Improving the accuracy of landmark tracking by the optical system of an unmanned aerial vehicle

The study examines the problem of autonomous navigation for a group of robotic transport vehicles in unstructured environments. The primary focus is placed on a method to use a computer vision system employing fiducial markers to determine the distance and angular position of reference objects, and to experimentally determine its range constraints. Conducted experiments demonstrate, that with the correct selection of camera parameters and identification algorithms, high measurement accuracy can be achieved, even under conditions of partial obstruction of reference objects and adverse external factors, at distances of more than 100 m. The presented results confirm the effectiveness of fiducial markers in autonomous control systems, making them a promising tool for application in the autonomous navigation of robotic complexes.