Quadrotor Control using Neuro-Adaptive Robust Generalized Dynamic Inverasion

Abstract

A two-loops structured control system is proposed for Quadrotor position and attitude tracking. The outer (position) loop of the control system implements small disturbance linearization of the roll and pitch attitude variables together with a Proportional-Derivative constrained control type of the position errors in the three dimensional space, and it provides attitude commands to the inner (attitude) loop for tracking the Quadrotor’ s center of gravity position coordinates in the instantaneous horizontal planes, in addition to the required thrust force. The Neuro-Adaptive Robust Generalized Dynamic Inversion (NA-RGDI) control is employed in the inner loop for attitude tracking and body rates stabilization. The NA-RGDI control law comprised of a baseline controller integrated with a sliding mode control term to provide robustness against unmodeled system dynamics and uncertainties. Additionally, a Radial Basis Function Neural Network is engaged to estimate the unknown attitude dynamics of the quadrotor. The weighting matrices of the neural networks are updated online using Lyapunov principle. Numerical simulations of the closed loop system reveal good positional and attitude tracking performance in the presence of uncertainties in the system dynamical parameters and in the presence of external wind disturbances.