SLAM-based Underwater Adaptive Sampling Using Autonomous Vehicles

Abstract

In order to achieve efficient and accurate sensing Abstract—In order to achieve efficient and accurate sensing coverage of water reservoirs and 3D ocean bodies in near real time, in this paper a novel adaptive sampling strategy using Autonomous Underwater Vehicles (AUVs) is introduced. The vehicles capture the spatial distribution of the specific manifestations—such as salinity, temperature, potential Hydrogen (pH), chlorophyll concentration—of the phenomenon in the field of interest with the help of Simultaneous Localization and Mapping (SLAM) algorithms for navigation. To enable adaptive sampling with the required accuracy, the vehicles, i.e., mobile nodes, need to adjust continuously their trajectories with the help of an external platform—the static nodes on the surface— based on the sampling information gathered by the on-board sensors and also on the localization information provided by the Speeded-Up Robust Feature (SURF) algorithm. Experiments were conducted on a mobile robot and a static surface node to verify the proposed solution. In the original scheme, the robot was connected to the user via a tether; instead, we use an onboard controller to perform adaptive sampling autonomously underwater, and on the water surface via wireless connection to the static node as the remote processor. coverage of water reservoirs and 3D ocean bodies in near real time, in this paper a novel adaptive sampling strategy using Autonomous Underwater Vehicles (AUVs) is introduced. The vehicles capture the spatial distribution of the specific manifestations—such as salinity, temperature, potential Hydrogen (pH), chlorophyll concentration—of the phenomenon in the field of interest with the help of Simultaneous Localization and Mapping (SLAM) algorithms for navigation. To enable adaptive sampling with the required accuracy, the vehicles, i.e., mobile nodes, need to adjust continuously their trajectories with the help of an external platform—the static nodes on the surface— based on the sampling information gathered by the on-board sensors and also on the localization information provided by the Speeded-Up Robust Feature (SURF) algorithm. Experiments were conducted on a mobile robot and a static surface node to verify the proposed solution. In the original scheme, the robot was connected to the user via a tether; instead, we use an onboard controller to perform adaptive sampling autonomously underwater, and on the water surface via wireless connection to the static node as the remote processor.