Vision-Based Tracking of a Ground-Moving Target with UAV

This paper presents a vision-based estimation and guidance method to enable a UAV to fly a circular orbit around a moving target. The target motion is estimated using a Kalman filter design combining an onboard GPS and inertial sensors with a target-pointing vector obtained from a vision system. The aircraft is guided to orbit the moving target by a simple guidance law that deploys the relative side-bearing angle and the relative velocity with respect to the target. The proposed method is demonstrated using simulations and flight tests.