Research On Fault Diagnosis Algorithm Based On Voting Mechanism In WSN

Wireless sensor network(WSN)is a multi-hop and self-organizing network composed of a large number of micro-sensor nodes deployed in the monitoring area through wireless communication.With the continuous development of micro-electronics and communication technology,WSN has been widely used in medical health,traffic management,environmen?tal monitoring and industrial manufacturing.However,in practical applications,there are always some constraints on the performance of sensor nodes.On the one hand,sensor n-odes usually use relatively cheap chips in order to save hardware costs,which limits the computing and communication capabilities of the nodes.On the other hand,nodes are often deployed in relatively harsh and complex environments,natural factors and electromagnetic interference will affect the performance of nodes.This makes nodes prone to failure in the process of operation.The fault node will not only detect the wrong data,but also interrupt the cormunication between the networks,thus resulting in decision-making errors of sys?tem.Therefore,in order to improve data quality,strengthen network security and extend network life,fault diagnosis in WSN has become an important research topic in the field of WSN.In this paper,the fault diagnosis algorithms of WSN are deeply studied.The main work is as follows:(1)A parallel and local diagnosis algorithm based on neighbor representative nodes is proposed.In a large-scale WSN,the global diagnosis algorithm needs to obtain the measure-ment information of all nodes in the network,which causes a large number of unnecessary tests and excessive consumption of node power.To solve this problem,this paper proposes a local diagnosis algorithm based on neighbor representative nodes.The algorithm performs diagnostic tests in parallel,which reduces the diagnosis time and avoids the signal collision between nodes.(2)A dual-clustering algorithm is proposed.In this paper,clustering algorithms based on WSN in recent years are studied.However,these algorithms do not take both sensing val?ues and geographical location of nodes into account.The dual-clustering algorithm divides nodes with similar measured values and close geographical location into the same cluster.It is also applicable to nodes monitoring multi-dimensional values at the same time,which makes the clustering algorithm more reasonable and improves the diagnostic performance of the network.(3)A fault diagnosis algorithm based on node reliability model is proposed.Traditional fault diagnosis algorithms pursue the accuracy of fault diagnosis unilaterally,but ignore the cost of sensor nodes participating in diagnosis.This paper designs a reliability model of the node based on the historical diagnosis results and residual energy of the node.The algorithm selects nodes that have recently been diagnosed as fault-free nodes and with the mOost residual energy to participate in the diagnosis.The experimental results show that this algorithm can not only improve the accuracy of diagnosis and reduce the false alarm rate,but also save the energy consumption of nodes,thus extending the life of the network.