Guidance and control of AUVs performing hydrobatic maneuvers

The recent field of Hydrobatics presents a set of formidable engineering challenges for autonomous underwater vehicles (UAVs). Previous work has shown quite small, high thrust AUVs, capable of performing various maneuvers with optimal control solutions. However, these AUVs are still unfit for tight spaces and narrow maneuvers. A metric for the AUV's maneuverability is its thrust-to-weight ratio (thrust divided by gravitational force), of which most AUVs show values less than 1.
This Thesis will show, as a proof of concept, that improved hydrobatic maneuvers and even multi-domain trajectories (here: jumps out of water) are possible. For this, a high thrust-to-weight (> 4) AUV "Arrow" will be used as hardware, together with optimal control methods. These methods consist of trajectory generation via direct transcription and time-varying linear quadratic regulators (TVLQR) for stabilization. The System will be validated, first in simulation and afterwards with real-world tests. Trajectories include a change of orientation (inverted pendulum), circular motion ("loopings"), and jumps out of water with planned ballistic trajectories.