Autonomous visual tracking and landing of a quadrotor on a moving platform

Abstract

This paper deals with the problem of autonomous visual tracking and landing of a quadrotor on a moving platform. The application of this problem can be seen in autonomous landing of UAV on a ship. The solution consists of two parts, a sensor framework to estimate the state of quadrotor using vision based approach and a controller design which generates appropriate actuator commands. A computer vision approach is proposed which tracks the pre-specifled oriented roundel object continuously while maintaining a fixed distance from the roundel and also simultaneously keeping it approximately in the center of the image plane. Quadrotor's relative pose is estimated through dead reckoning and control approach is implemented which seeks for full autonomy of the robot, by considering only internal sensors and processing unit and also, experimental results are presented.