Differential Image Technique for Proximity Navigation

In this paper we introduce the Differential Image Navigation technique to perform au-tonomous proximity navigation and docking, or landing, with an object whose shape and dynamics are initially known with a big uncertainty. In the proposed configura-tion the spacecraft only carries a fixed, single digital camera, and the control system, during the critical phase of landing/docking, updates the control commands based on some global parameters characterizing the image flow obtained from the differ-ence of two subsequent images. This results into a novel and passive vision-based framework to perform autonomous proximity navigation. In particular, we will show how to estimate the rotational dynamics and the global parameters used to provide feedback to the control system