Target state estimation for vision-based landing on a moving ground target

Abstract

This paper describes vision-based target state estimation approaches for autonomous landing on a moving ground target. The state of moving ground target is estimated by using vision information from a gimbaled camera on an Unmanned Aerial Vehicle (UAV). Using the information from vision system, the UAV estimates the state of a moving target on the ground using the Unscented Kalman Filter (UKF). In this paper, three types of process model are compared by numerical simulations: state in inertial frame, state in inertial frame with position uncertainty of UAV, and relative state with acceleration of UAV.