Abstract:

In general, the fault detection in wireless sensor networks (WSNs) could be happened in additional communication and computation load. Considering the feature of WSNs that those nodes in the same coverage range have the close reading data, a neighbordata analysis based fault detection schema is proposed. The confidence level of sensor nodes is taken by analyzing the history reading data in the first place. Comparing the node’s reading data with that of neighbor nodes results in good confidence level to determine that whether this node is failure or not. Simulations show that the proposed schema can catch a good performance in fault detection accuracy and transient fault tolerance, but do not bring any communication and computation overhead to sensor nodes.

Key words: wireless sensor network, fault detection, neighbor nodes, confidence level, fault tolerance