A Technique of Filtering Dirty Data Based on Temporal- Spatial Correlation in Wireless Sensor Network

Abstract The problem of filtering dirty data in wireless sensor network is studied in this paper. To solve the problem, the technique of filtering dirty data based on temporal-spatial correlation is proposed. The technique takes advantage of the temporal-spatial correlations of sensed data, and builds a temporal-spatial correlation model. The error data, namely dirty data which are produced by sensor nodes, are filtered out through the model and are divided into temporary bad data and permanent bad data. First of all, local nodes carry out the first-time filtering to filter the temporary bad data through temporal correlation. Secondly, second-time filtering is performed to filter the permanent bad data in a cluster through spatial correlation. Sensor nodes of every cluster are alternate to be the head nodes to balance the energy consumption, thereby reducing the death rate of the nodes to extend the life cycle of the whole network.