A Resilient Cooperative Localization Strategy for Autonomous Underwater Vehicles in Swarms

Mainstream terrestrial approaches to location awareness are curtailed underwater as radio frequency (RF) signals attenuate rapidly and accordingly invalidating reliant technologies for near bottom use. Sans RF, one predominant means for Autonomous Underwater Vehicles (AUVs) localization requires extrapolation from a known point. Arising from the dynamic nature of the subsurface milieu, with changing currents and obstacles, errors in these predictive calculations worsen over time; a phenomenon referred to as fading. In the event that multiple AUVs are operating in a swarm, fading is compounded and location-sensitive data collection are rendered impractical. In this paper, we contribute an approach to leverage swarm element interaction to reduce the effect of fading. An AUV ascertains a location fix when surfaced and subsequently may propagate their derived location when in proximity to peers underwater. The decision to propagate depends on the AUVs confidence in its reckoned location. Exploratory testing regarding the efficacy of the strategy was performed using a MASON based simulation platform. We found that while a naive streamlined approach had lower cumulative errors as swarm size increased, our strategy outperformed the control when error as malfunction was introduced within the test model. Our study of additionally performing purposed dispatch for scaffolding in the absence of global knowledge was unable to definitively disprove the null hypothesis for larger size swarms and behaved erratic when the search area was varied. We still include this strategy in our report as contribution to the community of negative results.