See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/305716879

Manipulation Tasks in Robotics Education

Article · December 2016

DOI: 10.1016/j.ifacol.2016.07.147

CITATIONS READS

2 161

7 authors, including:

Oleg I. Borisov Alexey Vedyakov

ITMO University ITMO University
49 PUBLICATIONS   126 CITATIONS    37 PUBLICATIONS   101 CITATIONS   

SEE PROFILE SEE PROFILE

Igor Petranevsky Alexey Bobtsov

ITMO University ITMO University
6 PUBLICATIONS   19 CITATIONS    222 PUBLICATIONS   1,525 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Magnetic bearing View project

This paper is supported by Government of Russian Federation (GOSZADANIE 2014/190 (project 2118), grant 074-U01) and the Ministry of Education and Science of
Russian Federation (project 14.Z50.31.0031) View project

All content following this page was uploaded by Igor Petranevsky on 19 January 2018.

The user has requested enhancement of the downloaded file.

11th IFAC Symposium on Advances in Control Education
11th
11th IFAC
June IFAC Symposium
Symposium
1-3, 2016. on
on Advances
Slovakia in
Advances
Bratislava, in Control
Control Education
Education
11th
June IFAC Symposium on Advances in Control Education
June 1-3,
1-3, 2016.
2016. Bratislava,
Bratislava, Slovakia
Slovakia Available online at www.sciencedirect.com
June 1-3, 2016. Bratislava, Slovakia

