Robust H Tracking Control for a Quadrotor UAV with Unknown Dynamics Via Integral Reinforcement Learning

This paper addresses the robust H infinity tracking control for a quadrotor unmanned aerial vehicle (UAV) subjected to external disturbance with unknown dynamics. It is worth emphasising that solving the optimal H infinity tracking control problem requires addressing a nonlinear Hamilton-Jacobi-Isaacs (HJI) equation, which is well known to be difficult to solve in practical applications. Furthermore, in this work, the system dynamics are assumed to be unknown; therefore, obtaining an optimal solution to the HJI equation using mathematical analysis methods is infeasible. To overcome this challenge, a learning algorithm based on Integral Reinforcement Learning (IRL) is developed for both translational and rotational controllers to achieve optimal control policies by using only the measured input-output quadrotor data. Numerical simulation results are presented to demonstrate the superior tracking performance and disturbance robustness of the proposed method.