关键词: 倾转四旋翼无人机, ADRC, RBF神经网络, 姿态控制, 参数自适应

Abstract: An enhanced active disturbance rejection control (ADRC) parameter adaptive control strategy based on radial basis function (RBF) neural network is proposed for the control stability and accuracy of the autonomous flight of tilt-quadrotor UAVs under complex disturbances. Firstly, based on the component mechanism modelling method, a nonlinear flight dynamics model of the tilt-quadrotor UAV covering the whole flight mode is established. Secondly, the RBF neural network with strong nonlinear function approximation ability is used to solve the problem of online adaptive tuning of nonlinear active disturbance rejection controller parameters. The con-trol input solved by the controller in real time and the state output of the tilt-quadrotor UAV are used as the input of the neural net-work. Based on the output of the neural network, the parameter adaptive adjustment rules are constructed, and the parameters of the extended state observer (ESO) part and the nonlinear state error feedback control (NLSEF) part of the active disturbance rejection controller are dynamically adjusted online to realise the effective estimation and compensation of model uncertainty and external dis-turbance. Finally, the attitude adaptive control system of the tilt-quadrotor UAV is constructed based on the ADRC-RBF controller, and the attitude control simulation is carried out in the typical flight mode. The simulation results show that compared with the tradi-tional ADRC controller, the ADRC-RBF controller designed in this paper has better anti-interference, adaptive ability and stability.

Key words: tilt-quadrotor UAV, ADRC, RBF neural network, attitude control, adaptive parameter adjustment