This study deals with the robust optimization of the transition maneuver for a vertical take-off and landing drone (VTOL) using a bi-level optimal control (OC) approach: The upper level, parameter optimization, problem is optimized using a differential evolution (DE) genetic algorithm. By this a global optimal solution is achieved. On the other hand, the lower level problem is solved by using a gradient-based direct OC scheme. The distinctiveness of this lower level setup is the fact that not only the general mean optimization problem is solved, but an OC problem (OCP) with uncertainties. This combination yields the possibility to calculate robust trajectories. The uncertainty modeling is done by means of generalized polynomial chaos (gPC). Thus, the lower level problems are not only solved at one specific set of parameters, but on multiple parameter sets to calculate the uncertainty influence on the optimal trajectory in distinct parameters. The developed algorithm is applied to the transition maneuver, including the climb to a safe altitude, of a VTOL. The results show an enhancement of the optimal trajectory in the sense of robustness with respect to wind influences regarding the safety of the transition maneuver.     «

This study deals with the robust optimization of the transition maneuver for a vertical take-off and landing drone (VTOL) using a bi-level optimal control (OC) approach: The upper level, parameter optimization, problem is optimized using a differential evolution (DE) genetic algorithm. By this a global optimal solution is achieved. On the other hand, the lower level problem is solved by using a gradient-based direct OC scheme. The distinctiveness of this lower level...     »