One-way Ranging for Mobile Underwater Acoustic Networks with Long Interaction Periods

One of the main challenges in underwater positioning is a correct estimate of the sound speed in the deployment area. In fact sound speed can be either directly measured with a velocity meter, or indirectly computed from Conductivity Temperature and Depth (CTD) measurements. While the former approach requires the use of very precise devices that need to be often calibrated, the latter requires the use of extremely precise CTDs that are usually very expensive. The availability of low-cost unmanned underwater vehicles and low-cost acoustic modems makes the integration of both velocity meters and CTDs impractical, as the cost of these devices will be in the same order of magnitude of the unmanned vehicle itself. On the other hand, the recent availability of large data-sets of sound speed measurements in various parts of the world, and the efficient implementation of Machine Learning algorithms that can nowadays run also in embedded devices, suggest a new approach based on sound speed pred...