Robust Direction-Preserving Control Allocation

This paper investigates robust linear control allocation for aircraft with parametric model uncertainty. We augment a standard linear programming-based allocator with a robustified direction preserving constraint, which ensures that the realized generalized moment is approximately collinear with the command over the specified uncertainty set. This results in a second-order cone program which can be solved reliably in real time. Using an example with the F-18 high-alpha research vehicle, the proposed allocator is compared to similar methods and demonstrated to best preserve realized moment directions under uncertainty while being able to realize nearly all physically-attainable moments.