🤖 AI Summary
Robots struggle to achieve stable object placement under geometric uncertainties—including object pose and shape variations, as well as discrepancies in environmental models.
Method: This paper proposes an end-to-end optimization framework grounded in differentiable contact dynamics. It integrates differentiable contact modeling with uncertainty-aware multi-hypothesis belief tracking to mitigate the sensitivity of gradient-based optimization to initial conditions; unifies the representation and closed-loop calibration of heterogeneous geometric uncertainties; and leverages real-time force/torque sensor feedback for online parameter estimation.
Contribution/Results: Evaluated on a physical robotic platform, the method demonstrates strong robustness against all three categories of geometric uncertainty. It significantly improves both the success rate of stable placement and the quality of contact interactions, enabling reliable manipulation under realistic modeling imperfections.
📝 Abstract
From serving a cup of coffee to carefully rearranging delicate items, stable object placement is a crucial skill for future robots. This skill is challenging due to the required accuracy, which is difficult to achieve under geometric uncertainty. We leverage differentiable contact dynamics to develop a principled method for stable object placement under geometric uncertainty. We estimate the geometric uncertainty by minimizing the discrepancy between the force-torque sensor readings and the model predictions through gradient descent. We further keep track of a belief over multiple possible geometric parameters to mitigate the gradient-based method's sensitivity to the initialization. We verify our approach in the real world on various geometric uncertainties, including the in-hand pose uncertainty of the grasped object, the object's shape uncertainty, and the environment's shape uncertainty.