Machine learning tools to produce 3D shape models of asteroids from radar observations

The near-Earth asteroids (NEAs) are a population of objects on orbits that cross or come near that of Earth; primarily rocky``rubble-pile''aggregates made up of fragments of rock held together by self-gravity and cohesion. NEAs' shapes are linked to their compositions, internal structures, and dynamical histories, giving us insight into the processes that form them. Detailed shape models are valuable for modelling of shape-sensitive non-gravitational forces which play a significant part in their dynamical evolution and moving them into Earth-crossing orbits. Knowing NEAs' shapes can also help facilitate future asteroid exploration and planetary defence mission planning.