Robotics in education: a tool for recruiting, engaging, retaining and educating students

Abstract

Robotics is being used in the Montana Tech Computer Science Department in a variety of ways: to increase awareness of the field, to assist us in recruiting students, to engage our existing students, and to provide an advanced research and development experience for our upper level students. The use of robotics in our Computational Thinking course has shown statistically significant effect in the tested increase in logical and mathematical thinking as measured with a pre-test and post-test, which has been administered over the last three years that the course has been offered. Other robotics projects that are underway will take time before results are expected, including the implementation of our NAOCam project, since it is still in its design phase. Thus far, though, our results are promising, and we are optimistic that as additional robotics efforts are completed, those results will also be positive.

Full text

Robotics in Education: A Tool for Recruiting, Engaging, Retaining and
Educating Students
MICHELE VAN DYNE, JACOB FJERMESTAD
Computer Science Department
Montana Tech of the University of Montana
1300 W. Park St., Butte, MT 59701
USA
mvandyne@mtech.edu jtfjermestad@mtech.edu http://cs.mtech.edu
Abstract: - Robotics is being used in the Montana Tech Computer Science Department in a variety of ways: to
increase awareness of the field, to assist us in recruiting students, to engage our existing students, and to
provide an advanced research and development experience for our upper level students. The use of robotics in
our Computational Thinking course has shown statistically significant effect in the tested increase in logical
and mathematical thinking as measured with a pre-test and post-test, which has been administered over the last
three years that the course has been offered. Other robotics projects that are underway will take time before
results are expected, including the implementation of our NAOCam project, since it is still in its design phase.
Thus far, though, our results are promising, and we are optimistic that as additional robotics efforts are
completed, those results will also be positive.
Key-Words: - Robotics, education, network control, NAO, Lego Mindstorm NXT, computational thinking
1 Introduction
Computer science as an educational discipline has
seen a decline in enrollment in the last decade,
although the trend appears to be reversing. As the
job market continues to be promising for the
graduates of the discipline, we see more and more
students interested in majoring in computer science.
Unfortunately, the Computer Science Department at
Montana Tech has found that many who show initial
interest become disillusioned with programming
early in their education, and there are many more
who don’t even consider it as a major. Thus, we
have problems attracting students since they are
unaware of the area, and problems retaining those
who become majors who don’t understand the
nature of the discipline.
These issues have been attacked on many fronts,
but the approach of interest in this research is that of
incorporating simple robotics into the pre-
curriculum and more advanced robotics for
undergraduate research and development.
Robotics captures the imagination, both as a
means of accomplishing tasks, and as a model of
human behavior. The area encompasses both the
electro-mechanical side of creativity, and the
software/behavioral side. We have used robotic
technology to stimulate interest of potential students
through incorporating robotic demonstrations and
hands-on experience. We have also used robotics
with current students, first as an approach to
teaching logical problem solving, and then as a tool
for advanced research and development. These
different uses of robotics complement each other,
thereby forming one method of addressing the
problems of recruiting and retention.
2 Problem Formulation
Montana Tech is primarily an engineering school,
and approximately half of our student population
comes from the state of Montana, although some of
our programs draw national and international
students. The computer science department
generally draws the majority of its students from in-
state. Montana is a sparsely populated state, with
most areas considered rural. Because of this,
computer science, and rigorous science beyond the
basics, is not emphasized in most pre-college
curricula. Consequently many students entering the
university are not aware of computer science as a
discipline or a career path.
In addition to lack of awareness of the field,
many high schools don’t prepare college-bound
students for the more rigorous mathematical
requirements of engineering and science required at
the university level. Many students entering our
program are underprepared mathematically. A
significant number of students declare computer
Recent Researches in Circuits, Systems, Multimedia and Automatic Control
ISBN: 978-1-61804-085-5
196

