Numerical Investigation of Feedback Path Selection for Envelope Shaping in Fixed-Wing UAVs

Envelope feedback can improve a fixed-wing UAV state margin while increasing the command and input demand needed to realize it. We compare all eight combinations of speed force, angle-of-attack force and angle-of-attack moment at common gains in one inversion-based architecture. On 48 reused development cases, nested differences and independent directional Taylor propagation give the same outcome classifications. Angle-moment-only CAT and speed-force/angle-moment CVAT complete all cases with positive sampled margins, whereas angle-force configurations exhibit background-dependent input-request terminations. A frozen protocol for independent-input numerical evaluation then assesses four configurations on 100 new cases in two sampling ranges. CAT and CVAT complete all 50 cases in each range with positive sampled margins. CFULL completes 49/50 and 42/50. The stronger angle-margin target remains unmet. Common-state diagnostics trace force-path changes through command acceleration to elevator demand. A second-order attitude command tracker removes six terminations in a separate five-case comparison but increases position RMSE by 4.49–7.03 m on common complete pairs. Prescribed state-estimation errors substantially reduce completion. These results show task- and command-realization-dependent tradeoffs within the tested architecture. Conditional power accounting identifies retained coupling terms without establishing stability or invariance. The findings are limited by finite numerical resolution, qualified sampling, instantaneous actuation and the absence of atmospheric or hardware validation.