中文摘要：

传感网络节点一般分布于环境复杂的区域,当前故障诊断方法诊断精度低,虚警率高,整体性能低。提出一种基于粗糙集的传感网络节点故障诊断方法,给出传感网络结构图,采用粗糙集方法,通过过滤传感网络节点故障数据不关键属性集,得到简化的节点故障属性集,保持了分类能力,提高了运行效率。通过对传感网络故障节点数据进行约简,减少传感网络节点故障数据特征向量维数。建立简化层次各异的传感网络节点故障诊断决策网络,对故障诊断决策规则覆盖度进行分析,求出所有符合既定采集规定的诊断决策规则集,完成新故障状态的处理,实现传感网络节点的故障诊断。实验结果表明,所提方法不仅具有很高的诊断精度,而且虚警率低、整体性能高。

英文摘要：

Sensor network nodes generally distributed in the environment of complex area,the current fault diagnosis methods in the diagnosis is low accuracy and high false alarm rate,the overall performance is low.A sensor network node fault diagnosis method based on rough set is proposed.The sensor network structure is given,rough set method is used,by filtering the key attrlbutes sensor network node failure data,node failure attribute sets are simplified.The classification ability is kept,the operation efficiency is improved.Through the sensor network fault node data reduction,sensor network node failure data feature vector dimensions was reduced.Different reduced levels of decision-making network,sensor network node fault diagnosis decision rules for fault diagnosis coverage analysis were established,all in accordance with established diagnosis decision rules acquisition rules was fond out,the processing of new fault state,the fault diagnosis of sensor nodes were completed.The experimental results show that the proposed method not only has high diagnostic accuracy,and low false alarm rate,the overall performance is high.