Compulsory Game based Robot Contest

for Embedded System Development Education

Harumi Watanabe Mikiko Sato Masafumi Miwa
Tokai University Tokai University Tokushima University
Tokyo, Japan Tokyo, Japan Tokushima, Japan
[email protected] [email protected] [email protected]

Makoto Imamura Shintaro Hosoai Nobuhiko Ogura

Tokai University Change Vision, Inc. Tokyo City University
Tokyo, Japan Fukui, Japan Tokyo, Japan
[email protected] shintaro.hosoai@change- [email protected]
vision.com
Hiroyuki Nakamura
Kyushu University Kenji Hisazumi
Fukuoka, Japan Kyushu University
[email protected] Fukuoka, Japan
[email protected]

ABSTRACT 1. INTRODUCTION
This paper proposes a compulsory game based robot contest Recently, various robot contests have become popular in a global
involving embedded system development lectures. Both context, because robot development is an attractive subject for
undergraduate and graduate computer science students participate many students, and is well-known for the useful educational
in this contest. For such students, the embedded system materials [1,2]. For the past fourteen years, we have also been
development is not easy to learn. Because, the development needs organizing a robot contest called Embedded System Symposium
to cover a wide range of knowledge in a variety of fields, Robot Challenge(ESS-RC) for graduate and undergraduate
including software, electronics, and control theory. Robot system engineering students.
development is an attractive subject for students and comprises
various technologies that are similar to practical embedded Additionally, it has become easier to learn embedded system
systems. Namely, robot system development is useful for learning programming because of various educational robots, such as
embedded system development. However, in single themed LEGO and robot cleaners [3-8]. Moreover, graphical
contests, we cannot evaluate the learning level of the students, and programming languages have become popular for use with
they cannot easy to understand how to construct a robot. To educational robots [5,9,10]. Because of its simplicity, robot
overcome this problem, we propose a compulsory game based programming has also been adopted in elementary schools
robot contest. The compulsory games consist of fundamental [5,9,10]. However, for students studying to become engineers,
techniques for developing robot systems and involves embedded system development is not easy to learn. Because, the
straightforward procedures for evaluating student learning level. subject needs to cover a wide range of knowledge and skills in
Additionally, we intended for students to learn how to construct a software, electronics, and control theory.
robot system by solving the compulsory games in a step-by-step To obtain these skills, students need to be highly motivated.
fashion. In this paper, we propose four compulsory games and Additionally, relevant educational materials are required to cover
evaluate the methods of this contest by examining its results. these fields. Robot system development is an attractive subject for
students and comprises various technologies that are similar to
CCS Concepts practical embedded systems. As mentioned in several related
• Social and professional topics➝ Computing education works [7,11,12,13,14,15], we believe robot system development is
• Computer systems organization➝Robotics useful for learning embedded system development.
Keywords However, in single themed robot contests, such as NHK Robocon
Embedded system development; education; robotics; project- [16], we cannot evaluate the learning level of students, and
based learning students cannot easily understand how to construct a robot. To
overcome this problem, we propose a compulsory game based
Permission to make digital or hard copies of part or all of this work for robot contest involving lectures. The compulsory games consist of
personal or classroom use is granted without fee provided that copies are fundamental techniques for developing robot systems and
not made or distributed for profit or commercial advantage and that
involved straightforward processes for evaluating student learning
copies bear this notice and the full citation on the first page. Copyrights
for third-party components of this work must be honored. For all other levels. Additionally, we intend for students to learn how to
uses, contact the Owner/Author. construct a robot system by solving the compulsory games in a
ICSCA 2018, February 8–10, 2018, Kuantan, Malaysia step-by-step fashion. In this paper, we introduce four of these
© 2018 Copyright is held by the owner/author(s).
ACM ISBN 978-1-4503-5414-1/18/02.
https://doi.org/10.1145/3185089.3185144
compulsory games and evaluate the methods of this contest by 2. ESS ROBOT CHALLENGE
examining its results. The section describes our robot contest called Embedded System
The remainder of this paper is organized as follows. Section 2 Symposium Robot Challenge (ESS-RC). Section 2.1 clarifies the
