IEEE Robotics & Automation Magazine - March 2017 - 84

publication. This can multiply by factors of hundreds the price
of the CAD license. This legal issue is totally underestimated
by both people participating in such projects and by the CAD
editors, and it is a potential threat for many projects.
Who Designs, Produces,
and Supports the Hardware?
In the definition of open hardware given in the "Background"
section, we stated that one should offer "hardware whose
design is made publicly available so that anyone can ... make
... hardware." Should "anyone" mean every single person or
only companies able to produce the product?
In our project, we decided to have two different types of
hardware: the robot itself and the accessories. The robot is the
expensive part and has a very neutral design, allowing adaptation to specific situations. This adaptation is achieved by custom
accessories that increase the attractiveness of the robot in its specific role, enabling activities for different ages and genders.
For the robot itself, we opted to interpret "anyone" as professional structures able to mass-produce hardware based on
the price and complexity of the product.
For the accessories, less technically challenging and stronger linked with creativity and educational value, we promoted
techniques that are accessible to anyone in the broader sense:
paper, cardboard, LEGO constructions, and 3-D printing.
This allows a much broader spectrum of contributors, including teachers and lay people.
Another strong element of our vision of open hardware is
that Thymio should be durable. As schools invest in longterm training of teachers, for instance, the lifetime of the
products should also be as long as possible. The open hardware approach fits well to this requirement, as it gives to the
user, or to a generic technician, better conditions to repair the
system, having the specifications of all components. This is
not the case of proprietary robots like Edison, for instance.
Supporting this type of operation has an impact on the robot
design; e.g., Thymio can be easily opened with four standard
screws, and we introduced connectors between key elements
such as motors, the speaker, the battery, and the main PCB.
Another key element in supporting repairs by end users is the
documentation of calibration methods. When choosing very
low-cost components, one faces large dispersion of characteristics. For example, in the Thymio, the right and left wheel
motors can differ in their electrical characteristics, resulting in
the robot not going straight for similar speed commands to
both wheels. To correct this problem, we introduced factory
calibration. To allow the user to replace a broken motor, it is
essential to also give him or her the possibility of recalibrating
the robot and adjusting the parameters of the new motor. In
Thymio, this results in the design and the documentation of
calibration processes that can be performed by anyone, getting close to the original definition of open-source hardware.
Conclusions
The introduction of robots in formal education is a very challenging task, not only because of technical requirements such
84

IEEE ROBOTICS & AUTOMATION MAGAZINE

March 2017

as low cost and interactivity but also because of factors
depending on the school environment, such as the diversity of
the educational programs, the dependence on local structures
and languages, or the required training of teachers. Most of
the current educational robotic activities capitalize upon one
strong element, e.g., the technical innovation, but miss matching the formal education requirements.
The open-source hardware approach used in the Thymio
project addresses this and several other issues found in educational robotics. The inclusion of education scientists, teachers,
and designers is possible because of the open nature of the project and the split between the core technology, which is produced in a central way, and accessories, which are accessible
with do-it-yourself approaches. This split also allowed ensuring
a gender- and age-neutral basic robot. This large inclusion of
users and contributors allowed the production, in parallel to the
robot technology, of a large set of pedagogic scenarios.
The philosophy of open source and free access to
information fits extremely well with the community of
users in education and was reinforced by producing the
robot in a nonprofit structure. This approach allowed
broad distribution of the robot with minimal charges due
to management of intellectual properties, royalties, financial support, etc. Moreover, the open-source approach allows
provision of a durable robot, easy to maintain and repair,
with a community of users providing educational material
and mutual support.
By conducting a survey among contributors to opensource hardware projects, we could observe that our project
shares some characteristics with the majority of the projects
represented in the survey. We identified, for instance, an underestimated legal issue for open-source hardware projects in the
licensing term of CAD software. Finally, we could show some
elements, specific to educational robotics, that differentiate
our project from other open-source hardware projects. In
particular, our project takes advantage of an alignment
between the principles underlying open source and the nature
of education institutions. We also found a solution to the
problem of production methods by splitting our hardware
into two categories, enabling both advanced technology for
the robot and a large variety of accessories. Hence, Thymio
appeals to a broad community of end users in education,
addressing durability and inclusion at several levels.
Acknowledgments
This research was supported by the Swiss National Center of
Competence in Research "Robotics" (Thymio robot development and deployment), GebertRuf Stiftung (design of
accessories), Swiss National Science Foundation Project
CRAGP2 151543 "Robotics in Schools" (educational material), and European Union FP7 Project ASSISIbf 601074
(survey on open hardware). Many thanks to Luc Bergeron
and his team at écal.ch for the industrial design of Thymio;
Didier Roy, David Sherman, and all of the French Institute
for Research in Computer Science and Automation team for
their contributions and diffusion in France; Gordana Gerber

http://www.cal.ch

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - March 2017