基于无线传感网络的通信数据异常检测方法

Communication Data Anomaly Detection Method Based on Wireless Sensor Network

环境变化及无线传感器自身参数畸变可使无线传感通信数据极易出现异常。为此,本文提出一种基于无线传感网络的通信数据异常检测方法。构建通信数据时间序列模型,引入滑动窗口实时更新传感器采样数据,通过置信区间和区间差异度对比正常数据和异常数据,并给出2个判断条件。在仿真实验中,选取正常节点和异常节点检测率/误报率作为算法的评判指标。测试结果表明,与其他3种算法相比,该方法具有最高的检测率和最低的误报率。

The environmental changes and distortion of wireless sensor parameters can make the wireless sensor communication data prone to abnormal. Therefore, a communication data anomaly detection method based on wireless sensor network is proposed. The time series model of communication data was constructed, and the sliding window was introduced to update the sensor sampling data in real time. The normal data and abnormal data were compared by confidence interval and interval difference, and two judgment conditions were given. In the simulation experiment, the detection rate/false alarm rate of normal nodes and abnormal nodes were selected as the evaluation index of the algorithm. The test results show that this method has the highest detection rate and the lowest false alarm rate compared with the other three algorithms.