Study Of Hierarchical Fault Diagnosis Methods In Wireless Sensor Network

Wireless sensor network is a self-organizing network which consisted of large-scaleof tiny sensor nodes and deployed in environment to monitor some interest things oftarget area. Due to numerous nodes, working in complicated environment is and itsuncertainty, wireless sensor network prone to failure, and fault symptoms are complex.Fault diagnosis monitor the status of the wireless sensor network in real-time, preventfault occur and eliminate fault and so on. So the fault diagnosis to the wireless sensornetwork apply reliability is of great significance.Nodes and the network are closely coupled in wireless sensor network system,especially in the case of large-scale applications, node failures and network failuresinteraction increases the complexity of the fault diagnosis. From the perspective of ahierarchical network, the characteristics were studied in each layers of fault diagnosis inthis article, Designed a methods of fault diagnosis system with fault detection capability,fault identification capabilities. The main research results in this paper are described asfollows:①Through the research work of existing induction and consolidation, The mainmethod of fault diagnosis for wireless sensor networks is targeted only to the nodes oronly consider the case of a network failure, To solve this problem, Proposed a methodfor wireless sensor networks hierarchical fault diagnosis, wireless sensor network isdivided into nodes layer and system layer, Analysis of the characteristics and difficultiesof each layer, propose different solutions for different layer.②For the most present fault detection methods for wireless sensor networks haveassume that the fault node are based on outliers data, it have limitations. Proposed awireless sensor network node failure detection method based on similarity measure innode layer, according to data which is correlation of time and space of node, andproposed an improved similarity measure that acted as the basis for the credibility of theadjustment which detects node failure to improve the detection accuracy and reliability.Discussed the general problem in different types of failure of sensor node failuredetection scheme.③For the difficulty in obtaining faulty sample characteristics in wireless sensornetwork with large-scale of nodes, proposed wireless sensor network fault diagnosismethod based on immune mechanism in system layer. Integrated biological immune mechanism into the wireless sensor network fault diagnosis, and proposed detectorgeneration, cloning, mutation and immune suppression corresponding solutions.Analysis antigen characteristic parameters through wireless sensor network simulationby NS2, and verify the effectiveness of the wireless sensor network fault diagnosismethod based on immune mechanism.Proposed similarity-based node failure detection based on a hierarchical network,and immune theory for wireless sensor network fault diagnosis problem tentatively.Simulation results show that the fault diagnostic capability of the two methods.