Robust docking control and safety evaluation of autonomous aerial refueling for unmanned aerial vehicles

This paper introduces a novel tracking control scheme based on additive state decomposition (ASD) to address the challenges of precise docking control in autonomous aerial refueling (AAR) for unmanned aerial vehicles (UAVs) under external disturbances, model uncertainties, and actuator faults. Firstly, using ASD theory, the complex control problem of aerial refueling docking is decomposed into two subproblems: a simple linear robust tracking problem with disturbances and a nonlinear system stabilization problem without disturbances. Then, a robust $H_{\infty }$ anti-disturbance fault-tolerant composite controller is designed for the primary system, while a feedback linearization controller is applied to the secondary system. Furthermore, calculating the relative docking distance between the probe and drogue under external disturbances involves an extremely complex process. To address this, we develop a predictive model using a deep learning data-driven approach, integrating the sparrow search algorithm (SSA) with a long short-term memory (LSTM) network. Utilizing the predictions from this model, we construct a safety assessment network (SAN) to evaluate the future safety of AAR docking operations. Finally, the robustness of the proposed control method and the accuracy of the network’s prediction results are validated through comparisons with various control methods and other network models.