基于分布式压缩感知的WSNs异常节点检测
|
摘要
无线传感网的节点极易被物理捕获而遭受恶意攻击,因此如何评测、识别并及时剔除内部异常节点是WSN亟待解决的问题。提出一种基于分布式压缩感知的无线传感网异常节点检测算法。首先,通过在每个异常节点检测周期中划分多个检测时隙,利用检测时隙的节点电量损耗向量之间的联合稀疏性,建立联合稀疏模型JSM-2;其次,对多个检测时隙的电量损耗向量进行压缩感知,并在检测周期结束时刻通过DCS-SOMP算法进行联合重构并判决,从而识别出无线传感网中的异常节点;最后,通过仿真验证了该算法可以降低数据收集的采样点数,有效延长网络生命周期。
Nodes in wireless sensor networks are vulnerable to malicious attacks due to physical capture. Therefore,how to evaluate,identify and eliminate internal abnormal nodes in time is an urgent problem for WSN. In this paper,an anomaly detection algorithm for wireless sensor networks based on distributed compressed sensing is proposed. First,the joint sparse model JSM-2 is established by dividing multiple detection slots in each detection period of abnormal nodes,utilizing the joint sparsity between the energy loss vectors of detection slots. Second,the energy loss vectors of multiple detection slots are compressed and sensed,and at the end of detection period,they are reconstructed and judged jointly by DCS-SOMP algorithm,so as to identify them. Finally,simulation results show that the algorithm can reduce the number of sampling points in data collection and effectively prolong the network life cycle.
Nodes in wireless sensor networks are vulnerable to malicious attacks due to physical capture. Therefore,how to evaluate,identify and eliminate internal abnormal nodes in time is an urgent problem for WSN. In this paper,an anomaly detection algorithm for wireless sensor networks based on distributed compressed sensing is proposed. First,the joint sparse model JSM-2 is established by dividing multiple detection slots in each detection period of abnormal nodes,utilizing the joint sparsity between the energy loss vectors of detection slots. Second,the energy loss vectors of multiple detection slots are compressed and sensed,and at the end of detection period,they are reconstructed and judged jointly by DCS-SOMP algorithm,so as to identify them. Finally,simulation results show that the algorithm can reduce the number of sampling points in data collection and effectively prolong the network life cycle.
引文
[1] Culler D E,Estrin D,Srivastava M B. Guest editors’ introduction:overview of sensor networks[J].Computer,2004,37(8):41-49.
[2]刘青林,艾红.基于Zig Bee的无线传感器网络监测系统设计[J].北京信息科技大学学报,2014,29(3):59-63.
[3]杨光,印桂生,杨武,等.WSNs基于信誉机制的恶意节点识别模型[J].哈尔滨工业大学学报,2009(10):158-162.
[4]刘双.基于分层的无线传感器网络入侵检测算法研究[D].乌鲁木齐:新疆大学,2017.
[5]胡宇.无线传感器网络中恶意节点的检测和安全路由技术[D].成都:电子科技大学,2015.
[6]陶砚蕴,徐萃华,林家骏.无线传感器网络的安全性研究[J].计算机安全,2007(4):8-13.
[7]杨光,印桂生,杨武,等.无线传感器网络基于节点行为的信誉评测模型[J].通信学报,2009,30(12):18-26.
[8]肖政宏,陈志刚,李庆华.WSN中基于分布式机器学习的异常检测仿真研究[J].系统仿真学报,2011,23(1):181-187.
[9] Coluccia G,Ravazzi C,Magli E. Distributed compressed sensing[M]. Compressed Sensing for Distributed Systems,2015.
[10]程银波,司菁菁,候肖兰.适用于无线传感器网络的层次化分布式压缩感知[J].电子与信息学报,2017(3):539-545.
[11]梁跃祖.无线传感器网络智能分簇及数据融合算法研究[D].南昌:华东交通大学,2015.
[12]吕方旭,张金成,石洪君,等.WSN中的分布式压缩感知[J].传感技术学报,2013(10):1446-1452.
[2]刘青林,艾红.基于Zig Bee的无线传感器网络监测系统设计[J].北京信息科技大学学报,2014,29(3):59-63.
[3]杨光,印桂生,杨武,等.WSNs基于信誉机制的恶意节点识别模型[J].哈尔滨工业大学学报,2009(10):158-162.
[4]刘双.基于分层的无线传感器网络入侵检测算法研究[D].乌鲁木齐:新疆大学,2017.
[5]胡宇.无线传感器网络中恶意节点的检测和安全路由技术[D].成都:电子科技大学,2015.
[6]陶砚蕴,徐萃华,林家骏.无线传感器网络的安全性研究[J].计算机安全,2007(4):8-13.
[7]杨光,印桂生,杨武,等.无线传感器网络基于节点行为的信誉评测模型[J].通信学报,2009,30(12):18-26.
[8]肖政宏,陈志刚,李庆华.WSN中基于分布式机器学习的异常检测仿真研究[J].系统仿真学报,2011,23(1):181-187.
[9] Coluccia G,Ravazzi C,Magli E. Distributed compressed sensing[M]. Compressed Sensing for Distributed Systems,2015.
[10]程银波,司菁菁,候肖兰.适用于无线传感器网络的层次化分布式压缩感知[J].电子与信息学报,2017(3):539-545.
[11]梁跃祖.无线传感器网络智能分簇及数据融合算法研究[D].南昌:华东交通大学,2015.
[12]吕方旭,张金成,石洪君,等.WSN中的分布式压缩感知[J].传感技术学报,2013(10):1446-1452.