Improved Compressive Sensing based on Sampling Rate Adjustments in Wireless Sensor Networks

This paper presents compressive sensing (CS)-based approach, that aims at finding a more energy-efficient deployment strategy of a wireless sensor network (WSN). The proposed approach takes advantage of the sparse nature of most natural data in order to minimize the number of the retrieved compressed samples, and thereby reduces the energy consumption of the deployed sensor nodes. Throughout the work, various CS-based reconstruction methods have been investigated and tested with regard to their efficiency in order to select the optimal approach. Then, a real-time scenario that applies the CS process in a small network has been investigated, comprising resource-limited nodes. Furthermore, the paper proposes a sample rate adjustment (SRA) scheme that adapts CS methods to the characteristics of the data. It exploits the variability of the data sparsity to further reduce the number of samples and minimize energy consumption. To test the performance of our approach, we implemented small programmable wireless modules, that are widely used for environmental monitoring scenarios. The final results prove that the energy consumption was immensely reduced, up to 61.96%, which contributes to prolonging the lifetime of the deployed batteries.