Modeling And Analysis Of High Reliable Wireless Sensor Networks For Transient Faults

Wireless sensor networks are widely used.These require actuator in large areas and are sensitive to transient faults.In the two different views,we study the influence of the transient faults in the wireless sensor network.In the peer to peer node's level,this paper proposes double control flow method for transient fault detection.The instruction tag is the kernel of the proposed method.This method not only detects the control flow errors for instruction data errors in communication by storing and comparing the program execution path,but also participates in detecting the control flow errors from internal node.Compared with the RSCFC(Relationship Signatures for Control Flow Checking)algorithm and SCFC(software-based control flow checking)algorithm,the proposed method was able to detect effectively control flow errors of communication.Moreover,it was characterized by higher fault coverage and lower overhead.It is a pure software method without hardware constraints.Therefore,the proposed method can be widely applied to wireless sensor networks.In the network's level,this paper proposes a quantitative analysis method for transient fault propagation in wireless sensor networks.This method refers to the research of transient faults in field of software.We extend the research method to the field of wireless sensor network.This quantitative analysis method can compute the influence rate of the input and the average drop rate of the node to find key inputs and key nodes.The method is proved by simulation experiment for chain network,star network and semi-autonomous network.These contributions are in favor of providing more effective algorithms to improve network reliability.In the future research,we will analysis the structural characteristic of loops in the program and improve the design of the double control flow detection algorithm to reduce blind spots.Moreover,we will also analysis the correlation of transient faults when they propagate in the wireless sensor network and consider the differences of nodes to improve the quantitative analysis method for transient faults.