science as a major based on their enjoyment in using
computers either for schoolwork or for gaming.
Unfortunately, neither of these uses of computers
truly shows the logical and mathematical nature of
the discipline. Many students who enroll in the
introductory course sequence become disillusioned
early when they discover the exacting nature of
programming.
In this paper, we discuss the problem areas we
addressed by incorporating robotics in this section,
then the approaches used to address those problems
in Section 3. Finally, we talk about work under
development and future directions in Section 4.
2.1 Computer Science Awareness
During the 2008-2009 academic year, as part of its
investigation into recruiting efforts, our department
hired consultants to conduct focus group studies of
local high school attendees. We were particularly
interested in finding out whether the cultural
perception of computer science majors by young
women was “geeky”, and if that was the reason we
were attracting few women into the program. The
results of the focus group work were eye-opening. It
was discovered that young women in local high
schools had not even heard of computer science as a
potential career path, and therefore had not even
considered it as a major.
Additional contact with high schools outside of
our local area uncovered that very few schools
included any kind of computer science in their
curricula, and those that did had very few students
enrolled. In one of the more populous cities in the
state, three of the high schools (jointly comprising a
population of several thousand students) joined
together to offer a single course in computer
science. The class had an enrollment of 7 students.
This is not an indictment of high schools – with
education funding tight, the schools must focus on
the core subjects: English, mathematics, and basic
sciences. Computer science was a subject that very
few could afford.
This revelation posed a challenge in our
recruiting efforts. If students are unaware of
computer science as a potential path of study, how
could we make our program attractive to them?
McKee and Maunders found an interesting
approach to allowing the public to interact with
robotics using the internet [1]. They used a
customized environment in which students
developed a web-interactive application where a
user controls a toy “digger” to find, move to, and
pick up a ball within the environment. Their work
supports one of the approaches we are taking to
making robotics control available online as an
approach to raising awareness of the computer
science discipline in general, although we are
approaching it using the Aldebaran NAO H25 robot
and off-the-shelf technology.
2.2 Recruiting
Assuming we could make more potential students
aware of computer science as a major, the next
question becomes “how can we attract them into our
program?” What about our program makes the
curriculum more engaging than another computer
science program? And how can we spread the word
about our program?
2.3 Engaging and Educating
Many of the students who enter our program are
mathematically underprepared for the computer
science curriculum. In fact, this is not unique to the
computer science program, but applies to other
engineering disciplines as well. Montana Tech has
addressed this problem on a campus-wide scale by
designing the Fundamentals of Engineering and
Science Program (FESP). FESP is designed to bring
the math and science skills up to a college entry
level for all students who are under-prepared. At its
inception, however, the program focused on
mainstream engineering students, and didn’t meet
the needs of our computer science and software
engineering students.
The FESP students who wished to enter
computer science or software engineering programs
don’t meet the prerequisites to enter our standard
degree track program, and in the past have been
routed into programming courses that don’t count
toward their degree, but ones that would hopefully
give them a taste of computer science; courses such
as Applications Programming (Visual Basic) or
Matlab Programming. While these courses do teach
programming, they don’t teach the more disciplined
and mathematical science of computation.
The lack of preparation, and potentially the lack
of understanding of computer science as a discipline
led us to ask “how could we engage these students,
teach them the basics of logical thinking, and
introduce them to programming at the same time?”
In younger students (ages 9-15), Pucher and
Hofmann found that the “one-way, top-down
transmission of knowledge” is not as effective as a
more hands-on project-based approach, particularly
in stimulating and motivating learners to solve real-
world problems [2]. In their research, the Lego
Mindstorm NXT and its graphical programming
environment were used to engage students in
learning and problem solving. Although our
Recent Researches in Circuits, Systems, Multimedia and Automatic Control
ISBN: 978-1-61804-085-5
197

