Control of Quadrotor Based on RBF Neural Network Adaptive Fast Terminal Sliding-Mode Strategy

Abstract

I The purpose of this paper is to propose a fast adaptive terminal sliding mode control method for quadrotor unmanned aerial vehicles (UAVs) under dynamic uncertainties and external interference. Firstly, a complete mathematical dynamics model of a quadrotor UAV is presented. Secondly, a robust dual-loop nonlinear control law for the quadrotor velocity and attitude tracking is designed. Through the introduction of equivalent control inputs, the attitude control channels can be decoupled and attitude stabilized controllers can be designed separately. For velocity tracking and altitude tracking control, coupled controllers are designed to ensure velocity tracking based on the stability of altitude tracking control. The attitude loop adopts RBF neural network sliding mode control, to compensate for model disturbance uncertainty, adjust its gain in real time, and suppress chattering problems. Finally, the experimental results show that the designed controller can not only suppress external interference, but also achieve accurate tracking of the desired flight command.