Vision-Based Autonomous Aerial Refueling

Abstract

Aerial refueling tasks are very challenging due to the high risk of aircraft close proximity. Currently, within the drogue-probe method, the receiver aircraft pilot manages the refueling task in accordance with the tanker aircraft pilot. Therefore, autonomous aerial refueling is still an unaccomplished task for aircrafts. In this paper, a fully automated aerial refueling procedure based on digital visual inspection is proposed. A nonlinear dynamic model of receiver aircraft is derived to track the motion of drogue. In order to control the receiver aircraft affected by tanker aircraft vortex during approach, and ensure the receiver aircraft to automatically track and dock the tanker aircraft, an autopilot system that considers visual sensing of drogue motion is designed. The receiver aircraft is controlled by the autopilot system via translational motion of tanker aircraft projected by a camera mounted on the receiver aircraft. Thanks to this vision-based controllers, the need of tanker aircraft positioning is denied since camera projection has the capability of perception of three-dimensional direction of tanker aircraft. In order to test the autopilots include vision-based controllers and algorithms, the vision-based autonomous aerial refueling is operated under presence of turbulence and vortex. Finally, the simulation results demonstrate that the proposed guidance-navigation-control system achieve aerial refueling autonomously, and make it feasible and realizable for aircrafts.