Optimal control simulations of two finger grasping

Grasping is a complex human action simulated here as an optimal control problem with a three‐dimensional rigid multibody model composed of two fingers along with the wrist and the forearm. The dynamics is described by a hybrid dynamical system with a given switching sequence (reaching or tree kinematic structure and grasping or closed loop contacts) and unknown switching times. The optimal control problem is solved using the direct transcription method DMOCC (discrete mechanics and optimal control), see [1], leading to a structure preserving approximation. An objective function such as the distance of contact points from the grasped object center of gravity or the sum of normal contact force is minimised subject to the discrete Euler‐Lagrange equations, boundary conditions and path constraints. The object dynamics is also taken into account.