Lifetime Maximization of an Internet of Things (IoT) Network Based on Graph Signal Processing

The lifetime of an Internet of Things (IoT) system consisting of battery-powered devices can be increased by minimizing the number of transmissions per device while not excessively deteriorating the correctness of the overall IoT monitoring. We propose a graph signal processing based algorithm for partitioning the sensor nodes into disjoint sampling sets. The sets can be sampled on a round-robin basis and each one contains enough information to reconstruct the entire signal within an acceptable error bound. Simulations on different models of graphs, based on graph theory and on real-world applications, show that our proposal consistently outperforms state-of-the-art sampling schemes, with no additional computational burden.