Practical finite-time prescribed performance adaptive NN backstepping control for quadrotor hovers based on disturbance observers

In this paper, a practical finite-time prescribed performance adaptive neural network (NN) backstepping control based on a disturbance observer is presented for quadrotor hover system. In the proposed method, a finite-time disturbance observer was constructed to estimate and compensate for external disturbances, and radial basis function NNs with gradient descent training are applied to deal with uncertainties. On the contrary, through the incorporation of a prescribed performance function and an error transformation function, the proposed control method ensures that the tracking error remains consistently confined within the boundaries of the prescribed performance function. Based on practical finite-time Lyapunov stability criterion, it can be proved that the system tracking error is practical finite-time stable. Finally, experimental results on a quadrotor hover are presented and verify the advantages and effectiveness of the proposed control method.