Neural Adaptive Coordinated Docking Control With Improved Prescribed Performance for UAV Aerial Recovery

Existing master–slave docking strategies only rely on controlling the unmaned aerial vehicle (UAV) to accomplish aerial docking and suffer from low efficiency and accuracy. To this end, by synchronously controlling the UAV and towed drogue, this article proposes a neural adaptive coordinated docking control scheme with improved prescribed performance for UAV aerial recovery. First, affine nonlinear dynamics models of heterogeneous docking system, including the drogue and UAV, are established. Then, by constructing envelopes with an asymmetric structure and repeatable definition transient processes, a novel command switching-oriented appointed-time prescribed performance control (CSAPPC) method is proposed. CSAPPC can guarantee that the docking errors converge to a user-defined domain within an appointed time while preventing overshoot, and it can effectively handle the strike of velocity command switching during docking. To compensate for unknown disturbances, a neural network (NN)-based adaptive approximation approach with a low computational burden and decoupling from the control loop is developed. Based on the above works and backstepping theory, coordinated controllers for the drogue and UAV are designed. Finally, simulations and ground experiments are conducted to verify the effectiveness of our proposed framework.