Data Aggregation for Failure Tolerance in Wireless Sensor Network
One of the critical tasks in designing a wireless sensor network is to monitor, detect, and report various useful occurrences of events in the network domain which are determined by the result of data aggregation. Fault tolerance is critical to the efficiency of data aggregation scheme. One important reason is that sensor nodes are neither reliable nor stabile. In this paper, we present an improved k-means data aggregation algorithm considering the proposal of isolated point. Each cluster includes three types of sets: aggregation data, fault data set and abnormal data set. Aggregation data comes from normal sensors in this cluster through the improved K-means aggregation algorithm and abnormal nodes can be detected according to the aggregation result.