Neural network adaptive inverse model control method for quadrotor UAV

Abstract

For the highly nonlinear, strong coupling and underactuated, an effective controller for quadrotor unmanned aerial vehicle (UAV) is not easy to obtain, especially in conditions of the parameters variations and external disturbance. In this paper, an adaptive inverse model control method (AIMCM) is proposed for this plant. Two BP networks are used in control system to achieve the model identification and control. Both offline trained and online learned are used to ensure the fast learning and the robustness. The convergence of the learning algorithm is proved based on Lyapunov function. At last, a quadrotor UAV control simulation based on the full model shows the superiority and robustness of the control system.