Research On WSN Fault Detection For Spatial-Temporal Correlation Analysis

With the promotion of intelligent construction and development,Wireless Sensor Network(WSN)has broader application prospects in military,medical health,environmental science,smart home and other fields,which also puts forward higher requirements for WSN performance.Due to its own characteristics such as dynamics,limited energy,and complex environment,sensor nodes that are in a long-term monitoring state are prone to death or damage,resulting in distortion of monitoring data.This is a big challenge for WSN to provide high-quality services,and timely detection of network failure nodes is an important prerequisite for ensuring network reliability.Node-aware data has a high spatiotemporal correlation in applications involving observing changes in physical quantities.This article is to study the WSN fault detection method for this feature.The work is mainly as follows:In order to solve the problem of abnormal data collection caused by sensor node failure,this paper proposes a WSN fault detection method based on node multi-attribute association relationship,which combines sensor data spatial correlation and multi-attribute correlation to perform deeper fault detection Explore.Based on the non-uniform clustering network structure,the fusion distance factor and the spatial correlation analysis between adjacent nodes estimate the cluster head confidence interval.Under the condition of reliable cluster heads,a saliency test is used to make a feedback decision on the state of the nodes in the cluster.Finally,the abnormal source of the faulty node is determined according to the correlation between the attributes of the node,so as to avoid the interference of the abnormal event on the node detection.Simulation experiment results show that the algorithm improves the failure detection rate of nodes under various failure scenarios and can effectively reduce the communication cost.Distributed fault detection is an important issue in WSN.For large-scale WSNs with dense deployments,this paper combines the time-space correlation characteristics of WSN nodes with the learning idea of timing difference strategy to propose an intelligent node fault diagnosis algorithm to ensure network reliability.The algorithm uses a linear programming model to achieve uniform clustering of the network,uses the perceived state of neighboring nodes in the cluster as environmental feedback information,and incorporates a hybrid leapfrogging mechanism to estimate the data sequence of nodes within the time period ?.After determining the fault state of the node,short-term correction of the faulty node is performed in the manner of prediction sequence substitution.Simulation experiment results show that the improved time-space correlation detection algorithm improves the detection accuracy and adaptability of the diagnostic mechanism.