IEEE Robotics & Automation Magazine - December 2021 - 68

augmented reality in robotic surgery are reviewed in [51]. A
recent review of gesture analysis in surgical robotics summarized
the state of the art in this field [266].
Discussion
This review article summarized the research facilitated
throughout the first decade of the dVRK by providing a comprehensive
collection of the papers that have been published
thus far in a wide range of research topics. Overall, 253 papers
have been classified into five different categories based on
their application paradigms. In each category, the publication
was then grouped, based also on the type of data it relied on,
and Figure 6 depicts the percentage use of a given type of data
for each research field.
Starting from the au -
A future improvement for
the dVRK platform would
be enabling researchers to
collect and store data with
minimum effort so that it
can be reused for different
applications.
tomation research category,
nearly all of the
papers we reviewed rely
on the use of endoscopic
images and/or KD from
the encoders as the primary
data sources. A
similar trend can be
observed in the imaging
and vision classes, even if
research items based on
KD are slightly fewer. For
training and skill assessment
and gesture recognition,
most of the papers
rely on KD, using any
other type of data in fewer than 50% of the cases or exploiting
external-sensor ED. When it comes to hardware implementation
and integration, nearly all of the types of data cross the
50% threshold, preserving a good balance, except for the KD.
For system simulation and integration, it is possible to notice
how KD and DD are used in the vast majority of publications,
100
75
50
25
RI KD DD SD ED RI KD DD SD ED RI KD DD SD ED RI KD DD SD ED RI KD DD SD ED
Automation
Training, Skill
Assessment, and
Gesture Recognition
Hardware
Implementation
and Integration
System Simulation
and Modeling
Imaging and Vision
leaving the other data types to fewer than 25%. In general, the
correlation between the type of data and each application area
shows the increasing importance of images in RMIS, because
in nearly all of the categories, raw image (RI) crosses 50%.
The extensive use of KD and DD also highlights the significance
of having a research platform, such as the dVRK, which
facilitates the ability to exploit the robot as a haptic interface
and to make use of the systems' data-generation capabilities.
Furthermore, the open access design of the dVRK incentivizes
and enables researchers to integrate it with different types
of hardware and software, as demonstrated by the extensive
usage of external data in nearly all of the classes we covered.
These considerations on data usage and research fields in
the work facilitated by the dVRK can be an interesting stimulus
for reflecting on the optimization or acceleration of surgical
robotic research. Despite the nonexhaustive nature of this
review and analysis, we believe that the information collected
provides a useful basis of the research directions explored by
and enabled through the dVRK. It offers adopters of the
dVRK a comprehensive overview of the research outputs and
a synopsis of activity of the different consortium stakeholders
across the globe.
This article highlighted the importance of data accessibility.
A future improvement for the dVRK platform would be
enabling researchers to collect and store data with minimum
effort so that it can be reused for different applications. For
example, all of the experiments carried out in papers around
the category of surgeon training and skill assessment could be
recorded in centralized data storage and used as a demonstration
to train algorithms for task automation. This links to
areas of active development with research institutions under
research agreements with ISI where data can be recorded
from the clinical setting (using custom recording tools such as
the dVLogger by ISI, as in [270]). Another interesting addition
to the dVRK considering the recent developments in the
automation area, would be to integrate a fully enabled simulation
environment, giving researchers the possibility to test
100
75
50
25
Figure 6. A histogram of data usage (in percentage) for each category based on the publications listed in Table 1. The percentage
refers to the number of publications involving a certain data type out of the total number of publications in a certain research field.
68 * IEEE ROBOTICS & AUTOMATION MAGAZINE * DECEMBER 2021
(%)
(%)

IEEE Robotics & Automation Magazine - December 2021

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - December 2021