Vision-based Autonomous Tracking Control of Unmanned Aerial Vehicle

During the past decades, the field of aerial robotics has become the hot-spot for researchers, especially control of autonomous robots navigating in indoor environments. A lot of work has been published but still needs further efforts. In this paper, the problem of marker-based tracking control of a ground moving target and its solution is discussed. The main motivating application of this work is tracking a ground moving target while observing the environment at the same time. So the whole work mainly consists of two parts: Object recognition and tracking control. In the object recognition part, recognizing the marker and calculating the position of the target using the identified marker is discussed. While tracking control part, mainly focused on applying a control algorithm to track the moving target. A drone with a full HD camera is utilized for marker tracking and using it as a bird's eye view. With a larger field of view, surrounding can be observed. No external tracking system is used. In the end, experiments (including indoor and outdoor environments) and results are discussed, which show the effectiveness of the proposed approach. For tracking of the target, markers of different sizes and numbers (including single and multiple markers) are used. The performance of multiple marker-based tracking control in terms of states versus time plots is compared with the single marker tracking approach to elaborate our presented scenario.