The asteroid belt is believed to be grinding down through collisions, yet, surprisingly, only a few impacts between asteroids have ever been witnessed. Even more surprisingly, the record of these long sought collisions might be already out there. This project aims at finding evidence of past collisions between small to medium sized asteroids (<100 km) from the dynamical analysis of their orbits, as retrieved from ESA's NEOCC database (DB). To this aim a tailored pruning algorithm will be built, capable of scanning the orbits DB to identify pairs of medium size dynamical outliers, i.e. objects marginally fitting the standard dynamical model, with a possible encounter geometry. On all identified cases a further dynamical analysis will then be performed, to better define the sudden velocity change. To discriminate between impacts and flybys, precoveries of the deflected asteroid will be sought after. For the cases where a deflecting asteroid is not found, and can thus be attributed to an actual impact, the dynamics of the collision event will be studied, while for other cases a precise estimation of the deflector mass will be obtained as a by-product of discovered flyby, enabling an estimation of the asteroid bulk density. This would provide insights on the interior structure and composition of the object, in a size regime (tens of kilometers) that is poorly known today. The obtained precise mass estimations would gain particular interest for objects of rare spectral classes (e.g. D, A, E) the composition and structure of which are either largely unknown or is thought to be peculiar.
