MVAr. This paper will describe the reasons for its size, location, its chain-link configuration and give examples of its operating performance. The paper will also describe the features that allow this STATCOM to deliver much more than reactive compensation in support of a wider transmission service objective, as system conditions require.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.2
• /
• pp.450-470
• /
• 2023

Traditional cloud computing faces some challenges such as huge energy consumption, network delay and single point of failure. Edge computing is a typical distributed processing platform which includes multiple edge servers closer to the users, thus is more robust and can provide real-time computing services. Although outsourcing data to edge servers can bring great convenience, it also brings serious security threats. In order to provide image retrieval while ensuring users' data privacy, a privacy preserving image retrieval scheme in edge environment is proposed. Considering the distributed characteristics of edge computing environment and the requirement for lightweight computing, we present a privacy-preserving image retrieval scheme in edge computing environment, which two or more "honest but curious" servers retrieve the image quickly and accurately without divulging the image content. Compared with other traditional schemes, the scheme consumes less computing resources and has higher computing efficiency, which is more suitable for resource-constrained edge computing environment. Experimental results show the algorithm has high security, retrieval accuracy and efficiency.

• Journal of Internet Computing and Services
• /
• v.24 no.1
• /
• pp.17-26
• /
• 2023

A smart city is a massive internet of things (IoT) environment, where all terminal devices are connected to a network to create and share information. In accordance with massive IoT environments, millions of IoT devices are connected, and countless data are generated in real time. However, since heterogeneous IoT devices are used, collecting the logs for each IoT device is difficult. Due to these issues, when an IoT device is invaded or is engaged in malicious behavior, such as infection with malware, it is difficult to respond quickly, and additional damage may occur due to information leakage or stopping the IoT device. To solve this problem, in this paper, we propose identifying the attack technique used for initial access to IoT devices through MITRE ATT&CK, collect the logs that can be generated from the identified attack technique, and use them to identify the cause of malware infection.

• Annual Conference of KIPS
• /
• 2010.11a
• /
• pp.126-129
• /
• 2010

WBAN(Wireless Body Area Network)기반의 의료 응용으로 실시간 모니터링 시스템을 구현하였다. 특히 산소포화도 생체 센서들로부터 연속적으로 전송되는 스트림 데이터에 대해 다양한 조건을 포함하는 질의들이 실행 되는데 이러한 실시간 모니터링 질의들을 효율적으로 식별하기 위한 질의 인덱스를 설계하였다. 매번 모든 질의들을 실행하기에는 시간이 많이 걸리기 때문에 Interval Skip List 를 이용하여 빠르고 효율적으로 식별하도록 설계하였다. 이로써 위급한 상황의 환자의 건강에 문제가 생겼을 때 신속하게 대처할 수 있는 환경을 제공한다. 본 논문에서는 방대한 양의 스트림 데이터와 이 데이터를 실시간으로 감시할 수 있도록 Interval Skip List 를 스마트 메디컬 스페이스(m-MediNet)에 적용한 방법을 기술하고 있다.

• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.3
• /
• pp.166-175
• /
• 2023

Reinforcement learning can be applied to a wide variety of problems. However, the fundamental limitation of reinforcement learning is that it is difficult to derive an answer within a given time because the problems in the real world are too complex. Then, with the development of neural network technology, research on deep reinforcement learning that combines deep learning with reinforcement learning is receiving lots of attention. In this paper, two types of neural networks are combined with reinforcement learning and their characteristics were compared and analyzed with existing value-based reinforcement learning algorithms. Two types of neural networks are FNN and CNN, and existing reinforcement learning algorithms are SARSA and Q-learning.

• Annual Conference of KIPS
• /
• 2009.11a
• /
• pp.315-316
• /
• 2009

차량용 네비게이션에 사용되는 지도는 새로운 도로를 추가하거나 폐쇄된 도로를 삭제하는데 어려움을 겪는다. 이는 네비게이션 벤더에 의해 변경된 도로가 업데이트 되기 때문인데 벤더가 변경된 도로를 알아내기 위해서는 실제로 도로를 조사하는 등의 경비가 소모되므로 잦은 업데이트가 어렵다. 이런 문제를 해결하기 위해서 본 논문에서는 블루투스를 이용한 차량간 애드혹 네트워크를 이용해서 실시간으로 주변 도로의 형태를 생성하는 방법을 제안한다.

• Annual Conference of KIPS
• /
• 2009.11a
• /
• pp.63-64
• /
• 2009

자동차 내에는 여러 종류의 차량용 네트워크가 사용되고 있다. 자동차에서 요구되는 서비스의 종류가 다양해짐에 따라 차량용 네트워크의 특성 또한 다양해지고 있다. 최근 이들 네트워크 간의 협업이 중요하게 인식되어 차량용 네트워크 게이트웨이가 등장하게 되었으나 서로 다른 특성의 네트워크를 효율적으로 지원하기 위한 방안에 대해서는 아직 충분히 연구되지 않았다. 그 대표적인 예로서 서로 다른 네트워크의 통신 실시간성 요구를 만족하기 위한 효율적인 시스템 소프트웨어의 구조를 들 수 있다. 본 논문에서는 차량운전 및 안전성과 직결되는 전장용 네트워크 도메인과 인포테인먼트 시스템을 위한 네트워크 도메인을 구분하고 전장용 네트워크의 실시간성을 지원하기 위한 리눅스 기반의 차량용 네트워크 게이트웨이를 제안한다. 성능 측정 결과 제안된 게이트웨이는 인포테인먼트 네트워크 도메인의 통신에 영향을 받지 않고 전장용 네트워크 도메인의 실시간성을 보장해줄 수 있음을 보인다.

• Annual Conference of KIPS
• /
• 2009.04a
• /
• pp.856-858
• /
• 2009

최근 휴대폰, PMP 와 같은 모바일 장치를 개발하는 데에 그 성능과 저전력화 SoC 기술을 적용하고 있다. 또한 화상통화와 같은 영상 및 음성 멀티미디어 서비스가 확장되고 있다. 그러나 현재 모바일 SoC 기술에서 멀티미디어 서비스의 실시간 요구사항을 고려한 네트워크 프로토콜 설계에 대한 연구가 부족하다. 이에 본 논문에서는 실시간성 모바일 SoC 에서 실시간성을 제공하는 네트워크 프로토콜 스택을 설계하는 기법을 제안하고자 한다. 그리고 제안한 기법이 구현된 실시간 네트워킹 SoC 플랫폼의 성능을 실험하였으며, 그 결과 기존의 기법보다 더 좋음을 확인하였다.

• Proceedings of the Korea Institute of Applied Superconductivity and Cryogenics Conference
• /
• 2002.02a
• /
• pp.303-305
• /
• 2002

SSTF(Samsung Superconducting Test Facility) has been constructed at Samsung Advanced Institute of Technology to test the KSTAR(Korea Superconducting Tokamak Advanced Research) superconducting magnets and conductors. The SSTF DAC(Data Acquisition and Control) system basically consists of VME I/O modules, host PCs, and Ethernet links. VxWorks is used for the real-time OS of the VME IOC(Input/output Controller). EPICS (Experimental Physics and Industrial Control System) provides a software architecture for the communication between IOCs and host PCs. For the efficient management of measured data, the database management programs through NFS(Network File System) have been developed and successfully operated. In this paper, the current status of the SSTF DAC system, DBMS(DataBase Management System), recent test results, and future plans are presented.