粗糙集-神经网络集成的WSN节点故障诊断

提出一种粗糙集-神经网络集成的无线传感器网络(WSN)节点故障诊断新方法。根据无线传感器网络的应用环境和故障特征得到诊断决策表,利用改进的粗糙集中的归纳属性约简算法对诊断决策表进行属性约简,用Hamming网络建立一套故障分类的方法。仿真实验结果显示,该诊断算法在进行WSN节点故障诊断时,诊断准确性高,通信代价小,能耗低,鲁棒性高。

Fault Diagnosis of Node in WSN Based on Integration of Rough Sets and Neural Network

In the paper, a novel fault diagnosis method integrating rough sets and neural network in wireless sensor network (WSN) is presented. Firstly, the decision-making table of fault diagnosis can be gained by application environment and fault characteristic in WSN. Secondly, the attribute reduction of the table of decision-making is implemented by inductive attribute reduction algorithm in rough sets theory. Finally, a set of method for fault classification is founded by hamming network. The result of simulation shows that this method has the feathers such as high veracity of diagnosis, a little expenditure of communication, low energy consumption, and strong robustness.