In the last sixty years of space exploration, only a handful of missions targeted Phobos and Deimos, the two moons of Mars. All of these missions failed, preventing them from successfully achieving their objectives. The planned Mars Moon Exploration mission is the first to be specifically designed to gather samples from Phobos and return them to Earth.

At the state of the art, this kind of mission relies on on-ground teams to monitor the spacecraft’s state, gather knowledge of the target body, make decisions and send instructions to perform the manoeuvres needed to approach the target body. In this control loop, navigation is one of the processes that may be automated by exploiting the latest advancements in computer vision, favouring autonomy and reducing the dependence on ground teams. This work aimed at exploring possible solutions to achieve that goal.

With the objective of autonomously creating a map of points of interest on the surface of small bodies, a set of novel feature extraction algorithms based on deep learning techniques were integrated into simultaneous localization and mapping algorithms and tested on real and synthetic optical data. The results showed a significant robustness of the proposed deep-features, in tracking the observer motion and mapping the environment, compared to standard feature extraction algorithms.

In the context of the MMX mission, it was also investigated the best orbital solutions to maximize the scientific return of the mission. It was shown that flying quasi-periodic orbits around Phobos, instead of the currently planned periodic ones, and integrating optical data from the spacecraft’s camera, can significantly improve the observability of the moon’s gravitational field and distribution of the moments of inertia.

To conclude, the feasibility of a dedicated autonomous system to navigate the spacecraft around Phobos, using only spaceborne sensors and computer vision algorithms was demonstrated, highlighting the possible navigation setups and combination of sensors that ensure a safe and reliable navigation of MMX around Phobos.