Interception of an aerial manoeuvring target using monocular vision

Abstract

In this work, a vision-based guidance algorithm for interception of an aerial manoeuvring target is proposed. Vision-based aerial engagement in 3D space in an outdoor environment is challenging due to uncertainty in estimating the target location. The vision-based guidance algorithm proposed here generates the desired velocity commands for the interceptor drone based on the orientation of the target pixel in the image plane. The desired velocity at the centre of gravity of the interceptor is calculated such that the velocity at the camera centre is aligned towards the line joining the target pixel and the camera centre. It is shown that the proposed guidance strategy forces the target pixel towards the centre of the image plane and leads to low miss distance. The algorithm developed is first tested in the ROS-Gazebo platform and then implemented in field experiments.