WSN中基于被动端到端的故障链路集推断算法  

Fault link set inference based on passive end-to-end measurements for wireless sensor networks

作　　者：许可[1] 王剑[1] 尚凤军[1] 

机构地区：[1]重庆邮电大学计算机科学与技术学院,重庆400065

出　　处：《计算机应用研究》2015年第7期2123-2127,2132,共6页Application Research of Computers

基　　金：重庆市自然科学基金资助项目(cstc2012jj A40038);重庆市基础与前沿研究计划资助项目(cstc2013jcyj A40023);物联网发展专项资金支持项目(工信部科[2012]583号)

摘　　要：为了节省故障定位所需的网络能耗,给出了基于被动端到端的启发式贪婪故障链路推断算法。该算法基于被动端到端的数据建立故障链路推理模型,推断网络中最可能故障的链路集。使用端到端的数据计算各条路径的丢包率,通过与阈值的比较对路由矩阵进行简化,目的是去掉根据端到端的数据可以判定为好的路径。由于该推理模型需要各条链路的故障概率,通过简化矩阵的方式优化算法LIPM(loss inference based on passive measurement),给出了推测效果更好的LISC(loss inference based on passive measurement and set-cover)算法。仿真结果显示,新算法具有更好的精确度和运行效率。最后,迭代计算各条链路的权值,通过每次选择最优权值的链路并删除包含此链路的路径,直到所有的故障被解释为止。通过仿真验证了新算法具有更好的故障覆盖率。This paper proposed a greedy heuristic algorithm using passive end-to-end measurements and the link loss rate as an indicator. In order to further save energy, it infered the maximum probability of the faulty link sets based on passive end-to- end data. Firstly, it reduced the dimension of the route matrix to remove the good path from the path set by comparing the pass loss rate and threshold. Secondly, it optimized existing algorithms to infer the failure probability of the bad link set. The result shows that the new algorithm has better accuracy and CPU efficency. Finally, it iteratively calculated the weight of each link, and removed the link which had the optimal weights and the path which included the above link, until all the faults were interpreted. Experiments show that the new algorithm has higher fault coverage.

关 键 词：无线传感器网络 故障链路定位 故障链路推断 启发式算法 丢包率 

分 类 号：TP393.03[自动化与计算机技术—计算机应用技术] TP301.6[自动化与计算机技术—计算机科学与技术]