File:VISION-BASED RELATIVE POSITION ESTIMATION AND INTERCEPT TRAJECTORY PLANNING FOR SMALL UNMANNED AIRCRAFT SYSTEMS (IA visionbasedrelat1094563509).pdf

The proliferation of unmanned aircraft systems (UAS) has contributed to the asymmetric threat of malevolent actors exploiting this technology for mischief or harm. Existing ground-based solutions are limited by line of sight, while human-operated responder drones can be less responsive and more labor-intensive. Hence, there is a capability requirement for autonomous vision-based pursuit and interception of unauthorized drones. To address this, the author developed a computer vision (CV) algorithm to detect, track and estimate the relative position and range of a hovering and moving airborne small UAS target in field conditions. CV-based measurements were compared against GPS data, to assess the range and angular estimation performance of the CV algorithm. Then, the CV-estimated range and angular information was processed by a flight control algorithm utilizing simple angular guidance principle to pursue and intercept the target. Field tests of the algorithm were done using a prototype drone. This research will inform the conceptual design and choice of hardware implementation for a commercial-off-the-shelf-based counter-UAS capability. More broadly, the research contributes to the body of knowledge in autonomous object tracking applications.

Subjects: computer vision; object detection; object tracking; position estimation; navigation; trajectory; UAS; drone