IEEE Robotics & Automation Magazine - June 2021 - 10

FROM THE GUEST EDITORS
Emerging Paradigms for Robotic
Manipulation: From the Lab
to the Productive World
By Maria Pozzi, Virginia Ruiz Garate, Arash Ajoudani, Kensuke Harada, Monica Malvezzi, Leonel Rozo, and Maximo A. Roa
N
ovel robotic developments
are going hand in hand with
the need for innovation in
manufacturing and service
applications. Large industries increasingly
need fast-adapting production lines,
whereas small and medium-sized enterprises
are adopting collaborative manipulators
to gain a competitive edge
in global markets. Additionally, a
growing number of companies are
producing and using service robots
for a variety of applica tions, from agriculture
to surgery. In this emerging
framework, robots should be able to
grasp and manipulate different tools
as well as operate in fast-changing
or even unstructured surroundings,
possibly interacting with human users
or coworkers. These real-world scenarios
require going beyond tra ditional
parallel-jaw grippers and preprogrammed
or teleoperated tra jectories
and require searching for innovative
grasping and manipulation paradigms
that can cope with the needs of uncertain
environments. Proposed solutions
include soft manipulation, data-driven
grasp planning, and human-robot collaborative
manipulation. But how long
will it take to see these novelties outside
of research labs?
For this quest to be successful, we
cannot focus on solving only purely
technological and application-driven
issues; we need a holistic, interdisciplinary
view of the robotic system. Basic
research on single components (robotic
Digital Object Identifier 10.1109/MRA.2021.3067685
Date of current version: 11 June 2021
10 * IEEE ROBOTICS & AUTOMATION MAGAZINE * JUNE 2021
grippers, sensing, and grasp control)
and a deep understanding of the
scenario(s) in which the robot will work
are both fundamental to achieving the
desired result. In this context, a tight
collaboration among research scientists
in different fields and final end users or
producers of the technology is required.
This special issue aims to stimulate
and gather publications describing how
new approaches in the field of robotic
manipulation can be (or have already
been) transferred from research labs to
the productive world. Nine articles have
been included in this collection, and we
briefly summarize them in the following.
Robotic hands and grippers represent
the ultimate interface between
robot arms and manipulated objects.
Their design and control still pose
important challenges to the research
community, and there is a growing
trend to embed part of the intelligence
of the manipulation system directly into
the hardware design of the end effector
without demanding it completely from
the robot software. For example, the use
of compliant elements or soft materials
ensures an intrinsically safe interaction
with the environment as well as a passive
shape adaptation between the gripper
and the grasped object. The first
three articles of the special issue focus
on the development of new end effectors
based on different design principles.
Kim and Deshpande propose a
method to optimize the nonlinear parallel
compliance of a robotic gripper to
meet stability goals and further reduce
the overall stiffness of a device through
a coupled tendon routing. McCann et al.
present the design, modeling, dimensional
synthesis, and experimental
characterization of a robotic hand
having a parallel structure inspired by
the Stewart-Gough platform. Wang et al.
describe the design and fabrication of
a novel humanoid hand made of
intrinsically soft materials and endowed
with pneumatic fingers and a palm
capable of adapting to several different
grasped objects.
Alongside suitable end effectors
that allow the physical handling of
objects, deploying a manipulation system
into a real-world scenario, it is
fundamental to endow the system
with proper sensing capabilities.
Because tactile sensing requires direct
contact between surfaces and due to
the lack of a solid and robust tactile
sensor that can resist everyday use,
tactile sensing in industrial settings
has so far been discouraged, and
vision-based solutions are preferred.
The research field of robotic vision
has reached a high level of maturity,
and its application to manipulation
tasks has recently been expanded by
the introduction of machine learning
algorithms capable of identifying,
localizing, and classifying objects
based on large training data sets of
images. In the article by Koskinopoulou
et al., a deep-trained computer
vision module for the categorization
of recyclables is developed and integrated
with a robotic manipulation
system (a delta robot equipped with a
blower-based vacuum gripper) for
sorting recyclables in a material
recovery facility.

IEEE Robotics & Automation Magazine - June 2021

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