IEEE Robotics & Automation Magazine - December 2021 - 4

FROM THE EDITOR'S DESK
Learning to Adapt
By Yi Guo
M
any students in the United
States have returned to the
classroom since the start
of the fall semester after
18 months of online or hybrid classes
caused by the COVID-19 pandemic.
It is a big relief to switch from Zoom
to the classroom, and in-person interactions
in classrooms and laboratories
The survey article
by D'Ettorre et al.
reviews the da
Vinci Research Kit
and its impact in
accelerating surgical
robotics research.
provide better support to students
inte llectually and psychosocially. On
the other hand,
instructors and
students found
that they need
to adapt to a
new normalcy
with face mask
mandates, social
distancing
prac tices, and
occasional ly,
hybrid modes
of t e ac hin g
when needed.
We need to learn to adapt to the changing
classroom and the changing world
more comfortably.
There was no lack of disruption and
rapid changes during the past two years.
The good news is that what doesn't kill
you makes you stronger. Supply-chain
disruptions have affected people's daily
life and their bills. However, experts
have predicted that global supply chains
will survive, and they are adapting but
not falling [1]. A side effect of the panDigital
Object Identifier 10.1109/MRA.2021.3117355
Date of current version: 9 December 2021
demic lockdown is the forced
use of digital teaching materials
that teachers were
hesitant to adopt, many of
which will remain as students
head back to the classroom
[2]. Perhaps the best lesson
we have learned is to adapt to the hard
school of life, embrace these changes,
and build resilience [3].
In classical control theory, the
mechanism of feedback is designed to
control the system to some desired
behavior in the face of uncertainty of
the system and in the presence of
uncontrollable external disturbances
acting on the system. Adaptive control
was developed to dynamically tune
unknown and changing parameters so
that the system under control adapts to
various operating conditions. Machine
learning algorithms embed experience
replay into the learning process so that
the learned system adapts to observation
more efficiently. It is interesting to
observe the interplay between modeland
learning-based techniques in decision
and control and how artificial
intelligence-based methods complement
model-based control toward better
adaptive systems.
Three articles in this December issue
of IEEE Robotics and Automation Magazine
have responded to the call for this
special issue on autonomous vehicles:
artificial intelligence and model-based
techniques in decision and control.
Mitriakov et al. design reinforcement
learning-based control for tracked
robots to traverse staircases, and they
4 * IEEE ROBOTICS & AUTOMATION MAGAZINE * DECEMBER 2021
demonstrate that learningbased
techniques are particularly
promising to solve
problems where it is hard to
model the dynamics of physical
interaction. Yuan et al.
propose a learning method for
highway-driving decision making based
on prioritized experience replay mechanisms
and multireward architectures.
Kim et al. present control of functional
electrical stimulation bikes for rehabilitation,
which can compensate for the
user's muscle fatigue.
In addition to the three themed articles,
this month's issue contains one
regular article and four survey articles.
The regular article by Lepora and
Lloyd describes a new way of controlling
robots using soft tactile sensors
that combine deep neural networks for
pose estimation with visual servoing
by tactile sensing. The survey article by
D'Ettorre et al. reviews the da Vinci
Research Kit and its impact in accelerating
surgical robotics research. Sciutti
et al. survey how the robotics community
in Italy responded to the pandemic
outbreak in the early stages,
sharing the experience and lessons
learned. Birk and Simunovic present
an international survey on the effects
of COVID-19 on robotics laboratory
courses and other forms of hands-on
education in tertiary robotics education.
Although COVID-19 is fresh in
our memory, these two articles offer
retrospectives on robotics research and
education during the pandemic. The
final featured article in this issue,
IEEE Robotics & Automation Magazine - December 2021

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