describes ESS-RC. Section 3 proposes the compulsory games. educational goals, background, and schedule. Section 2.2
Section 4 evaluates the method of the contest and section 5 illustrates the main game of the contest.
discusses whether the contest method solved the problems. Finally,
section 6 concludes this paper. 2.1 Outline
The goal for ESS-RC is to provide an education of embedded
Table 1. Spring School system development. The contest is supported by Information
(1) Smart mobile robot (2) Rover trial Processing Society Japan (IPSJ) and enPiT, which is an
competition competition educational project of Ministry of Education, Culture, Sports,
20 May Science and Technology-Japan (MEXT).
09:00-09:30 Opening and course guidance. In 2016, the enPiT project began its second season. Accordingly,
Lecture 1: Ability to discover and solve
problems with PBL learning
for ESS-RC 2017, we renewed the educational materials robot and
for improving practical and research skills. the contest theme. Until now, the main theme has been the auto-
09:30-09:50 Lecture 2: Relationship between general theory cruising airship, the auto-vacuuming robot, and the multi-copter.
of system development and the The new robot is relatively small and is equipped with a camera.
profession of embedded software. The detail illustrates in Section 2.2.
09:50-10:35 Lecture 3: Trends and characteristics of
development technologies in The educational periods of ESS-RC consist of Spring School,
embedded systems. Summer School, and supplementary lectures. During Spring
10:45-12:00 Lecture 4: Introduction to control theory. School, we provide lectures for fundamental knowledge of robot
13:00-16:30 Exercise 1: Exercise 2: developments as shown in Table 1. Namely, we teach project
Robot control exercise Robot control exercise management, as well as embedded software design, robot
for iRobot Create2. for Zumo. programming, and control theory. Figure 1 shows one scene in the
16:30-17:00 Short presentation on industry–university collaboration. SCRUM project management exercise, which is a framework for
21 May developing and sustaining complex products [17]. Through this
09:00-14:00 Exercise 3: Exercise 4: management exercise, students can divide the contest games into
System development Basic measurement and small issues and goals. As a result of this lecture and exercise, we
methodology. implementation of
expect that students will become to understand how to construct
embedded systems.
14:00-16:00 Exercise 5: Project management using Scrum.
the main game by using the compulsory games. Related works
16:00-17:00 Closing. [13,15] applied SCRUM to project management in the PBL. In
22-26 May Practicing mini PBL exercise at home universities. particular, the curriculum of [15] is similar to ours.
24 May Intermediate presentation of exercise progress. Summer School consists of the contest involving compulsory
27 May Final presentation of exercise results. games, a poster session, two keynote speeches, and a student
workshop, as shown in Table 2. In the poster session of the first
Table 2. Summer School day, students explain the features of the system and the
11 August development of the contest robot. The contest score includes the
13:20-13:30 Opening results of evaluating the poster. In the student workshop on the
13:30-15:30 Compulsory games second day, students reflect on the development. The group of the
16:00-16:30 Main game (Auto-Cruising Game) workshop is organized as mixed Universities. Because students of
17:00-17:15 Short talk for Poster all teams aim to the same goal, they develop friendly relations
17:20-20:00 Poster with one another, despite their different fields and university
12 August affiliations.
09:30-10:45 Keynote 1: Big data and AI in Era of Industrial 4.0
10:50-12:05
Keynote 2: System Control with Machine Learning and 2.2 Auto-Cruising Game
AI The auto-cruising game is the main game of ESS-RC. The final
13:00-14:50 Student Workshop (for reflecting the contest) goal of this game is to learn the fundamental technologies for
15:00-15:50 Presentation of Workshop modern embedded systems, such as IoT. In a few years, we plan
to create a subject of ESS-RC involving those technologies. This
year’s contest is the first step toward the final goal. Therefore, we
consider that a robot requires comprising some sensors and
communication functions, as follows.
Figure 2 depicts the auto-cruising game. The robot is equipped
with a camera, a three-axis acceleration sensor, a magnetic sensor,
a three-axis gyro sensor, and a line-tracing sensor. The robot
traces a route from its starting place on the map to the goal. While
running, the robot distinguishes some AR-markers and adapts its
behavior based on the markers’ instructions. In ESS-RC, the
teams compete to determine whose robot is faster and whose robot
correctly follows the instructions given by the AR-markers.

