The analysis of long-term collision risk space activities relies on the modelling of the evolution of orbital populations. Our ability to estimate the composition and individual states (position, velocity, etc.) of an initial population after decades or centuries of propagation has important implications in the quality of various risk analyses, such as the identification of orbits with a high density of debris, the collision risk induced by a fragmentation, or the construction of environmental indexes. Evolution models often involve a full representation of the population – i.e., every object is described with a full state – at an initial epoch, and the individual propagation of each object to subsequent evaluation epochs. These approaches are computationally prohibitive, and they do not account for the fact that individual states are often unnecessary for long-term analyses focusing instead on collective metrics (e.g., orbital densities, mean number of collisions, etc.). Point process models allow us to establish statistics on the initial population, rather than individual states for each member object, and to propagate these statistics through a dynamical model covering the size of the population (re-entries, launches, etc.) as well the orbital dynamics of objects. Statistics on the evaluated risk are then drawn from the statistics of the population, thus sparing us the need for expensive MonteCarlo simulations involving the sampling of various initial populations, followed by drawing of object births/deaths during the propagation horizon, etc. We will propose in this presentation some use cases of the point process formulation in to support various risk analyses, such as the construction of environmental maps, or the long-term evolution of orbital densities under several hypotheses of mitigation/remediation options.
Delande et al. (Fri,) studied this question.