改进的马尔科夫模型的异常节点检测算法

Abnormal node detection strategy based on improved Markov model

为降低异常节点对移动自组织网络路由性能的影响,提出一种基于马尔科夫模型的异常节点检测策略。将移动自组织网络中各个节点的状态转换过程看作一个马尔科夫过程,采用改进的马尔科夫模型计算各个节点的马尔科夫预测因子,依据马尔科夫预测因子判断节点是否异常。仿真结果表明,在异常节点比例不同的情况下,该策略的异常节点检出率均高于87%。采用该策略改进的AODV路由协议具有报文送达率高、端到端平均延时小和网络吞吐量大的优点。

To decrease the influence on the performance of mobile ad-hoc network routing caused by abnormal nodes,an abnormal node detection strategy based on Markov model was proposed.The nodes’ states transition in mobile ad hoc network was taken as a Markov process,and Markov model was used to calculate the Markov predictor of each node,for the purpose of determining whether a node is abnormal or not based on the Markov predictor.Simulation results show that the abnormal nodes detection rate of the proposed strategy is above 87% under different proportions of abnormal nodes.The improved AODV routing protocol using this strategy has advantages such as high packet delivery rate,low end-to-end average delay and large throughput.