Q-Learning Algorithm for F16 Aircraft: An approach to Discrete-Time Systems

This brief presents the application of On policy Q-learning algorithm for perturbed F-16 aircraft based on the discrete-time model. In this modified Q-learning approach, the two variables Q-function is developed from Bellman function and the disturbance is integrated into optimal control problem with zero-sum game consideration. Moreover, the stability effectiveness and $H_{\infty}$ attenuation performance are satisfied by theoretical analysis. Thanks to the computation of Q-function without solving again the optimization problem, the presented solution is able to deal with the complete model uncertainty. Finally, the simulation results show that the calculation is correct and feasible of the proposed algorithms.