IEEE Robotics & Automation Magazine - September 2023 - 137

benchtop machines and ubiquitous rapid-prototyping methods
within 2 h. It can be integrated with a variety of sensors, the
data from which can be processed through the Robotic Operating
System-capable computer on board. Taking the idea of
simple fabrication further, the US$20 chassis of the BigANT
robot [84] requires only a foam board and minimal hand tools
to craft the structure of the robot. The plate and reinforced
flexure (PARF) joints in this design can be employed in any
number of functional robot designs and facilitates a quick and
inexpensive fabrication of robot bodies.
In addition to the aforementioned applications and morphologies,
legged robotics research also spans domains of
kinematics and dynamic control as well as lower-limb prostheses
and exoskeleton designs. The HOPPY kit [71] [see Figure
4(c)] for robotics education allows students to experiment
hands on with topics in robotics that range from control and
trajectory generation to simulation and kinematics on modular
and robust physical hardware. The low cost of the kit
makes access easier for researchers and educators studying
dynamic behavior in robotic systems. The Open Source Leg
(OSL) [72] [see Figure 4(d)] project likewise aims to lower
barriers of entry in leg prosthesis research by eliminating the
time and resources required in developing one's own robotic
leg system from scratch, and presenting a common testbed
across the field for benchmarking control approaches. The
OSL project also demonstrates some developer practices that
can encourage adoption within the community, such as creating
an online forum [85] for users to share modifications and
ask questions and a step-by-step video assembly guide [86].
Users also have the option to purchase a fully built OSL from
a vendor, and the released design and documentation files can
help users with future upgrades and repairing the leg. Similar
to the OSL, the Alice pediatric exoskeleton [87] was developed
with the goal of improving access to lower-limb assistive
technologies. Alice can be controlled using simple, low-cost
electronics, and the physical components of the exoskeleton
are 3D printed or purchased off the shelf, and Alice has even
been clinically validated on patients with different medical
conditions. That said, lower-limb exoskeletons for adults are
still missing from the open source landscape, likely because
they would require fabrication with the metal and other highstrength
material to sustain the heavier loads during use.
Legged robotics research and education can be cost prohibitive,
especially if the robots require advanced manufacturing
methods or rely on domain-specific knowledge to build and
operate. However, these ORH projects are improving access
to physical robot platforms in their respective fields through
their comparatively lower cost of entry and minimal fabrication
requirements. Across the different robotic leg systems,
one can now find an ORH morphology suitable for his or her
application, either as it is from the developer or, after some
modifications, permitted by the open source license.
MOBILE ROBOTS
Mobile robots are an effective, low-cost, and versatile platform
in robotics education and research, and a number of
mobile robot platforms at every scale have been proposed,
some of which have also been open sourced. One of the most
popular mobile robots, E-puck [88] [see Figure 5(a)] was
designed with robotics education in mind, and teaches students
concepts in signal processing, embedded programming,
autonomous control, and distributed systems. The simple
structure of the robot, equipped with a microcontroller and a
wide variety of sensors, makes for an easy-to-use robot that
also has a lot of flexibility for implementation. Although the
E-puck robot may be directly purchased from vendors, its
part files robot are also publicly available for users to make
design changes or build on their own. Other small, twowheeled
open source mobile robots such as Mona [89] and
Miniskybot [90] have also been successfully implemented in
educational settings, both of which offer lots of flexibility in
programming and an inexpensive construction. A tracked
mobile robot with 3D-printed structures and mostly off-theshelf
components have also been proposed as benchmarking
tools in computer vision and artificial intelligence research in
addition to education; Zumy [91], Veter [92], and Nanosaur
[93] robots are furnished with more capable onboard computing
for users in these domains to test their software on a
physical platform, which can be quickly and easily built. To
provide students interested in mechanics, electronics, and
programming hands-on experience in building and operating
a robot, the European Space Agency (ESA) and NASA's Jet
Propulsion Laboratory (JPL) have also released their own
scaled-down versions of six-wheel Mars rover designs [see
Figure 5(b)], along with detailed instruction guides of the fabrication
and assembly processes on their websites [94], [95].
Beyond educational applications, robot vehicles have also
been used as popular testbeds in a range of robotics research
fields. A popular family of robot vehicles, the TurtleBot3 [96],
offers a low-cost mobile robot kit that can be adapted to be
employed in a range of research applications [see Figure 5(c)].
Although users commonly purchase the TurtleBot3 hardware
directly from distributors, developers have released the design
files for the robot so that users can customize and modify
the robot as needed. A lot of open source mobile robots rely
heavily on off-the-shelf components so that the hardware can
be effortlessly put together by the user. The Zumy [91] mobile
platform is built mostly from off-the-shelf components and
is intended for development of multirobot systems and computer
vision. Finally, scaled-down miniracecars like the MIT
RACECAR [97] and the Multi-agent System for non-Holonomic
Racing (MuSHR) [98] can serve as powerful platforms
for implementing autonomous navigation, localization, and
planning algorithms in real-world environments for robotics
research and education [see Figure 5(d)]. But, users can also
modify these open hardware cars to adapt the testbeds, for
instance, by adding a gripper to the MuSHR to investigate collaborative
multirobot manipulation [98].
Mobile robots serve as a very effective tool for testing im -
plementations of software algorithms in a range of robotics
domains quickly with a reduced cost of materials and a simple
construction requiring nominal expertise and tools. Their ease
SEPTEMBER 2023 IEEE ROBOTICS & AUTOMATION MAGAZINE
137
IEEE Robotics & Automation Magazine - September 2023

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - September 2023
