Multi-Objective Gain Design for Enhanced Nonlinear Closed Loop Performance and Robustness for eVTOL Applications

This paper presents a multi-objective gain design approach considering both nonlinear closed-loop dynamics, such as accurate axes coupling, and linear stability and robustness requirements. In particular, Kreisselmeier-Steinhauser vector performance indexing is used in conjunction with pattern-search optimization to summarize both requirements from the nonlinear model as well as stability and robustness requirements from the linear model at the same trim point to formulate the multi-objective target. Consistency between the linear and nonlinear closed-loop systems is facilitated by conditional re-linearization throughout the optimization. The performance as well as the stability and robustness objectives are considered within the cost function using specifically tailored sub-goal metrics. Although the suggested approach, requires increased computational cost for the repeated simulation of the nonlinear closed-loop system for the optimization, it has the potential for saving computational effort by reducing t e overall gain design-verification iterations needed, this is because the inherent inclusion of the nonlinear dynamics within the optimization ensures a more accurate representation of the implemented closed-loop behavior, as opposed to discovering deficiencies in verification afterwards. The proposed approach is applied to and tested on a piloted, electric vertical take-off and landing (eVTOL) vehicle with an approximate MTOW of three tons, currently under development. Superior results compared to a standard linear model based approach are presented and discussed. Finally, future improvements to the suggested method are indicated.