06.1 - Flight Dynamics and Control (Control & Modelling)
LEARNING-BASED CONTROL ALLOCATION FOR OVER-ACTUATED VTOL AIRCRAFT WITH NONLINEAR DYNAMICS
D. Kim¹, J.D. Lee¹, H. Lee¹, Y. Kim¹, H. Bang¹; ¹KAIST, South Korea
This study presents a learning-based control allocation method for over-actuated VTOL aircraft while explicitly accounting for nonlinear dynamics. A deep auto-encoder is employed to learn the mapping between actuator commands and the resulting control effects, enabling effective utilization of available control authority under actuator constraints. Simulation results demonstrate improved allocation performance compared to conventional methods.
DOI: 10.71945/icas2026_0567