Adaptive Learning Control for a Quadrotor Unmanned Aerial Vehicle Landing on a Moving Ship

Abstract

The shipboard landing problem of a quadrotor without prior knowledge of the ship is investigated in this paper. The relative motion model of the quadrotor and the ship is established and transformed into an affine form for the landing controller design. To improve the adaptability to different ships, a neuron-adaptive neural network is adopted to estimate the model uncertainties caused by unknown ship parameters, and the disturbance observer is developed to measure the lumped disturbance, including the external disturbances and the approximation error. Moreover, a novel backstepping integral evolution sliding mode controller is developed for the relative position, and nonsingular fast terminal sliding mode control is adopted for others. Combined with the auxiliary systems, the input saturation and filter errors are considered in the closed-loop system. The theoretical analysis illustrates that the states of the relative motion system are guaranteed to be bounded. The simulation examples demonstrate the effectiveness and advantages of the proposed method.