Visual Servoing of Micro Aerial Vehicles with the Cooperation of Ground Vehicle

A ground-to-air cooperation system is proposed in this article aiming to enhance the performance of Micro Aerial Vehicles(MAVs) in GPS-denied environment. The MAV cooperates with the ground vehicle to complete flying task. Specifically, the ground vehicle uses a monocular camera to track the marker points on the MAV with background subtraction and optical flow algorithms to avoid interference from unrelated backgrounds, and calculates the position relative to the ground vehicle. After receiving the position of the ground vehicle, the MAV fuses the observed position with its own accelerometer information, and performs self-state estimation and feedback control. Compared with the past methods, the proposed method can effectively remove the interference of complex background, perform point-tracking on the MAV markers to improve the robustness of positioning, and reduce the state estimation delay by fusing multi-sensor information. Experiments are carried out to verify the flight path of the MAV observed by the ground vehicle with the trajectory from the Vicon motion capture system, and thus prove the accuracy and effectiveness of the system.