本文采用Markov线性预测模型,为无线传感器网络设计了一种基于流量预测的拒绝服务攻击检测方案——MPDD。在该方案中,每个节点基于流量预测判断和检测异常网络流量,无需特殊的硬件支持和节点之间的合作;提出了一种报警评估机制,有效提高方案的检测准确度，减少了预测误差或信道误码所带来的误报。仿真实验结果表明,Markov模型具有较高的预测精度,能够实时地预测传感器网络流量;MPDD方案能够快速、有效地检测拒绝服务攻击且消耗资源较少。

In this paper,the authors propose a lightweight denial of service (DoS) attacks detection scheme based on  a simple and efficient Markov traffic prediction model,MPDD (Markov Prediction based DoS attack Detection). In MPDD,each node detects independently Through predicting the traffic of Wireless Sensor Network,which neither special hardware nor node’s cooperation .An algorithm evaluating the reliability of alert is proposed to reduce the false alerts caused by prediction or channel errors. Simulation results has shown that the Markov model can predict sensor network traffic precisely; MPDD is an efficient DoS attacks detection scheme which can quickly detect various DoS attacks with less resource overhead.