Redundancy control in large scale sensor networks via compressive sensing

In wireless sensor networks for smart city or smart planet applications, massive volumes of real-time sensory data are being generated in every second, which pose great challenges to the power-limited sensor nodes, bandwidth-limited transmission links, and require high data storage and management costs. To deal with these challenges, compressive sensing (CS) converts the the spatially and temporally correlated information to sparse signals in some transformed domains (Such as DCT and FFT), and conducts cost-efficient, low-rank sensing. This paper presents a cost-centric comparison between recent compressive sensing solutions, i.e., Compressive Data Gathering (CDG) and Compressive Sparse Function (CSF), with traditional sensing technologies, in the means of sensing, transmission, storage and computation costs. It shows by a city temperature collection example that CDG performs similarly to CSF, both of which can prolong the network lifetime for almost one magnitude than traditional multi-hop sensing, while providing enough information for recovering the temperature distributions.