Figure 1. SCRUM Project Management

Additionally, students explain the robot’s behavior at runtime as
shown in Figure 3.

Figure 5 Compulsory Game: Line Trace

Figure 2 Main Game: Auto-Cruising Game

Form for
Robot IP

Interpreted value
of AR-marker

Map
Value of the
AR-maker

View of AR-Marker
Remote Robot
Controller for
Robot Figure 6 Compulsory game: AR-Marker
3.1 Spin Turn
This game aims to evaluate whether students identify quantities of
Figure 3 View of Robot behavior robot behavior. Figure 4 shows the game. The issue of this game
is as follows:
In a 20cm square area, a robot must rotate 180 degrees
counterclockwise by the spin turn. Continuously, the robot
rotates clockwise, counterclockwise, clockwise, and
counterclockwise. When switching direction, the robot must
stop over 3 seconds.
This game evaluates the running time and the angle of the gap
between the goal and the robot’s final stopping position. To
accomplish this issue, students need to measure the behavior of
the robot and identify the quantities required for movement.

3.2 Line Trace

This game aims to evaluate the student learning level of
fundamental control theory. Figure 5 shows the game. The issue
of this game is as follows:
Figure 4 Compulsory game: Rotate robot A robot traces by the double black lines: the course distance is
2 m, the line width is 1 cm, the width between lines is 10 cm,
3. COMPULSORY GAME and the course does not contain gaps.
This section proposes the compulsory games of ESS-RC.
Previously mentioned, the compulsory games contribute the This game evaluates running time and whether the behavior is
following issues: (1) identify learning level and (2) students learn smooth. To accomplish this issue, students adopt a fundamental
how to construct a robot system in a step-by-step fashion. Thus, control theory, such as PID. If students have never used any of the
the compulsory games are fundamental technologies for building control theories, the robot will vibrate or easily stray off the
a robot system used in the main game. In the following, we course.
illustrate the four compulsory games.
 Table 3. Result of the games (3) Maze: Two teams gained scores. The reason for this is that
other teams exceeded the time limit. During the contest, students
Team
were required to finish all games within five minutes.
Compulsory Game A B C D
Time 35 30 33
(4) AR-marker: All teams failed. After the contest, we found that
Spin Turn AR-markers could not be technically identified. Outside of the
Degree 20 25 180 contest, one team was able to identify the AR-marker as shown in
Time 42 38 20 23 Figure 3. However, this team could not distinguish the position of
Line Trace
Stray off 0 1 0 0 the AR-marker.
Time 48 47
Maze 5. DISCUSSION
Obstacles 0 1
Previously mentioned, we invoked two issues for single themed
Time contests. Firstly, we cannot evaluate the learning level of students.
AR-Marker
Distance Secondly, students cannot easily understand how to construct a
robot. In this section, we discuss whether the contest method
contributes to the issues based on the results detailed in Section 4.
3.3 Maze
This game aims to enable the consideration of algorithms. The Reflecting on the game of 14 years ago, we had not yet adopted
course field of this game is same as the main game of Figure 1. compulsory games in the contests. In those contests, we evaluated
The issue of this game is as follows: the behaviors exhibited in one game that was similarly
constructed to the main game mentioned in section 2.2. The
In the game field, there are a start point, a goal point, and some scoring method was designed around whether an airship was able
obstacles. A robot must run from the start point to the goal to reach a certain height, move to a specific position, and then
point avoiding obstacles. land in the goal position. Those contests had two significant
problems. First, judges could not identify whether students
This game evaluates running time, whether the robot reaches the
adopted theories to their systems. Even if an airship reached the
goal, and the number of times it avoids obstacles. If students
required height, the behavior could have been coincidental.
desire the robot to run speedily, they need to consider some
Secondly, the contest game, such as the main game, was too
algorithms.
complicated and expansive for students. If students did not divide
3.4 AR-Marker the game program into smaller components, they were not able to
This game aims to combine multiple knowledge. Figure 6 shows obtain scores. Indeed, several teams did not achieve any scores at
the game. The issue of this game is as follows. all.

