Design of a CubeSat mission for the detection of small-dimension space debris

Space debris, also known as space junk, constitutes a major challenge for the entire space community. These artificial objects in orbit around the Earth are mainly made up of not-removed spacecraft at the end of their service life, upper stages of lunch vehicles and fragments produced after explosions or collisions. Being out-of-control space waste ready to crash into other orbiting objects, space debris represents a serious threat for all the space activities, present but especially future. Hence, several researches have been carried out to tackle the problem, identifying various mitigation strategies. Some of these involve an active management of end-of-life satellites, parking them into graveyard orbits or deorbiting them until complete disintegration in the Earth’s atmosphere. Other researches suggest the removal of potentially dangerous debris through the use of capture mechanisms, requiring dedicated space mission. By contrast, other lines of research address the problem in a passive way, accepting space debris and developing strategies to mitigate the fragmentation risk. For example, satellite shielding and end-of-life passivation are techniques that aim to reduce the severity of a collision and the probability of explosions, respectively. On the other hand, another solution aims at reducing the probability of collision through collision avoidance manoeuvres. Such a solution requires a database of catalogued and continuously tracked debris, enabling possible in-orbit conjunctions to be calculated in advance. But the currently active ground-based detection and tracking networks can hardly observe debris smaller than 10 cm. This PhD project fits into the latter research context, aimed at expanding the knowledge on 1-10 cm space debris. By means of a dedicated CubeSat space mission, this class of debris would be populated for the first time with observational data collected directly from space. This would allow for improved models of the orbital environment, enabling not only the release of tailored mitigation guidelines, but especially more accurate estimations of in-orbit collision risk, useful for both space operators and insurance companies. Therefore, the research activity ramified into two main directions, both crucial for achieving the project’s objectives: on the one hand the preliminary design of the space mission, and on the other the development and testing of debris detection algorithms. The first research direction led, through the definition of Mission Objectives and Requirements, to the definition of the Concept of Operations, analysing the orbit and estimating the number of detectable space debris during the mission. In addition, two possible types of detectors were analysed and compared, one based on the JEM-EUSO technology and analogous to the Mini-EUSO telescope on board the ISS, and a second neuromorphic one, sizing the spacecraft for each of them. As for the second research direction, the Stack-CNN detection algorithm was analysed, tested and optimised in order to reduce its computational time, targeting an online space debris trigger. Once optimised, the algorithm was tested on real data, showing excellent performance on both Mini-EUSO-like detector and neuromorphic sensor data. A simulation framework capable of generating synthetic space debris traces was also developed, essential for developing a second deep learning detection algorithm. The latter, characterised by extremely short computational times, showed promising results when tested on Mini-EUSO data. This thesis work has focused on the design of both hardware and software aspects of a CubeSat mission to investigate a category of space debris about which very little is known, aspiring to pave the way for a possible future space observatory consisting of not just a single CubeSat, but of an entire constellation of satellites for the revelation of small-dimension space debris.