Simulation-based Bayesian inference for robotic grasping
This addresses the challenge of robotic grasping in uncertain environments, though it appears incremental as it builds on existing simulation and optimization methods.
The paper tackled the problem of controlling robotic grippers under uncertainty by computing 6-DoF grasp poses using simulation-based Bayesian inference, achieving a high success rate in benchmarks.
General robotic grippers are challenging to control because of their rich nonsmooth contact dynamics and the many sources of uncertainties due to the environment or sensor noise. In this work, we demonstrate how to compute 6-DoF grasp poses using simulation-based Bayesian inference through the full stochastic forward simulation of the robot in its environment while robustly accounting for many of the uncertainties in the system. A Riemannian manifold optimization procedure preserving the nonlinearity of the rotation space is used to compute the maximum a posteriori grasp pose. Simulation and physical benchmarks show the promising high success rate of the approach.