时空协作的WSNs节点异常检测算法

为了降低无线传感器网络中异常节点产生的错误信息对WSNs服务质量的不利影响，利用节点数据时空相关性建立了可靠邻居筛选模型及节点感知数据稳定性评估模型。针对现有的依赖数据时空相关性的异常检测技术中对邻节点数据及其本身可靠性的忽略，建立了一种基于历史数据及可靠邻居协作的两步节点异常数据检测算法NADST。实验结果表明，该算法能在实现高检测精度的同时将虚警率控制在较低水平，且算法具有较强的容错能力。

Node Anomaly Detection Algorithm Based on Spatial-Temporal Cooperation in WSNs

In order to reduce the adverse effect of the faulty information generated by abnormalnodes on the QoS of WSNs, a reliable neighbor screening model and data stability assessment model are established based on spatial-temporal correlation of node sensed data. Aiming at the neglect of the data of neighbor nodes and their own reliability, a two-step node anomalydetection algorithm NADST based on historical data and the cooperation of reliable neighbors is proposed. The experimental results show that the algorithm can achieve higher correct detection accuracy while controlling the false alarm rate at a low level, and has a strong fault tolerant ability.