Improved BP network based sliding model tracking control for a quadrotor UAV

Abstract

This paper introduced a novel method for trajectory tracking control of a quadrotor UAV under the parametric uncertainties. In order to improve its stability during trajectory tracking, the proposed method combines the advantages of back propagation (BP) network and sliding model control (SMC). Firstly, the expected roll angle and the expected pitch angle are decoupled by the nonlinear dynamic model of quadrotor and the known desired state. Secondly, a sliding model four-channel control law is designed for the nonlinear dynamics model of quadrotor, and the stability of the designed control law is proved by the Lyapunov theorem. Finally, the designed sliding manifold coefficients are solved by the improved BP network. In the Matlab/Simulink environment, compared with other algorithms, the proposed method can accurately track the desired trajectory, effectively compensate the parameters uncertainty, and greatly suppress the chattering phenomenon with the maximum overshoot of the SMC.