A two-stage data fusion model for wireless sensor networks

Wireless sensor networks (WSNs) are widely applied in many industrial and consumer fields, and data fusion arises as a critical discipline concerned with how data collected by sensors can be processed. However, existing research results on data fusion cannot achieve the optimal performance of the accuracy, the processing speed and the network life-span simultaneously. In this paper, a two-stage data fusion model is established. On the basis of this model, a fusion matrix is constructed to get rid of the redundant data so as to reduce the data fusion time at the first stage. Then strategies of BP neural network are adopted at the second stage to fuse data for more confident ones, which guarantees the fusion accuracy further. Simulations and experiments show that the performance of both the accuracy and real-time property is much improved.