Optimal Design and Nonlinear Control of Autonomous Underwater Vehicle

This project as a bilateral research collaboration Germany-Croatia is a joint effort of research teams from Technical University of Munich (TUM) and University of Dubrovnik (UNIDU). Within this project, we aim to devise an optimal design for Autonomous Underwater Vehicle (AUV) in terms of hydrodynamics and actuator allocation given mission requirements. We will focus on establishing a generalized model for an underwater robot with customizable configurations. Afterwards, the corresponding AUV prototype will be developed at TUM, while a nonlinear mathematical model of the corresponding AUV dynamics will be derived from the generic model with optimal parameters and analyzed at UNIDU. Our intention is to employ nonlinear control methods, analyze robustness of the obtained control algorithms and test them both numerically and experimentally. The resulting control performance will be compared with the existing control strategies. Both research teams have already considered and investigated similar problems in terms of control and dynamics, but approached them from different viewpoints (aerial and underwater). 


At the moment, the design of an Autonomous Underwater Vehicle (AUV) is typically an iterative design by experts on specific requirements posed on the mission the AUV should perform. AUVs are highly integrated systems with tight coupling in many of the design parameters and a significant influence of these parameters on the dynamics (and thus on the control of the system), which in turn makes their design challenging. In research applications, owing to varying mission requirements, corresponding costs and the absence of standard platforms, a solution that fits a specific mission is a limiting factor in utilizing the designed UAV in other use cases of interest.

Project goals

In this project, we will first devise tools for optimal vehicle design in terms of hydrodynamics and actuator allocation for AUVs given mission requirements (e.g., mission duration, goals, payloads, dynamic constraints and requirements, terrain configuration, obstacles). We plan to use trimming trajectories (a combination of helical trajectory segments) to specify the mission requirements in terms of duration and the necessary vehicle dynamics to track the trajectory. We will focus on establishing a generalized model for underwater robots with customizable configurations. In other words, the amount, position and orientation of fins and thrusters are free to be allocated along a parameterizable hull configuration. The internal configuration of the AUV is subject to matching the sensing, energy and volumetric (i.e., buoyancy, space) requirements of the mission. The parameterized generic model can then be used to find an optimal solution to the envisioned use-case through multi-objective optimization.