Artificial Intelligence Applications in Identifying Close Approaches Between Satellites and Space Debris, Avoiding Collision in Orbit: A Scoping Review

Space debris poses a growing threat to space exploration (Yap, Heiberg & Truffer, 2023), as it can collide with active satellites at high speed, generating thousands of new fragments and exacerbating the scenario known as "Kessler syndrome" (Thompson, 2015; Barry, 2022). This debris comes from inactive satellites, rocket stages, and other equipment left in orbit. In addition to accidental collisions, other sources of generation include mission terminations, anti-satellite weapons tests, and degradation caused by natural factors (Colvin, Karcz & Wusk, 2023). Although research is being done into debris mitigation and removal (Migaud, 2020; Hoeher, 2023; Viavattene et al., 2022), the primary strategy remains collision avoidance through the identification of critical approaches. To this end, data such as Conjunction Data Messages (CDM)—which predict the point of closest approach (TCA) up to seven days in advance—and Two-Line Elements (TLEs), updated daily (Merz et al., 2017) are used. In this context, Artificial Intelligence (AI) techniques have shown promise in analyzing large volumes of orbital data, enabling trajectory prediction and identification of close approach risks. Therefore, considering that scoping reviews are recommended for mapping emerging areas, this review aims to characterize AI-based approaches applied to satellite-debris collision avoidance, focusing on the techniques used, their performance, and their contribution to increasing the accuracy of satellite-debris collision prediction.