http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
무선 애드혹 네트워크에서 우선순위를 고려한 분산 노드 스케줄링
이원석(WonSeok Lee),김태준(Taejoon Kim),김태홍(Taehong Kim),이재생(JaeSeang Lee),함재현(Jae-Hyun Ham) 제어로봇시스템학회 2020 제어·로봇·시스템학회 논문지 Vol.26 No.2
Wireless networks are deployed in a variety of environments and must deliver high-priority data to a destination node as quickly as possible. For this reason, several node-scheduling algorithms have been proposed and some distributed scheduling algorithms make the load among nodes semi-equal through slot exchange. In spite of the distributed scheduling nature, they achieve similar performance with centralized scheduling schemes. However, the enhanced processing for high-priority packets has not been considered. Accordingly, this paper proposes an algorithm that uses multi-queues for the purpose of slot preemption in high-priority packets. Using a network simulator, we compared the performance of the proposed algorithm and the existing scheduling algorithm under various simulation environments in which the generation ratio of high priority packets was changed. Through the simulation, we can confirm that the proposed algorithm showed excellent performance.
Time Synchronization Robust to Topology Change Through Reference Node Re-Election
( Young Jeon ),( Taehong Kim ),( Taejoon Kim ),( Jaeseang Lee ),( Jae-hyun Ham ) 한국정보처리학회 2019 정보처리학회논문지. 컴퓨터 및 통신시스템 Vol.8 No.8
애드혹 네트워크에서는 하나의 기준노드를 선정하여 이를 중심으로 네트워크 내에 모든 노드들의 시간을 동기화하는 방법을 사용할 수 있다. 이러한 기준노드를 중심으로 하는 대표적인 시간동기화 알고리즘은 Flooding Time Synchronization Protocol (FTSP) 이다. 메시지를 주고 받는 과정에서 예측할 수 있는 지연과 예측할 수 없는 랜덤 지연이 발생하게 되는데, 이러한 지연은 정확한 동기화를 방해하기 때문에 제거해야 한다. 멀티홉 기반의 통신에서 메시지를 주고받는 과정에 지나는 홉 수에 따라 홉 지연이 발생하게 된다. 이러한 홉 지연은 노드 사이의 동기화 성능을 크게 떨어뜨리게 된다. 따라서 이러한 홉 지연을 줄이고 동기화 성능을 높이기 위한 방법이 필요하다. 기존 FTSP 방식에서 기준노드가 ID를 기반으로 가장 ID 값이 작은 노드가 기준노드로 선정되기 때문에 기준노드의 위치가 어디냐에 따라서 홉 지연에 따른 성능 저하가 크게 발생할 수 있다. 본 논문에서는 홉 지연을 줄이기 위한 최적의 기준노드 재선정 알고리즘을 제안하고 OPNET 네트워크 시뮬레이터를 사용하여 기존 FTSP와의 성능을 비교한다. 추가적으로 토폴로지 변경에 따른 성능을 측정하여 제안된 방식이 토폴로지 변환에 강인한 성능을 갖고 있음을 확인할 수 있었다. In an Ad-hoc network, a method of time synchronizing all the nodes in a network centering on one reference node can be used. A representative algorithm based on a reference node is Flooding Time Synchronization Protocol (FTSP). In the process of sending and receiving messages, predictable and unpredictable delays occur, which should be removed because it hinders accurate time synchronization. In multi-hop communications, hop delays occur when a packet traverses a number of hops. These hop delays significantly degrade the synchronization performance among nodes. Therefore, we need to find a method to reduce these hop delays and increase synchronization performance. In the FTSP scheme, hop delays can be greatly increased depending on the position of a reference node. In addition, in FTSP, a node with the smallest node ID is elected as a reference node, hence, the position of a reference node is actually arbitrarily determined. In this paper, we propose an optimal reference node election algorithm to reduce hop delays, and compare the performance of the proposed scheme with FTSP using the network simulator OPNET. In addition, we verify that the proposed scheme has an improved synchronization performance, which is robust to topology changes.
김용정(Yong Jeong Kim),Linh-An Phan,김태준(Taejoon Kim),김태홍(Taehong Kim),이재생(JaeSeang Lee),함재현(Jae-Hyun Ham) 제어로봇시스템학회 2019 제어·로봇·시스템학회 논문지 Vol.25 No.12
Gradient Time Synchronization Protocol (GTSP) is a consensus-based time synchronization protocol in which each node adjusts its logical clock by averaging the relative clock rate of neighbor nodes in every beacon interval. This process is repeated until the network achieves convergence of time synchronization. Therefore, a short beacon interval can reduce the convergence time, although it causes high energy consumption because messages are sent frequently. In contrast, a longer beacon interval can reduce energy consumption but may result in a longer convergence time. In this paper, we propose the Adaptive Gradient Time Synchronization Protocol (AGTSP), which enables dynamic adjustment of the beacon interval of each node in a fully-distributed manner to reduce convergence time and energy. A performance evaluation demonstrates that AGTSP provides superior performance to GTSP as measured by convergence time, synchronization errors, and beacon transmission overhead.