Abstract
This thesis investigates the development and implementation of different parts of an autonomous sampling framework for marine robots to study the water column, with a focus on zooplankton detection, mapping, and sampling. By leveraging Gaussian Processes (GPs) and collaborative missions between Autonomous Underwater Vehicles (AUVs) and Unmanned Surface Vehicles (USVs), this research addresses key challenges in real-time detection, spatial mapping, and autonomous sampling of zooplankton.
The primary contributions of this research include the creation of a Sound Scattering Layer (SSL) detection algorithm for real-time zooplankton identification, the application of GPs to generate spatial maps of zooplankton distribution from sparse observation, and the development of semi-autonomous and collaborative sampling strategies. The efficacy of these methods was validated through field experiments conducted along the coast of Mausund, Norway.
During the field experiments, two missions with AutoNaut robot were conducted. The first mission spanned 12 hours, while the second extended over 25 hours, resulting in substantial data collection. Specifically, mission 1 lasted 12-hour with recording acoustic data while detecting SSL in real-time, and mission 2 collected four distinct echosounder datasets, processed in real-time. These experiments demonstrated the consistent reliability of the SSL detection algorithm across varying conditions and depths, with accurate identification of upper and lower SSL boundaries.
Collaborative missions framework between AUV and USV for autonomous sampling of zooplankton was tested in Mausund. The experiment demonstrates the ability of the framework to carry more efficient sampling strategy. The AUV was able to receive orders from the USV and act upon these orders autonomously.
In conclusion, this work establishes a foundational approach for employing advanced robotic systems in marine ecology, providing valuable insights and tools for future research and operational deployments, ultimately contributing to more informed and effective marine conservation efforts.