Methods
Peak Association Analysis
The system consists of three hydrophones, each one suspended from a different surface platform, forming a large-aperture array with separations of about 1 km. Although this configuration tends to produce small localization errors, it poses significant challenges when trying to associate arrivals across the different hydrophones. The association problem is addressed using a pattern cross-correlation approach that has been developed in the framework of this project. This approach requires a very accurate time synchronization among all hydrophones, which is obtained by exploiting the PPS (pulse per second) signals – periodic box-shaped pulses of duration 0.1 s and period 1 s – of the GPS devices installed on the surface platforms.
Localization Analysis
Localization is the next step following peak association. It is based on novel Bayesian localization methods accounting for sound refraction effects (bending of acoustic rays and associated effects on travel times) and allowing for location uncertainty estimation. For the association of signals at the different hydrophones pattern recognition methods are used in combination with propagation modelling. The use of 3 hydrophones allows for a variety of localization approaches, from simple localization (range and bearing estimation) based on direct arrivals to 3D localization (range, depth and bearing estimation) exploiting direct and surface-reflected arrivals at the 3 hydrophones, whereas it also allows for useful fallback options, e.g. in case of suspended operation of one of the stations.
For further details on the algorithms publications section