Navigation of AUVs based on Ocean Fields Variability

A key to enable long-endurance and large area missions for Autonomous Underwater Vehicles is the development of novel navigation methods. Typical solutions based on periodic surfacing or the deployment of static beacons suffer from a number of limitations ranging from interrupting the vehicle task to constrain the operational area. To overcome some of these limitations, this paper moves the first steps into a different direction and aims at using marine environmental information (e.g. temperature, salinity, etc.) as navigational aid for the robots. Towards this aim the paper presents a Particle Filter able to use temperature and salinity maps produced by state-of-the-art ocean models, and assesses the navigation performance over a week long simulated mission. The obtained numerical results show that the proposed approach is able to substantially bound the navigation error, and hence to support the navigation of underwater robots for long-range missions. Discussion on advantages, limitations and promising ways forward are also presented.