A selective data fusion approach for coded cooperative wireless sensor networks

In this paper, we study relay and cooperative data gathering wireless sensor networks, where the system´s output is the outcome of information fusion that is not only performed at the final destination, but is first distributed among several intermediate processing nodes. We propose a selective stream fusion procedure at clustering and sink nodes, that aims to enhance the decoding performance as well as to increase channel availability. The optimal selection takes into account several parameters, like the network configuration, the sensing quality, the transmission conditions in the different hops, etc., and yields important gains in bit error rate performance compared to the default fusion process.