The 12th CAA Symposium on Fault Detection, Supervision, and Safety for Technical Processes (CAA SAFEPROCESS 2021)

1 / 2021-06-04 20:32:42

Robust fuzzy RBF neural network PID controller for quadrotor based on extended state observer under external perturbation

Fuzzy PID, Radical basis function neural network, Extended state observer, 6-DOF unmanned quadrotor

The problem of controlling the orientation and position

of a quadrotor is considered in the presence of parametric

uncertainties and external disturbances. In this article, a complete

mathematical model of a quadrotor UAV with four inputs and

six degrees of freedom (DOF) is modeled based on the Euler-

Newton formulation. A robust nonlinear fast control structure

for the quadrotor position and attitude trajectory tracking is

designed. The position loop generates the thrust to control the

altitude of the quadrotor and provides the desired pitch and

roll angles to the attitude loop, which allows the control of the

quadrotor center of gravity in the horizontal plane. A fuzzy

radial basis function (RBF) neural network PID control scheme

combined with the extended state observer has been performed

for the 6-DOF unmanned quadrotor. The upper bounds of the

attitude uncertainties and disturbances have been eliminated

through the extended state observer. To estimate the proposed

controller parameters of the system, fuzzy inference rules and an

RBF approximate method are proposed. Furthermore, Lyapunov

analysis is used to guarantee the stability of the quadrotor

UAV system and to ensure the robustness of the controllers

against variation. Finally, different simulations are performed

in the MATLAB environment to show the efficiency of the

constructed controller. The sovereignty of the proposed controller

is highlighted by verifying its performance with simulations.