Politecnico di Milano

Although robustness and safety of industrial robots have considerably increased
in the last decades, common industrial manipulators are still largely unable to
cooperate with human beings in a natural way.
A novel fact in recent years which may contribute to alleviate this situation is
the commercial availability of redundant manipulators (i.e. robotic arms having
more degrees of freedom than those strictly necessary to perform a certain task).
It is well known that such robotic arms allow to achieve high levels of dexterity:
since the same task can be performed in several ways, redundant manipulators of-
fer a wide range of flexibility in motion planning. Moreover, thanks to the extra
Degree(s) of Freedom (DOF), it is possible to modify in real-time the behavior
of the robot, in order improve safety in the cooperation with human co-workers,
without modifying the end-effector motion.
The aim of this research is to study strategies to exploit the kinematic redun-
dancy of robot manipulators in order to make their motion as natural as possible.
Since the robot is supposed to cooperate with humans, its motion should be pre-
dictable, safe, and intuitive. This work is then focused on the development of cri-
teria to design, and enforce on an industrial robot control architecture, a human-
like kinematic control for redundant manipulators. In this way, a human-friendly
behavior of the robot will ensure a natural cooperation between robotic manipu-
lators and humans, which is not obtained with today commercial industrial ma-
nipulators.