Study Of Distributed Fault-tolerant Algorithms Based On Event Detection In Wireless Sensor Networks

Recent advances in embedded systems, wireless communications and digital circuits have enabled the development of multi-function, low-power sensor nodes, which are small in size and communicate in short distances. Wireless sensor network has a highly dynamic system behavior and the characteristic of Data-centric and Dynamic network self-organization.. The unique features of wireless sensor networks offer unprecedented opportunities for broad applications. Event detection is one of the most important applications. However, in real applications, The fault of sensor node itself and the rugged environment it works in often lead to the failure of sensor node, which further lead to the decline of network quality and detection accuracy. Obviously, the fault-tolerance will become a crucial problem in event detection for wireless sensor networks.In order to improve the event detection accuracy, Most of existing and typical fault-tolerant algorithms are in a distributed manner, such as distributed Bayesian fault-tolerant algorithms, improved Bayesian fault-tolerant algorithms and median based method. In these algorithms, so far achieve extremely high performance for a relatively low fault probability. As the fault probability increases, however, their performance degrades considerably. In order to overcome the low scalability problem of existing fault tolerant scheme, we present a distributed adaptive fault tolerant event detection scheme for wireless sensor networks. The starting point is to make full use of space redundant information, each sensor nodes compare its own readings with those neighbors and to decide whether the event happens or not. On the one hand, confidence levels of sensor nodes are used to dynamically adjust the threshold for decision making. Resulting in consistent performance even with increasing number of faulty nodes. In addition, the scheme employs a moving average filter to tolerate and correct most transient faults in sensor readings, reducing the effective fault probability. Simulation result shows that in the circumstance of increasing probability of faulty nodes, the faulty sensor detection accuracy and false alarm rate of this scheme achieved to a ideal result. Also some experiment is used to research the parameters in this scheme.