This paper addresses the problem of finite-time trajectory tracking control for a class of uncertain quadrotor unmanned aerial vehicles (UAVs) with input saturation and actuator failures. A reinforcement learning (RL)–based controller is designed to eliminate the effects of model disturbances and structural uncertainties. A robust controller is also implemented to ensure stability during the early stages of RL training. To handle input saturation, a hyperbolic tangent function is adopted to smooth transitions, limit amplitude, and enhance response performance. Combined with the proposed finite-time controller, a new robust trajectory tracking control scheme is constructed, which ensures that the tracking error enters a small neighborhood around zero in finite time. The simulation results show that the convergence times of the proposed method along the x-, y-, and z-axes are 1.4, 1.96, and 1.59 seconds, respectively, and the tracking errors are − 0 . 0547 , − 0 . 0521 , and − 0 . 000176 , respectively. Even in the presence of severe disturbances, unknown actuator failures, and input saturation, the tracking errors can converge to a small neighborhood near the equilibrium point within a limited time.
Zhang et al. (Thu,) studied this question.