Innovative 3D techniques for marine environmental monitoring through drone swarms

The monitoring of the human impact on the marine environment has emerged as a topic of growing interest in scientific research over recent years. Traditional methods, such as the use of oceanographic vessels, are increasingly becoming obsolete. Technological advancements have led to the introduction of solutions based on multi–vehicle autonomous marine drone systems, with the aim of reducing costs, accelerating execution times, while simultaneously expanding the monitoring area. This thesis proposes innovative techniques that naturally apply to this field, addressing two fundamental macro–themes. The first concerns trajectory planning based on the solution of 3D local interpolation schemes with a particular class of polynomial curves, known as Pythagorean–hodograph (PH) curves. Due to their intrinsic properties, PH curves are particularly advantageous for the precise and efficient computation of the estimated time of arrival, a key feature for synchronizing the motion of drone swarms. Another aspect of particular interest in this field is the definition of a suitable reference frame that incorporates the unit tangent of the prescribed curvilinear path. This kind of frames, called adapted frames, describes the motion a vehicle should perform to follow a trajectory while maintaining its roll axis aligned with the curve tangent. In particular, the thesis focuses on Rotation–Minimizing Frames (RMFs), a class of adapted reference frames that do not exhibit instantaneous rotations about the unit tangent vector. RMFs provide a natural motion description for under–actuated autonomous vehicles that cannot perform roll rotations. The second macro–theme concerns the development of a path–following guidance law designed to enable the tracking of these trajectories while accounting for the properties of the marine environment, such as the presence of currents, and the characteristics of underwater vehicles. Finally, the thesis work presents the results obtained by applying some of the developed algorithms to the motion planning of a multi–vehicle autonomous marine system intended for monitoring environmental pollution.