Feature-Level Data Aggregation for Underwater Acoustic Networks Based on Distributed Function Approximation and Cross-Layer Design

In underwater acoustic sensor networks (UASNs), feature-level data of targets from each sensor often need to be collected for further processing. Given that feature-level data of each sensor come from the same target, spatial correlation among these data is natural. Surprisingly, there has been a very limited exploration into leveraging this spatial correlation for UASN data aggregation. In this article, we have addressed feature-level data aggregation for UASNs by exploring and leveraging the spatial correlation of data from different sensors. We have proposed a model where various sensor data can be effectively represented through spatial functions. We have also proposed a feature-level data aggregation algorithm based on distributed function approximation, and thus, only a small number of sensors need to transmit their data. We also utilize a cross-layer design combined with the carrier sense multiple access (CSMA) protocol to further accelerate the data aggregation. We have provided simulation results to demonstrate the performance of the proposed algorithm.