利用置信规则库构建WSN节点故障检测模型

在无线传感器网络(wireless sensor network, WSN)节点故障检测领域的研究过程中，故障检测准确率会受节点数据的不确定性和专家知识模糊性的影响。针对这一问题，本文提出了一种基于置信规则库(belief rule base, BRB)的WSN节点故障检测方法。首先，根据WSN工作原理及节点工作特性描述WSN节点故障检测过程；然后，从空间和时间2个维度对节点数据提取特征，建立基于空间和时间相关性的WSN节点故障检测模型；最后，利用Intel Lab Data无线传感器数据集进行案例研究以验证模型的有效性。结果证明，本文方法能够统筹利用专家知识和节点数据实现WSN节点故障检测。

Constructing a WSN node fault detection model using the belief rule base

The WSN node fault detection accuracy is affected by uncertain factors, including the uncertainty of node data and the ambiguity of expert knowledge. This paper proposes a WSN node fault detection method based on the belief rule base. First, the WSN node fault detection process is described according to the working principle of the WSN and the working characteristics of the node. Node data are extracted from two dimensions of space and time, and then the WSN node fault detection model is established based on space and time correlation. We use the Intel lab data wireless sensor data set to conduct a case study to verify the effectiveness of the model. The experimental results indicate that the method proposed in this paper can coordinate the use of expert knowledge and node data to realize the fault detection of WSN nodes.