students are older once they enter the university
system, we believe that some of the same
approaches can be used to engage them.
Hsu and Weng also used the Lego Mindstorm
NXT robots to teach robotic control through
programming using the Java language [3]. They
found that first introducing programming through
the graphical programming environment and then
through annotated Java code enabled students to
understand the programming process more easily.
Again, we feel that this approach is more applicable
to our entry students who are under-prepared for the
computer science curriculum.
2.4 Advanced Education and Research:
Addressing Retention
Once students have fully entered our program, the
next question becomes “how do we continue to
engage students, and provide a state of the art
education that will be applicable to either graduate
study or the career market?” That is, how do we
ensure that we retain those students who are
prepared and interested in a career in computer
science?
Ray et.al. discuss engaging online learners by
considering learning style [4]. They discuss that
learning styles of computer students tends to be
active, sensing, visual, and sequential, as opposed to
reflective, intuitive, verbal and global, and they use
these findings in structuring their online computer
science course to engage students. We do not have
an online program, but the fact remains that
computer science students are more engaged when
we approach teaching by considering appropriate
learning styles.
Jehlicka discusses the interdisciplinary nature of
programming [5], and used Lego Mindstorm NXT
robots to illustrate and teach students about physical
phenomena such as the light spectrum and light
combination. While we use the robotics to teach
programming skills, it is a nice side-effect that our
students are learning about the physics of robotic
motion and control also.
3 Problem Solution
The introduction of robotics into our department
curricula was not the initial approach tried, nor is it
the only method we are using to address the issues
discussed. It is, however, one of the more visible
aspects of our approach, and the literature has
shown that robotics is an effective approach to
stimulating, motivating and engaging learners, a
result we wish to achieve in our program.
In each of the identified problem areas, we have
incorporated robotics technology in different ways,
with different intents. We discuss each of our uses
of robotics in the following sections.
3.1 Raising Computer Science Awareness
with Robotics
In order to increase the awareness of computer
science to potential students, we have incorporated
the use of robotics both as demonstration tools and
for hands-on experience.
4-H Club is an international organization for
students ages 9-19, and its mission is to provide
leadership, citizenship and life skills to its members.
Local grade school and junior high school 4-H
members started using our Lego Mindstorm NXT
robots last year. The department provides
mentorship to assist these students with their
projects through our own computer science students.
The department has also sponsored workshops
for high school students using the Mindstorm
robots. In this case, a project was devised in which
students were to build one of three types of robots,
program their robot using the graphical
programming environment available with the kits,
and then have their robots play a competitive game.
The Mindstorm robots are also used as
demonstration tools when potential students visit
our department.
Finally, with the acquisition of the Aldebaran
NAO H25 robot, we are developing a way for
potential students to interact with and control the
robot via the internet. This project is discussed more
fully in section 3.4
Since we are working with younger age learners,
results of our efforts will not be known until these
potential students reach university age.
3.2 Robotics for Recruiting
Many of the same strategies for raising computer
science awareness are expected to act as a venue for
recruiting also. Again, results of our efforts may not
be conclusive for a few years, when learners enter
the university system.
3.3 Robotics for Engaging and Educating
In working with students who are not prepared for
the computer science curriculum, we added a class
that would be targeted toward those who had an
interest in learning programming.
Following the lead of Jeannette Wing of
Carnegie Mellon University (CMU) [6], we
introduced a course called Computational Thinking
designed to teach problem solving skills to
Recent Researches in Circuits, Systems, Multimedia and Automatic Control
ISBN: 978-1-61804-085-5
198

