In the industrial field a lot of tasks can be executed automatically by means of robotic manipulators, e.g. pick and place, welding, assembly or product inspection. For this reason the way the trajectory is “taught” to the robot becomes crucial. Teaching techniques involving manual guidance of the robot maximize the possibilities of re-programming the robot in a intuitive and easy way, because they don’t need the knowledge of the robot programming language.
To supervise the physical human-robot interaction (pHRI) during tasks requiring cooperation between the human operator and the robotic manipulator, with an efficient and cheap setup, compliant motion control strategies (i.e. admittance/impedance control) can be implemented. The forces/torques arising during the interaction are measured by a Force/Torque (F/T) sensor typically mounted on the robot wrist flange.
Thanks to this kind of controllers, the “walk-through programming” can be exploited: the operator becomes like a teacher that physically follows (“walks”) the end-effector through the desired positions. At the same time the robot’s controller records the joints coordinates, “learning” the desired trajectory, and can play the trajectory back thereafter. In order to create an industrial setup, the tool attached at the robot end-effector has to be considered and its dynamics has to be compensated.
This research area started from the development of walk-through programming techniques for manual guidance of robotic manipulators and is currently focused on the development of advanced techniques of physical human-robot interaction.
ADAPTIVE Manufacturing – CLUSTER NAZIONALE FABBRICA INTELLIGENTE