RflySimAAR: An Integrated High-Fidelity Simulation Platform for Autonomous Aerial Refueling

Aerial refueling enhances aircraft endurance but involves complex challenges across disciplines, including aerodynamic modeling, hose dynamics, wind disturbances, and airflow effects. Despite advancements, a unified platform for studying and validating the entire refueling process remains lacking. Therefore, this paper introduces an integrated high-fidelity simulation platform for autonomous aerial refueling. This platform finely models various aspects of aerial refueling, including tanker and receiver models, flexible hose drogue models and wind disturbance models (including bow wave effect). Additionally, it has developed a highly-fidelity 3D simulation model using Unreal Engine 4 for simulating scenes, aircraft and refueling equipment, providing various sensors for developing vision and LIDAR-based perception algorithms for fully autonomous unmanned aerial docking. Furthermore, the platform supports versatile interface control, allowing for both fully autonomous unmanned docking and manual control using a joystick. Two demos implemented on the platform present the different methods, respectively. Furthermore, this platform employs a modular approach, enabling the development of individual models and facilitating joint simulations. In short, this platform aims to provide a comprehensive and reliable infrastructure for aerial refueling, thereby greatly facilitating research in the field.