underprepared students entering the curriculum.
Diverging from the CMU approach, we also
incorporated simple robotics. The course is taught as
two hours of lecture and three hours of lab per
week. The lecture portion stresses logical and
mathematical thinking and problem solving. The lab
portion uses the Lego Mindstorm NXT robotics kits
to illustrate some of the concepts taught during
lecture. The robotics approach was chosen rather
than general programming because of the interactive
and very visual demonstration of results that
students get. Students in the lab must build their
own robots and program them, first using the
graphical programming environment provided with
the kits, and then moving on to programming them
in Java (using LeJos, a Java library for the NXT
robots). We have offered this course for three years
now, and have been able to gather concrete
measurements on its success.
To test the effectiveness of incorporating the
Computational Thinking course and robotics into
the pre-curriculum, the Whimbey Analytical Skills
Inventory (WASI) pre-test and post-test were used
in each of the three years of offering the course [7].
On the first day of class, the WASI pre-test was
administered. It consists of 37 problems that test
mathematical and logical thinking in various forms.
During the semester, instruction is given which first
addresses the types of problems the WASI test
covers, and then delves into deeper logical concepts
such as algorithm development, using structured
logic, invariants, and recursion. Students are not
informed they will be taking the WASI post-test; in
fact, they are told that the final exam will cover
class material. On the last day of classes, the post-
WASI was administered. This is a second version of
the WASI test, also consisting of 37 problems of
similar composition as the pre-test.
Comparison of WASI pre-test and post-tests was
done over the three years of administering the
course. In the results, scores were included only if a
student took both of the tests. (Not all students are
present for the first day of classes, nor do all
students attend class, particularly if they are
unaware that an exam will be administered.) An
additional pair of scores was discarded because the
student misunderstood the instructions and
answered none of the questions on the pre-test,
resulting in a 33 point gain between pre- and post-
test scores. Including this data point would have
skewed the data quite favorably, but would be an
inaccurate measure of the true success.
Initial (pre-WASI) test scores ranged from a low
of 12 to a high of 28, while post-test scores ranged
from a low of 19 to a high of 37. The average point
improvement over the course of a semester was
6.92. Running the student’s t-test (single-tailed
matched), the improvement between pre-and post-
test scores turned out to be highly significant, with
p<0.01 (p=0.0000267).
The significant difference between performance
on the WASI pre- and post-tests indicates that the
course, with its use of robotics, impacts the problem
solving ability of these students. Over time, we
intend to track the effectiveness of the course long-
term, that is, to see how students who took the
course perform in our courses as opposed to those
who were similarly under-prepared, but didn’t take
the course.
3.4 Robotics for Advanced Education and
Research
Once students understand computer science as a
discipline, and they have entered our degree track
program, we wish to continue to engage them and
make their learning experience as relevant as
possible. To this end, we have used the Lego
Mindstorm NXT robots in some course work and as
independent research tools. However, the
capabilities of the Mindstorm robots reach a limit.
To this end, we purchased a NAO H25 robot for
student research use. We hope to eventually grow
our robotic capability to be able to enter the
RoboCup competition as has [2].
Figure 1: NAO H25 Robot
NAO is a humanoid robot manufactured by
Aldebaran Robotics. NAO is roughly 58 cm tall, has
25 degrees of freedom, a variety of sensors (vision,
tactile, auditory, sonar, etc.), and is capable of
network connectivity via Bluetooth, 802.11, and
infrared [8], [9]; all of which make NAO an
Recent Researches in Circuits, Systems, Multimedia and Automatic Control
ISBN: 978-1-61804-085-5
199

extremely versatile robot, and capable of engaging
students at a much more technical level than the
Mindstorm robots.
NAO H25 is the standard humanoid platform for
the RoboCup competition, and has been used in
many other areas, including voice recognition, audio
and visual signal treatment, and trajectory
optimization. Aldebaran Robotics has been
encouraging NAO development in academic
settings as well. Their premise is, which agrees with
the literature, and complements our approach, is that
robotics can be an engaging approach to stimulate
student interest in computer science and similar
fields of study.
NAO provides a platform for both beginners and
advanced users. Choreographe is a graphical
programming environment that allows users to
control the robot with little programming
experience. Additionally, Aldebaran includes a
software bundle and documentation for their
“NaoQi” middleware pre-installed on NAO. NAO
may be programmed with a wide variety of
languages including C++, C, Python, Java, and Urbi.
Figure 2: NAO and NXT “Dog”:
The dog has a touch sensor that is depressed if the
harness pulls, providing a signal to the dog to stop
moving.
3.4.1 NAOCam
The NAO robot is a fairly large investment, and
it is unlikely that potential students would have
access to such technology. The idea for the
NAOCam project came about as an opportunity to
allow remote programming and control of NAO via
the web. The project therefore contributes to our
goal of increasing awareness to the public about
computer science, while at the same time engaging
our upper level undergraduates in developing the
NAOCam system.
The design of the system is as follows. The
server host will set up a web service. The server
host will need to have software to allow a
connection from the web server to NAO. With
connectivity, then, a remote client is able to control
NAO from anywhere on the Internet, given the web
server information.
Since operation will be remote, visual feedback
to the client is necessary. This will be accomplished
using a webcam feed. The remote client can
subscribe to the webcam feed, connect to the web
server, and begin testing either existing code on the
server, or their own code.
The web server will have the Aldebaran software
available for clients to use, including NaoQi and the
SDK for Python or C++. Python, in particular, is a
language that has been used in teaching early
programming.
The system is so far in the design stage, and
there are two issues that still need to be resolved.
Since there is only one physical NAO robot,
scalability of this system is an problem. Only one
script/program can be run at a time. However, the
webcam feed will allow as many viewers as the web
server permits, so multiple clients can still observe
NAO simultaneously. Depending on usage, it may
be necessary to allocate time slots to remote clients.
Ideally, additional NAO robots could be purchased.
The second issue with the system is the security
and safety of the NAO robot. It is not guaranteed
that the remote client will program or control NAO
in a safe manner. One possible solution is that
NAOCam would be made available with pre-
programmed code blocks for users to select and
combine, but which have been tested for safety with
NAO.
4 Conclusions and Future Work
Robotics is being used in the Montana Tech
Computer Science Department in a variety of ways:
to increase awareness of the field, to assist us in
recruiting students, to engage our existing students,
and to provide an advanced research and
development experience for our upper level
students. Some of our efforts have shown significant
effect, particularly with the use of robotics in our
Computational Thinking course. Other efforts will
take time before results are expected. And finally,
our NAOCam project is still in its design phase.
Thus far, though, our results appear promising, and
we are optimistic that as additional work is
completed, results will also be positive.
Recent Researches in Circuits, Systems, Multimedia and Automatic Control
ISBN: 978-1-61804-085-5
200

