Human Machine Interaction and Communication in Cooperative Actions - Bioinspired Approaches for Human-Centric Technologies - page 255

force-controlled manipulanda, consists of a mass (the “tool-tip”) and two handheld
nonlinear springs. The tool-tip operates in an unstable environment, characterized
by a saddlelike force field, with mediolaterally oriented unstable manifold and
anteroposterior oriented stable manifold. The nonlinearity of the two springs allows
the users to affect size and orientation of the tool stiffness ellipse, by using different
patterns of bimanual coordination of the two spring terminals: minimal stiffness
occurs when the two spring terminals are aligned and the stiffness size grows by
stretching apart the two terminals. The tool parameters are set such that minimal
stiffness is insufficient to provide stable equilibrium of the tool-tip, but asymptotic
stability can be achieved with sufficient stretching, although at the expense of a
larger effort. As a consequence, tool users have two possible strategies for stabi-
lizing the tool-tip in different regions of the workspace: (1) high stiffness strategy
aiming at asymptotic stability and (2) low stiffness positional strategy aiming at
bounded stability, similar to the manual stabilization of an inverted pendulum. The
behavior of na¨ve users is spontaneously clustered into two groups of approximately
equal size: a stiffness strategy group and a feedback strategy group (Saha and
Morasso 2012 ). In a following study (Zenzeri et al. 2013 ) subjects were trained to
become expert users of both strategies in a discrete reaching task. Then the
generalization capabilities were tested by means of a continuous stabilization task
which consists of tracking a target in the unstable workspace: the results show that
human subjects can learn to master complex interaction tasks alternating different
interaction strategies.
8.3.4 Human-Robot Interaction in Neuromotor
Rehabilitation
Robot therapy is slowly emerging as an acceptable technique for the routine
treatment of people affected by neuromotor diseases like stroke (Mehrholz
et al. 2012 ; Krebs and Hogan 2012 ). However, there is still little agreement on
the theoretical background that is necessary for overcoming the current empirical
approach and for attempting to optimize and personalize the treatment provided by
the robot. This is indeed the topic which is currently the focus of the research
carried out in the MLRR-lab.
After the studies on animal models of stroke by Nudo ( 2006 , 2007 ), it has
become clear that beyond time-dependent spontaneous neurological recovery, the
principal process responsible for functional recovery is the use-dependent reorga-
nization of neural mechanisms made possible by neural plasticity . So neural
plasticity is important, but is it the effect of intensive and repetitive robot therapy
or the other way around? We are in favor of the latter option, because in our opinion
such fundamental property of neural tissue is the prerequisite for assuming that
robot therapy may have a chance of success, inducing a reorganization of the
damaged brain of the patient via appropriate physical interaction patterns.