Developing of a Video-Based Model for UAV Autonomous Navigation

Abstract

The use of unmanned aerial vehicles (UAVs) has driven the research and development of multiple applications. Autonomous and cognitive navigation in remote environments requires the use of independent on board sensors. One advantage of these vehicles is that they have an on-board camera that allows them to collect visual information about the environment. This work shows a way to be aware of the UAV movement depending only on images. Therefore, a vision-based mathematical model was defined that describes the movement. System identification experiments and results are presented to verify the mathematical model structure and to identify model parameters comparing with state of art models. Finally a visual-based model compare with other methods and improve performance.