Key points are not available for this paper at this time.
Geospace is crowded due to the proliferation of satellites and space debris and will become more crowded with the increasing deployment of new space missions. This trend is rapidly increasing the probability of collisions between space objects. Space objects fly at extreme speeds; hence, the consequences of collisions are catastrophic. However, accurate and efficient conjunction assessment (CA) and collision avoidance (COLA) have long been challenging, even with the current space catalogues of O(104) size. As the space catalogue size increases owing to the increased number of new satellites, improved sensor capabilities, and Kessler syndrome, the situation will worsen unless a paradigm-transforming computational method is devised. Here, we present the SpaceMap method, which can perform real-time CA and near-real-time COLA for O(106) or more objects, provided that the spatiotemporal proximity amongst satellites is represented in a Voronoi diagram. As the most concise and efficient data structure for spatiotemporal reasoning amongst moving objects, Voronoi diagrams play a key role in the mathematical and computational basis for a new genre of artificial intelligence (AI) called space–time AI, which can find the best solutions to CA/COLA and other space decision-making problems in longer timeline windows. The algorithms are implemented in C++ and are available on GitHub as AstroLibrary, which has RESTful APIs and Python packages that can be called from application programs. Using this library, anyone with elementary programming skills can easily develop efficient applications for challenging spatiotemporal problems.
Choi et al. (Fri,) studied this question.