Human Machine Interaction and Communication in Cooperative Actions - Bioinspired Approaches for Human-Centric Technologies - page 239

visual, and vestibular feedback generated by both partners simultaneously and by
the forces of motion and gravity. CI directs the dancer's attention to sensation and
nonverbal communication rather than execution of specific movement sequences or
visible appearance. Remarkably, we know little about how this channel is used to
transmit information between partners. To date, only a handful of studies have
investigated physical interaction in joint actions (review in Reed 2012 ).
To a large extent, this chapter focuses on emergent coordination and implicit
communication in human-robot joint actions. Because robots differ from humans in
several ways, human-robot interactions raise novel questions with respect to
human-human interactions:
First, robots in general do not look like humans and, even more often, they do not
move like humans. This basic observation raises questions about whether emergent
coordination does or can occur in human-robot interaction since this form of
coordination is based on subtle cues about the body and the movement of the
partners. These questions are addressed in Sect. 8.2 , which analyzes whether and
under which conditions a robotic device can trigger this form of covert communi-
cation with a human partner.
Second, the body of machines and robots differ greatly from the human body, in
ways that go well beyond what can be observed visually. As a matter of fact, most
robots (e.g., all industrial robots) are rigid and powerful artifacts which are dan-
gerous to interact physically with. Moreover, physical interaction with these robots
is limited by the fact that they fully control the movements which must then be
passively followed by the partner. This is clearly a strong drawback in robot-
assisted rehabilitation where an active participation of the patient is crucial to
obtain good results. Section 8.3 is dedicated to physical interaction between
humans and machines. Section 8.3.1 proposes a classification scheme of tasks
involving physical interaction and a measure of the complexity based on the
concept of compliance. In particular, Sect. 8.3.2 reports work done to develop
novel compliant actuation technologies that aims at facilitating physical interaction
between a robot and a person or its environment. Section 8.3.3 analyzes human
control strategies when interacting physically with unpredictable environments
simulated by a robot. Section 8.3.4 reports work done to optimize the treatment
of people affected by neuromotor diseases like stroke in the context of robot-
assisted rehabilitation of the upper limb. The final section emphasizes the impor-
tance of human studies for the development of robots that can interact with humans
and alludes to the ethical and societal questions that partnerships between humans
and robots will have to be addressed as machines play little by little an ever larger
role in our daily life.