Autonomous Wingtip Docking of Fixed-Wing UAVs: Relative Navigation, Closed-Loop Control, and System-Level Challenges

Autonomous wingtip docking of fixed-wing unmanned aerial vehicles offers a potential approach to long-endurance flight, modular mission reconfiguration, and airborne assembly. Unlike conventional formation flight, this task requires two forward-flying aircraft to reduce the relative position, attitude, and closing-velocity errors between their actual docking points while satisfying fixed-wing flight-envelope, visibility, aerodynamic-interaction, and collision-avoidance constraints. This article organizes relevant research around two closely coupled themes: relative navigation and closed-loop docking control. Vision-based relative measurement methods are examined, including cooperative fiducial markers, structure-based non-cooperative perception, monocular and stereo pose estimation, and learning–geometry hybrid approaches; multi-sensor relative-state estimation is then discussed with emphasis on visual–inertial fusion, GNSS/RTK and UWB assistance, dual-moving-platform modeling, asynchronous measurements, lever-arm compensation, observability, and uncertainty representation. The control literature is further analyzed from the perspectives of position-based and image-based visual servoing, robust and prescribed-performance control, model predictive and safety-critical control, and learning-based disturbance compensation. The central technical challenge lies not in improving pose-estimation accuracy or tracking convergence in isolation, but in constructing an uncertainty-aware, constraint-compliant, and contact-transition-capable autonomous closed-loop system. Future research should therefore focus on dual-platform relative observability, confidence-aware closure control, visibility-preserving trajectory planning, real-time near-field aerodynamic modeling, hybrid contact dynamics, navigation–control–mechanism co-design, and progressively structured experimental validation.