中文摘要：

针对无线传感网（WSN）系统故障的诊断问题,提出一种基于时间加权K-近邻法的无线传感网系统故障诊断方法。该方法按照系统故障机理确立特征值,根据WSN系统故障的时间相关性设计基于时间加权的故障诊断分类规则,并结合K-近邻法建立系统故障诊断模型。实验仿真表明,针对WSN系统故障诊断问题,时间加权K-近邻法具有较好的抗干扰性和适应性,与普通K-近邻法、SVM（支持向量机）相比诊断分类性能明显提升。

英文摘要：

Addressing to problems of WSN system fault diagnosis,this paper proposed an approach for diagnosis of system fault based on time-weighted K-nearest neighbor for WSN. The method established characteristic value according to system fault mechanism,and designed time-weighted rule of fault diagnosis classification on the basis of time correlation of WSN system fault,built a diagnosis model of system fault combining K-nearest neighbor as well. The experiment simulations indicate that the presented method processes balance between anti-interference and adaptability and improves significantly on WSN system fault diagnosis than normal K-nearest neighbor and SVM.