In this paper, we explore the problem of data acquisition using compressive sensing (CS) in wireless sensor networks. Unique properties of wireless sensor networks require we minimize communication cost for efficient power usage. At first, a compressive distributed sensing (CDS) algorithm is proposed but is then modified to decrease communication costs. The final algorithm presented is compressive distributed sensing with random walk CDS(RW); an algorithm that combines the data gathering and projection generation process of CDS.CDS(RW) uses rateless encoding, graph algorithms, and belief propagation decoding to improve upon the communication cost associated with CDS. In the end, we show that the communication cost of CDS(RW) versus existing CS algorithms is far superior, while still having satisfactory decoding accuracy. |