Task constraints classiﬁcation and exploitation for safe and productive
Nicola Maria Ceriani,Andrea Maria Zanchettin and Paolo Rocco
Tight and continuous human-robot interaction will char-
acterize tomorrow’s robotic applications.Human and robotic
tasks will be less and less distinguishable and the separation
between human and robot working environments will disap-
pear with the removal of safety barriers.In such a scenario
novel safety systems will be required to combine workers’
safety and robots’ productivity.The contribution of this work
to human robot interaction is a method to classify task
constraints based on their relevance for task completion and
to deﬁne a task-consistent collision avoidance safety strategy.
Safe interaction and productive operation can therefore be
more effectively combined.
Fig.1.The distributed sensor system mounted on the ABB FRIDA dual
arm robot prototype.
II.SKILL CONSTRAINTS’ CLASSIFICATION
Industrial robot controllers normally execute tasks by
transforming task constraints,be they instantaneous or not,
into a set of instantaneous constraints,one for each of the ma-
nipulator’s degrees of freedom.For this reason some of the
constraints enforced by the controller may not be necessary
for task completion.A task relevance based classiﬁcation
of constraints has therefore been proposed ,which is an
evolution of natural/artiﬁcial constraints theory ,deﬁning
three classes.The ﬁrst one corresponds to constraints that
cannot be relaxed unless disrupting the skill (“hard”),the
second one to constraints that can be temporarily relaxed
suspending skill execution (“skill”) and the last one to those
The research leading to these results has received funding from the
European Community Seventh Framework Programme FP7/2007-2013 -
Challenge 2 - Cognitive Systems,Interaction,Robotics - under grant
agreement No 230902 - ROSETTA.
All the authors are with Politecnico di Milano,Dipartimento di Elettron-
ica,Informazione e Bioingegneria,Piazza L.Da Vinci 32,20133,Milano,
simply constituting a task over-constraining (“soft”) and that
can be relaxed without inﬂuencing skill completion.
III.TASK CONSISTENT SAFETY STRATEGY
Based on the introduced classiﬁcation,a safety strategy has
been designed.Depending on the potential dangerousness
of the interaction,assessed as proposed in ,constraints
are relaxed consistently with their task relevance allowing
evasive motions.Relaxed constraints are enforced again
by the safety system when danger falls below a deﬁned
threshold.The safety strategy has been applied to an ABB
FRIDA dual arm manipulator prototype which has been
equipped with a distributed proximity sensor.The strategy
has proven to be effective in combining safety motions with
task execution,exploiting available task redundancy:Figure
2 shows the robot evading from a human,without violating
the geometrical constraints imposed by the environment.
Fig.2.The robot evading from a human consistently with geometrical
constraints of the red button housing.
 N.M.Ceriani,A.M.Zanchettin,P.Rocco,A.Stolt,and A.Roberts-
son,“A constraint-based strategy for task-consistent safe human-robot
interaction,” in Intelligent Robots and Systems (IROS),2013 IEEE/RSJ
International Conference on,Accepted.
 M.T.Mason,“Compliance and force control for computer controlled
manipulators,” Systems,Man and Cybernetics,IEEE Transactions on,
 B.Lacevic and P.Rocco,“Kinetostatic danger ﬁeld - a novel safety
assessment for human-robot interaction,” in Intelligent Robots and
Systems (IROS),2010 IEEE/RSJ International Conference on,Oct.,pp.