Robust Co-design Optimisation for Agile Fixed-Wing UAVs

A research team including Adrian Andrei Buda, Xavier Chen, Nicolò Botteghi, and Urban Fasel has published a new paper titled "Robust Co-design Optimisation for Agile Fixed-Wing UAVs" on arXiv. The work addresses a critical gap in autonomous system design: existing co-design frameworks often fail to account for the robustness needed in real-world, unstructured environments. For agile fixed-wing UAVs operating at their performance limits, this oversight results in designs overly sensitive to perturbations and model inaccuracies.

The researchers propose a novel bi-level framework that integrates parametric uncertainty and wind disturbances directly into the concurrent optimization of both a drone's physical design and its control strategy. The high-level loop optimizes physical parameters, while a lower level discovers nominal trajectories via a constrained planner and evaluates performance across a stochastic Monte Carlo ensemble using feedback LQR control. This method ensures the final design is inherently resilient.

Validated across three distinct agile flight missions, their robust co-design strategy consistently outperformed traditional deterministic baselines. The results demonstrate that the framework automatically tailors key aerodynamic features—such as wing placement and aspect ratio—to strike an optimal balance between raw mission performance and the ability to reject disturbances like unexpected wind gusts.