分布式卫星网络故障诊断算法  

A Distributed Fault Diagnosis Algorithm for Satellite Network

作　　者：江玉洁[1,2] 姚晔[1,2] 梁旭文[1] 

机构地区：[1]中国科学院上海微系统与信息技术研究所,上海200050 [2]中国科学院研究生院,北京100049

出　　处：《小型微型计算机系统》2013年第11期2518-2523,共6页Journal of Chinese Computer Systems

摘　　要：结合低轨通信星座系统自主运行发展的需求,提出一种适合于星上自主处理的分布式网络故障诊断算法(DFDS).DFDS算法基于系统级故障诊断理论(SLD),给出M-概率的分布式测试模型,依据小概率忽略的诊断规则,通过测试图预修正、二次判决后处理等风险消减措施,达到识别故障的目的.仿真实验和分析表明该算法能够有效地提高故障诊断的正确率和完全率,同时在诊断开销和诊断时延方面也较同类算法性能有明显优势.For the demand of LEO constellation autonomous operation, the paper proposed a novel distributed fault diagnosis algorithm （DFDS） for autonomous processing in satellite network. Based on the theory of SLD, DFDS designs the M-probability of testing model and the diagnosis rules of ignoring small probability events, DFDS identifies the faults by pre-modifying the testing graph and re-judging, Simulation analysis indicates that DFDS can enhance the accuracy and completeness and it is also distinctly superior to the similar algorithms in overhead and latency.

关 键 词：卫星网络 系统级故障诊断 分布式 自主 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]