There are several directions into which we wish
to take the robotics work. The first is that of fully
developing the NAOCam system for public access.
A second direction is that of working with the
Montana Tech system to make the Computational
Thinking course count toward general education
requirements. This would provide two benefits. One
is that more students (not necessarily only computer
science students) would be willing to participate in
the class if it counted toward graduation, and thus
more students would gain awareness of computer
science as a discipline. Another benefit is that our
own majors who are not yet prepared for the
mathematics in the computer science curriculum
will have a course that teaches the science of the
discipline, and not just the programming piece.
Finally, we intend to program and use the NAO
robot as a recruiting tool along with the Mindstorm
robots. To this end we would like to make NAO as
interactive and human-friendly as possible.
References:
[1] G. McKee and R. Maunders, “Exploiting Toys
and the Internet for Robotics Education”,
Proceedings of the 2001 WSEAS/IEEE
MultiConference on Modern Information
Technologies and Robotics, 2001, Malta.
[2] R. Pucher and A. Hofmann, “Teaching with
Projects Increases the Intrinsic Motivation to
Learn – The RoboCup Project”, Proceedings
of the 10th WSEAS International Conference
on Computational Intelligence, Man-Machine
Systems and Cybernetics (CIMMACS '11),
2011, Jakarta, Indonesia, pp. 179-182.
[3] M. Hsu and T.-S. Weng, “Assisted Instruction
Case Design of Robot Creative Assembly and
Control Program Design”, Proceedings of the
7th WSEAS/IASME International Conference
on Educational Technologies (EDUTE '11),
2011, Iasi, Romania, pp. 106-111.
[4] S. Ray, A. Denton, C. Beseman, and K.
Nygard, “Learning Theory and Styles in Online
Computer Science Courses”, 2004 WSEAS
Workshop: Engineering Education, Venice,
Italy, November 2004.
[5] V. Jehlicka, “Interdisciplinary relations in
teaching of programming”, Proceedings of the
2010 WSEAS Conference on Applied
Computing, 2010, Timosaora, Romania, pp. 33-
38.
[6] J.M. Wing, “Computational Thinking and
Thinking About Computing,” Philosophical
Transactions of the Royal Society, vol. 366,
July 2008, pp. 3717-3725.
[7] A. Whimbey and J. Lochhead, Problem Solving
and Comprehension, 6th ed., Psychology Press,
Taylor and Francis Group, New York and
London, 1999.
[8] Aldebaran Robotics, http://www.aldebaran-
robotics.com/en/Pressroom/About/NAO.html,
2012
[9] Aldebaran Robotics,
http://developer.aldebaran-robotics.com/nao/,
2012.
[10] R. Ramli, M. Yunus, and N. Ishak, “Experience
of Robotic Teaching for Malaysian Gifted
Enrichment Program at PERMAT Apintar”,
Proceedings of the 9th WSEAS International
Conference on Education and Educational
Technology (EDU '10), 2010, Iwate, Japan, pp.
163-166.
Recent Researches in Circuits, Systems, Multimedia and Automatic Control
ISBN: 978-1-61804-085-5
201