Data aggregation and recovery in wireless sensor networks using compressed sensing

QoS support for data aggregation in large-scale multi-hop wireless sensor networks WSNs inevitably faces two crucial issues: packet loss and energy dissipation. Fortunately, most sensing data is spatially and temporally correlated and compressible. Therefore, compressed sensing CS is a promising reconstruction scheme having the potential of packet error correction with low-energy consumption. In this paper we present such a CS-oriented data aggregation technique for the multi-hop topology. Our scheme is balanced in energy consumption among the nodes and recovers lost packets at fusion centre without additional transmitting costs. Simulations show that our approach works well even for 50% data loss rate when environmental data is sparse in a certain domain. Comparing with the existing methods, our method achieves higher recovery accuracy and less energy consumption on TinyOS. Furthermore, the system is demonstrated in the experiment of monitoring grid computer facilities set up at Shenzhen Institutes of Advanced Technology.