A Fault Diagnosis Method based on Node Dynamic Behaviour Status Prediction in Distributed Sensor Network
Abstract
The geographical distribution of network systems is to provide a range of services that require constant monitoring and repair to maintain clock uptime. Failure of the service will result in a huge loss in business or a high degree of dissatisfaction with the user. Therefore, it is necessary to monitor regularly to identify faults and recover them to work perfectly. However, it is still challenging to diagnose the occurrence of faults infrequent behavior state changes of nodes with correctness and low latency. To address this issue, we proposed a Fault Diagnostic Method (FDM) that monitors node Dynamic Behavior State (DBS) predictions to predict the probability of faults to ensure node functional status within a limited time limit. FDM aims to achieve marginal accuracy by quantifying the dynamic state change predictions in the system through regular checks during the defined period, through implementing the DBS algorithm. The FDM-DBS algorithm has been evaluated in arbitrary topologies of wireless sensor network systems to achieve accuracy and low latency diagnostics compared to modern mechanisms.