Adaptive attitude control of UAV based on neural network compensation

Abstract Aiming at the problem of unmanned aerial vehicle (UAV) attitude control with insufficient robustness and anti-disturbance, this paper proposes a UAV method combining incremental nonlinear dynamic inverse (INDI) neural network (NN) adaptive compensation. Firstly, model time scale separation is proposed. On basis, INDI law derived, angular acceleration information required by feedback filter. Then, NN compensation structure introduced to eliminate uncertainty caused mismatch, sensor delay external disturbance. Finally, simulation carried out. The result shows that improves anti-disturbance original without changing parameters controller, has certain ability for system uncertainty.