The increase in aircraft structural flexibility, resulting from the use of slender wings and
composite light-weight materials, to improve aerodynamic efficiency, poses a severe control
problem. The dynamics of these vehicles, described by nonlinear mathematical models,
are complex, with large wing deformations that significantly affect the vehicle flight mechanics
and make them vulnerable to disturbances or adverse atmospheric conditions.
Therefore, advanced control methodologies capable of accounting for the nontrivial dynamics
of such aircraft are required to improve their manoeuvrability and ensure their
structural integrity.
We propose a control framework that employs a model predictive controller and its dual
moving horizon estimator. These strategies are highly suitable to address this challenge,
since they rely on solving optimal control problems which are constrained by arbitrary
nonlinear dynamics and can explicitly account for state and actuator constraints. With
available proofs on stability and algorithmic improvements to enable real-time application,
the focus of this work is placed on deriving efficient characterisations that encapsulate
the relevant dynamics of these complex systems for successful control.
The conceived internal models for control build upon a recently developed geometricallyexact
beam description, known as the intrinsic formulation. It uses velocities and strains
as main variables, which leads to a compact mathematical model, amenable to efficient
nonlinear model-order reduction using the structural modes of vibration. These are also
used to project the typically high-dimensional inputs and outputs of a linearised unsteady
vortex lattice formulation onto lower dimensional subsets. This enables efficient model order
reduction of its state-space representation and establishes a common fluid-structure
interface facilitating the aeroelastic coupling. Finally, data-driven modelling is explored
to improve known limitations of the coupled aeroelastic models and further optimise their
size for control.
The resulting framework, able to operate in the nonlinear regime, allows for exploration of
unconventional control and stabilisation mechanisms as shown in the numerical examples, which would be unattainable with traditional control methods.