基于粗糙集的温室无线传感网络节点故障诊断

Rough set based fault diagnosis of node in greenhouse wireless sensor network

针对系统资源有限、不适宜运行复杂故障诊断算法的问题,提出基于粗糙集的温室无线传感网络节点故障诊断方法.结合网络体系结构,以资源丰富的汇聚节点为中枢,在运行过程中在线判别其子节点故障征兆,通过简单地查故障诊断决策表的方法来诊断子节点故障类型;从系统的、模块化的角度构建的子节点故障征兆、故障类型具有相关性,在简化决策的同时也造成了决策表的冗余,运用粗糙集及其数据约简理论设计一致数据启发式子节点故障诊断约简算法来约简冗余,建立了更简洁的子节点故障诊断决策表.结果表明,基于粗糙集能够实现在线简便的故障诊断,并可节约子节点系统资源.

In view of the problem that complicated fault diagnosis algorithm was not applicable due to the limitation of node system resources in the greenhouse wireless sensor network, a method based on the network framework was proposed to diagnose the fault type of sub-node by looking up the fault diagnosis decision-making table. The resource-rich sink node was taken as hinge to identify the sub-node fault sign online. The decision-making table was redundant due to the sub-node fault sign and fault types were con- structed from the point of view of system and module. The heuristic fault diagnosis reduction algorithm for consistent data of sub-node was designed on rough set to reduce the redundancy, and then a more concise fault diagnosis decision-making table of sub-node was established. The results show that the online simple fault diagnosis based on rough set is realized and the system resource of sub-node is saved.