Energy-efficient localization in networks of underwater drifters

In their quest to study the intricate underwater ecosystems, oceanographic scientists need spatially-rich data that is collected within the moving reference frame created by the oceans' currents. To this end, we are developing a system of networked underwater ocean explorers that drift freely with the currents. However, effective interpretation of collected sensor data hinges on knowing the positions of these drifters while submerged. Due to their uncontrollable motion, positions have to be tracked in time. Unlike in static networks, position estimation therefore represents a recurring cost, which should be minimized for reasons of limited on-board energy supplies. In this paper, we present a low-overhead scheme that is able to trade accuracy for energy savings by cleverly selecting the specific links that are required for the self-localization algorithm. Our technique is also agile enough to quickly adapt to on-demand changes in accuracy requirements, and cuts the position estimation costs by 40% or more under various ocean current dynamics.