Evolving the Neural Controller for a Robotic Arm Able to Grasp Objects on the Basis of Tactile Sensors

Abstract

We describe the results of a set of evolutionary experiments in which a simulated robotic arm provided with a two-fingers hand has to reach and grasp objects with different shapes and orientations on the basis of simple tactile information. Obtained results are rather encouraging and demonstrate that the problem of grasping objects with characteristics that varies within a certain range can be solved by producing rather simple behavior that exploit emergent characteristics of the interaction between the body of the robot, its control system, and the environment. In particular we will show that evolved individuals does not try to keep the environment stable but on the contrary push and pull the objects thus producing a dynamics in the environment and exploit the interaction between the body of the robot and the dynamical environment to master rather different environmental conditions with rather similar control strategies.