An Autonomous Robotic Surgical System (ARSS) requires the integration of both situation awareness and cognitive skills . It also benefits from the presence of an up-todatesimulation of the surgical scenario, that provides support throughout task execution (e.g. complementing intra-operative sensor data) (Fig.1). The simulation is created with personalized physics-based models. However, such models need updating during the intervention based on real data, to correct for possible inaccurate initial parametrization and guarantee that the simulation continuously represents the current surgical state.