While moving in 1 m square area, a robot must discover three As shown in Table 3, each team in our most recent contest was
AR markers and convey the position. able to gain a type of score. Regarding the first problem of
previous contests, we considered how to exclude coincidental
This game evaluates running time, the kinds of markers, and the behavior when constructing the compulsory game. For example,
distance of the gap between the marker position and the position in the spin turn game, a robot was required to continue in the same
information. To accomplish this issue, students need to combine motion five times, as mentioned in Section 3.1. Therefore, we
communication and AR-marker algorithms. considered that the scores in Table 3 were indicative of learning
level. We regarded that all teams had checked whether they
4. EVALUATION reached their own goals. Regarding the second problem,
This section illustrates the results of the compulsory games’ compulsory game programs that excluded the AR-marker game
evaluation for considering whether the compulsory game based were simple and a suitable size for learning. Thus, all teams were
contest is adequate. challenged in the compulsory game according to their level of
understanding.
Seventy-eight students participated in the Spring School program,
and their fields were computer science and control mechanics. So far, we have mentioned the advantages of the compulsory
The students were divided into 19 groups, with each group having based contest. The rest of this section describes the disadvantages
three or five students. In Summer School, thirty-eight students we must reflect on. Firstly, none of the teams were able to obtain
participated, and four teams joined the contest involving the a score in the AR-marker game. With our experience, we had
compulsory games. In Table 3, the numbers in each column considered that many students did not achieve the learning level
indicate the results of the compulsory games as followed by the of understanding the functional decomposition design. To solve
evaluation method explained in Section 3. In the following, we this problem, we created the AR-marker game. Through the
consider these results: inquiry of students, we became aware that many of them lacked
an understanding of the interruption handling. Additionally, this
(1) Spin turn: Three teams gained scores. We consider that the
game system was hardly fit for use as a compulsory game because
reason for this is that students had experience with a similar
it was highly compounded with the AR-marker, the
exercise during the Spring School period.
communication functions, and the position calculation by sensors.
(2) Line trace: All teams gained good points. Because line trace We should have divided this game into smaller issues. Thus, we
games are very popular for learning robot programming, some should add the issue of interruption handling and divide the game
students had experience with line trace games before ESS-RC, into smaller issues in the next version of ESS-RC.
while others easily obtained the necessary algorithms from
Secondly, the teams that joined the contest were small in size. As
informative websites.
mentioned in Section 4, only 38 students participated in Summer
School, while 78 students participated in Spring School. We
reflect that we may not have explicitly presented the learning [6] Barry S. Fagin and Laurence Merkle. 2002. Quantitative
roadmap for the main game. Analysis of the Effects of Robots on Introductory Computer
Science Education. J. Educ. Resour. Comput. 2, 4, Article 2
6. CONCLUSION (Dec. 2002), 18 pages.
In this paper, we proposed a compulsory game based robot contest [7] J. O. Hamblen and G. M. E. van Bekkum. 2013. An
to learn embedded system development. The compulsory games Embedded Systems Laboratory to Support Rapid Prototyping
aimed to solve the following problems. First, we cannot evaluate of Robotics and the Internet of Things. IEEE Transactions on
the learning level of students. Secondly, students cannot easily Education 56, 1 (Feb 2013), 121–128.
understand how to construct a robot. We provided the compulsory [8] J. McLurkin, J. Rykowski, M. John, Q. Kaseman, and A. J.
