"All for One and One for All": Systematic Data Collection and Sharing to Advance Socially Assistive Robots
Over the past decade socially assistive robots (SAR) have shown great promises to
be used in healthcare domain for improving the quality of lives of elderly and people with various
physical and cognitive disabilities. Despite their promises, such robots are barely seen in the real
world serving people in need (Paro is probably the only SAR that has actually been used in various
clinical settings). A growing opinion explaining this gap is that the health-care oriented research
on SARs is less concerned about proving the utility of robotics systems and hence fails to meet the
ultimate goal of deployment in clinical settings. SARs are application oriented technologies and, therefore,
their deployment in clinical settings will not be possible until they can prove their utility.
The robotics community, however, believes that, in many cases, technologies are not ready yet to support the
deployment of SARs in clinical settings. The goal of this workshop is to bring robotics and clinical
researchers together in order to establish an understanding of and approaches to mitigate the issues
that are slowing down clinically meaningful advancements of SARs. In this workshop, we particularly
want to focus on issues related to data collection and sharing that would enable more speedy research
and evaluation in the SARs community. .
The workshop will focus on two themes:
- Human-robot interaction, human factors and ergonomics: Irrespective of application domains, a
user-centered (or bottom-up) development of SARs requires that the opinions of potential end-users are reflected in the design
and development process. Data from carefully designed HRI studies, surveys, or
focus group that provide useful information on making SARs clinically useful will be
discussed under this theme. Examples of such information include, but not limited to,
design guidelines for conducting studies with SARs, universal metrics to evaluate SARs
in clinical settings, utility/usability analysis strategy for SARs, etc. Possible
application areas include, but no limited to, autism, elder-care, personal assistance
for people with various physical and /or cognitive disabilities.
- SAR research database: Sharing of study data, study protocol, codes, participants? demographics, etc. is not common in research
on SARs. Legal issues and IRB restrictions are two reasons, among many, behind the lack of data sharing
culture among SAR research community. But this lack of sharing is causing every SAR researcher to start
their research from the scratch. In other words, everyone, to some extent, is ?re-inventing the wheel?.
Lack of data sharing also makes quantitative comparison of results between two research groups very
difficult to perform, if not impossible. This is particularly limiting in cases where the study population
is difficult to recruit from, which is often the case in assistive applications dealing with vulnerable participants.
We are living in a time of big data. The practice of code and data sharing has become a standard in many branches
of robotics and AI research and has contributed significantly to a rapid uptake of corresponding research domains.
It is, therefore, important to devise a way so that SAR researchers can share data, protocol, and codes related
to human subject studies while conforming to the legal issues and health-care regulations.
We encourage submissions of 2-page extended abstract or 4-page position paper on a topic related to any of the two themes of the workshop.
Contributions should be prepared in standard ACM conference format. Please send submissions
in PDF to HRI.SAR.firstname.lastname@example.org. All accepted position papers will be assigned 15 minutes for oral presentation.
All extended abstracts will be presented in a 1.5-hour poster session. Position papers and workshop notes will be
published on the workshop website. Accepted papers will be available on CD and published on the workshop website
- Jan 27 2017: Submission deadline
- Feb 06 2017: Notification of acceptance
- Feb 17 2017: Final version due
- Mar 06 2017: Workshop in Vienna, Austria