Morphological optimization of prosthesis' finger for precision grasping

J. Ramirez (1) , A. Rubiano (2) , N. Jouandeau (3) , L. Gallimard (2) , O. Polit (2)

1 UGR - Universidad de Granada = University of Granada
2 LEME - Laboratoire Energétique Mécanique Electromagnétisme
3 LIASD - Laboratoire d'Informatique Avancée de Saint-Denis

In this paper, we present the morphological optimization of our tendon driven under-actuated robotic hand prosthesis' finger, to improve precision grasping. The optimization process is performed with a black box optimizer that considers simultaneously kinematic and dynamic constraints. The kinematic is computed with the Denhavit-Hartenberg parameterization modified by Khalil and Kleinfinger and the dynamic is computed from the virtual displacements and the virtual works. All these constraints are considered as a fitness function to evaluate the best morphological configuration of the finger. This approach gives a way to introduce and improve soft and flexible considerations for the grasping robots such as hands and grippers. Theoretical and experimental results show that flexible links combined with morphological optimization, lead in more precise grasping. The results of the optimization, show us an important improvement related to size, torque and consequently energy consumption.

Mots clés: Morphological optimization; mechanisms prehension; precision grasping; soft robotic

Domaines: Intelligence artificielle [cs.AI]

J. Ramirez, A. Rubiano, N. Jouandeau, L. Gallimard, O. Polit. Morphological optimization of prosthesis' finger for precision grasping. 4th International workshop in medical and service robots, Jul 2015, Nantes, France. ⟨10.1007/978-3-319-30674-2_19⟩. ⟨hal-02317209⟩