games for students to learn the fundamental techniques of the Lynch. 2013. Using Multi-Robot Systems for Engineering
main game that is the auto-cruising game. The compulsory games Education: Teaching and Outreach with Large Numbers of
consisted of the spin turn, the line trace, the maze, and the AR- an Advanced, Low-Cost Robot. IEEE Transactions on
marker game. We believe that the compulsory games relevantly Education 56, 1 (Feb 2013), 24–33.
contribute to these problems. However, we also discovered issues, [9] K. Y. Chin, Z. W. Hong, and Y. L. Chen. 2014. Impact of
as described Section 5. Using an Educational Robot-Based Learning System on
Students’ Motivation in Elementary Education. IEEE
In future work, we will define the skills and the achievement goals
Transactions on Learning Technologies 7, 4 (Oct 2014),
for the total education of ESS-RC. Additionally, we will clarify
333–345.
the relationship between the lectures in Spring School and the
[10] C. Martinez, M. J. Gomez, and L. Benotti. 2015. A
compulsory games in Summer School by providing a learning
Comparison of Preschool and Elementary School Children
roadmap. Furthermore, to provide the seamless learning stages,
Learning Computer Science Concepts Through a
we will add some new compulsory games and improve the current
Multilanguage Robot Programming Platform. In Proceedings
games.
of the 2015 ACM Conference on Innovation and Technology
7. ACKNOWLEDGMENTS in Computer Science Education (ITiCSE ’15). ACM, 159–
We would like to thank Editage (www.editage.jp) for English 164.
language editing and to express their gratitude to R. Ogino, H. [11] N. Correll, R. Wing, and D. Coleman. 2013. A One-Year
Ebihara, K. Mogi, and M. Kawamura of Tokai University for their Introductory Robotics Curriculum for Computer Science
technical assistance with the experiments. Upperclassmen. IEEE Transactions on Education 56, 1 (Feb
2013), 54–60.
8. REFERENCES [12] H. Hassan, C. Dominguez, J. M. Martinez, A. Perles, J. V.
[1] O. O. Ortiz, J. A. Pastor Franco, P. M. Alcover Garau, and Capella, and J. Albaladejo. 2015. A Multidisciplinary PBL
R. Herrero Martin. 2017. Innovative Mobile Robot Method: Robot Control Project in Automation and Electronic
Improving the Learning of Programming Languages in Engineering. IEEE Transactions on Education 58, 3 (Aug
Engineering Degrees. IEEE Transactions on Education 60, 2015), 167–172.
2 (May 2017), 143–148. [13] L. Pinto, R. Rosa, C. Pacheco, C. Xavier, R. Barreto, V.
Lucena, M. Caxias, and C. Maurccio. 2009. On the Use of
[2] D. Rus. 2006. Teaching robotics everywhere. IEEE SCRUM for the Management of Practical Projects in
Robotics Automation Magazine 13, 1 (March 2006), 15–94. Graduate Courses. In 2009 39th IEEE Frontiers in Education
[3] S. Akayama, S. Kuboaki, K. Hisazumi, T. Futagami, and T. Conference. 1–6.
Kitasuka. 2013. Development of a Modeling Education [14] M. C. Rodriguez-Sanchez, A. Torrado-Carvajal, J. Vaquero,
Program for Novices Using Model-driven Development. In S. Borromeo, and J. A. Hernandez-Tamames. 2016. An
Proceedings of the Workshop on Embedded and Cyber- Embedded Systems Course for Engineering Students Using
Physical Systems Education (WESE ’12). ACM, Article 4, Open-Source Platforms in Wireless Scenarios. IEEE
8 pages. Transactions on Education 59, 4 (Nov 2016), 248–254.
[15] J. Shin, A. Rusakov, and B. Meyer. 2015. Concurrent
[4] D. J. Cappelleri and N. Vitoroulis. 2013. The Robotic
Software Engineering and Robotics Education. In 2015
Decathlon: Project-Based Learning Labs and Curriculum
IEEE/ACM 37th IEEE International Conference on Software
Design for an Introductory Robotics Course. IEEE
Engineering, Vol. 2. 370–379.
Transactions on Education 56, 1 (Feb 2013), 73–81.
[16] NHK Gakusei Robocon. Retrieved December 1, 2017 from
[5] P. De Cristoforis, S. Pedre, M. Nitsche, T. Fischer, F. https://www.nhk.or.jp/robocon/gakusei/index_en.html
Pessacg, and C. Di Pietro. 2013. A Behavior-Based [17] ScrumGuides.org. 2016. Scrum Guides. (2016). Retrieved
Approach for Educational Robotics Activities. IEEE December 1, 2017 from http://www.scrumguides.org/
Transactions on Education 56, 1 (Feb 2013), 61–66.