Distributed Successive Measurement Selection Based on Online Sparsity Inference

Considering the limitations on communication capability in the big data era, measurement selection plays an important role in obtaining the desired information by collecting only a part of data from the sensors. In this paper, we study the large-scale measurement selection problem, and propose a distributed algorithm exploiting the sparsity property extracted from the on-line data processing. Different to the existing works, we propose a mission-oriented framework to analyze the performance improvements for the specific mission of collecting new data. Specifically, a Bayesian hierarchical prior is adopted in order to quantify the importance of uncollected data by the on-line inference from the collected data. Based on the sparsity property obtained by on-line data processing, the sensors with important uncollected data will have high priority to access. Due to the massive number of sensors, the measurement selection algorithm is executed distributively at each device according to the common information broadcast by the fusion center. Simulation results demonstrate the performance gain of our proposed measurement selection method compared to the conventional schemes.