Research On Reliability Evaluation Method For Wireless Sensor Networks

With the rapid development of wireless sensor network technology and its extensive application in military,intelligent home,power grid and environmental monitoring,WSN reliabilityevaluation hasbecome one of the key points that can not be ignored.It is of theoretical significance and practical value to evaluate the reliability of WSN.Based on possibility theory,percolation theory and fuzzy neural network,the reliability of WSN is described in this paper.On this basis,the evaluation model of reliability is constructed,and the simulation experiment and analysis are carried out.The main contents are as follows:(1)Based on possibility theory,the reliability of WSN is defined as the probability that the state function of sensor nodes is not less than a certain value.The lifetime of sensor nodes is described by probability distribution,and the relationship between the lifetime of sensor nodes and the life of WSN is constructed by the constructor.Finally,the reliability function for evaluating the reliability of WSN is constructed.(2)Based on the percolation theory,the WSN failure process is regarded as a phase transition process.The relationship between node reliability and WSN reliability is established by the critical threshold of phase transition,and the WSN reliability function is obtained.(3)Based on the fuzzy neural network,the 14 important indexes of WSN are taken as the input of the model and the reliable matrix as the output of the model.A neural model was designed to evaluate its reliability from the WSN system level.