基于等待时间和信道状态的轮询多址协议  

Polling Multiple Access Protocol Based on the Waiting Time and Channel State

作　　者：张宇眉[1] 张欣[1] 杨大成[1] 赵东风[2] 

机构地区：[1]北京邮电大学电信工程学院,北京100876 [2]云南大学通信工程系,昆明650091

出　　处：《北京邮电大学学报》2008年第4期126-129,共4页Journal of Beijing University of Posts and Telecommunications

基　　金：国家自然科学基金项目(60362001,F0424104)

摘　　要：提出了基于分组等待时间和信道状态的动态限定轮询多址协议.该协议根据各节点内分组的当前等待时间及信道状态动态地控制轮询过程,每个轮询周期对具有分组最长等待时间和好的信道状态的节点进行服务.在假设信道为两状态马尔可夫模型的基础上,通过采用嵌入马尔可夫链理论和休假排队模型,给出该协议的理论模型和性能分析,并且通过与基本轮询系统对比分析,可以看出所提系统缩短了分组的平均等待时间,提高了系统的吞吐量.Dynamic limited polling multiple access protocol based on waiting time of packets and chan- nel state is proposed. This protocol adaptively controls the polling process according to the instantaneous waiting time of packets and the channel state for each node, and serves the one with the longest packet waiting time and good channel state first in each polling cycle. By means of imbedded Markov chain theory and vacation queue model, the model and its performance analysis of that protocol are given assuming that channel is modeled as a two-state Markov chain. By comparing with the basic polling system, it is shown that the proposed system can reduce the mean waiting time of packets and increase system throughput.

关 键 词：等待时间和信道状态 轮询多址 Gillbert-Elliott信道模型 休假排队模型 平均时延 吞吐量 

分 类 号：TN929.53[电子电信—通信与信息系统]