ScienceDirect
IFAC-PapersOnLine 49-6 (2016) 022–027
Manipulation Tasks in Robotics Education 
Manipulation Tasks Robotics Education 
Manipulation Tasks in Robotics Education 
in
Oleg I. Borisov ∗∗ Vladislav S. Gromov ∗∗ Anton A. Pyrkin ∗∗
Oleg
Oleg I.
I. Borisov
Borisov ∗ Vladislav S.
∗ Vladislav
VedyakovS. ∗ Gromov
Gromov ∗ Anton A. Pyrkin
∗ Anton A. Pyrkin ∗ ∗
Oleg I. Alexey
BorisovA.
Alexey A. Vladislav
Vedyakov S.∗ Igor
Gromov
∗ Igor
V. Petranevsky
V. Anton A. Pyrkin
Petranevsky ∗
∗∗
∗
Alexey
Alexey
Alexey A.
A.
A. Vedyakov
Bobtsov
Vedyakov
∗,∗∗ Igor V. Petranevsky
∗ Vladimir I. Salikhov∗
Igor V. Petranevsky
∗,∗∗
Alexey
Alexey A. Bobtsov ∗,∗∗ Vladimir I. Salikhov ∗∗∗
Alexey A. A. Bobtsov
Bobtsov ∗,∗∗ VladimirVladimir I. I. Salikhov
Salikhov
∗
∗ Department of Control Systems and Informatics, ITMO University,
∗ Department of Control Systems and Informatics, ITMO University,
∗ Department of
Department of Control
Kronverkskiy
Kronverkskiy Control pr.Systems
pr. Systems
49, St.
and
and Informatics,
49, St. Petersburg, 197101,
Informatics,
Petersburg, 197101,
ITMO
ITMO Russia
Russia
University,
University,
∗∗
Laboratory Kronverkskiy
“Control
Kronverkskiy pr.
of
pr. 49,
49, St.
Complex
St. Petersburg,
Systems”,
Petersburg, 197101,
Institute
197101, Russia
for
RussiaProblems of
∗∗
∗∗ Laboratory
∗∗ Laboratory
Mechanical“Control“Control
Engineering, of Complex
of Complex Systems”,
BolshoySystems”, Institute
pr. 61, St.Institute
Petersburg, for Problems
for Problems
199178, of of
Laboratory “Control
Mechanical Engineering, of Complex
Engineering, Bolshoy Bolshoy Systems”,
pr. 61, Institute
St. Petersburg, for Problems
199178, of
Mechanical
Mechanical Engineering, Bolshoy Russia pr.
pr. 61,
61, St.
St. Petersburg,
Petersburg, 199178,
199178,
Russia
Russia
E-mail: [email protected], Russia [email protected],
E-mail:
E-mail: [email protected],
[email protected],
[email protected] [email protected],
[email protected],
E-mail: [email protected],
[email protected] [email protected],
[email protected]
[email protected]
Abstract: In this paper a series of practical tasks on control for manipulators is presented. This
Abstract:
Abstract: In
In thisa paper a series of practical tasks on controlMethods
for manipulators is presented. This
series represents
Abstract:
series In this
represents thisa paper
paper a
a series
laboratory
laboratory seriespart
part
of
of practical
practical
of the
tasks
of the course on
on control
“Control
tasks “Control
course
for
for manipulators
controlMethods for Robotic
manipulators
for Robotic
is
is presented.
Applications”
presented.
Applications”
This
This
series
included
series represents
represents a
in the curriculum
a laboratory
laboratory of thepart
part of
master’s
of the
the course
degree
course “Control
program
“Control Methods
atMethods
the Departmentfor
for Robotic Applications”
of Control
Robotic Systems
Applications”
included
included in
in the curriculum of the master’s The degree program at thetoDepartment
attract and of Control Systems
and Informatics
included
and in the
Informatics the curriculum
of ITMO
of
of the
of ITMO University.
curriculum the master’s
University.master’sThe
degree
exercises
degree
exercises
program
are aimed
program
are
at
at the
aimed theto
Department
Department
attract and
of
of Control
motivate
motivate
Systems
Controlstudents,
Systems
students,
and
improve
and Informatics
their
Informatics of
of ITMO
practical
ITMO University.
competency
University. andThe
The exercises
simplify
exercises are
their
are aimed to
understanding
aimed to attract
attractof and motivate
theoretical
and motivate students,
background.
students,
improve
improve their
theirarepractical
focused competency
practical competency and
and simplify their understanding of theoretical background.
The tasks
improve
The tasks theirare practical
focused
on such problems
competency
on such and simplify
problems
as path
simplify
as
their understanding
and
theirand
path understanding
trajectory
of
of theoretical
trajectory planning, theoretical
planning,
background.
automatic
background.
automatic
code
code
The
The tasks
generation
tasks are
and
are focused
design ofon
focused on such
a tracking
such problems
system as
problems as path
for path
moving and
and trajectory
target.
trajectory planning,
planning, automatic
automatic code
code
generation and
generation and designdesign of of aa tracking
tracking systemsystem forfor moving
moving target.
target.
generation and design of a tracking system for moving target.
© 2016, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Keywords: Control education, Control applications, Robotics, Manipulation tasks, Robotic
Keywords:
Keywords:
manipulators. Control
Control education,
education, ControlControl applications,
applications, Robotics,
Robotics, Manipulation
Manipulation tasks, tasks, Robotic
Robotic
Keywords:
manipulators. Control education, Control applications, Robotics, Manipulation tasks, Robotic
manipulators.
manipulators.
1. INTRODUCTION teach students principles and practical concepts dynamics
1. INTRODUCTION
1. INTRODUCTION teach
teach students
students of principles
principles and
and practical
network.concepts dynamics
1. INTRODUCTION and control
teach students
and control
control of
manufacturing
principles
manufacturing and practical
practical
network.
concepts
concepts
The et
dynamics
The Mechatronic
dynamics
Mechatronic
and
Control Kit
During the course of any applied science it is extremely and control of manufacturing network. The Mechatronic of
in manufacturing
Spong et al. network.
(2001); The
Bobtsov Mechatronic
al. (2009)
During the course of any applied science it is extremely Control Kit in Spong et al. (2001); Bobtsov et al. (2009)
During
important
During the course
tocourse
theto develop of any
both
of both applied
theoretical
any applied science
scienceand it it is
practical expe- Control Kit in Spong et al. (2001); Bobtsov et al. (2009)
extremely
is extremely
Control
is Kit
considered inas Spong et
configurable al. (2001);
applicationBobtsovfor et al.
testing vari-
(2009)
important
important to develop
develop both theoretical
theoretical and
and practical
practical expe-
expe- is
is
ous considered
considered
modeling, as
as configurable
configurable
identification, application
application
nonlinear for
for
control testing
testing
approaches.vari-
vari-
rience of
important students.
to develop Understanding
both theoretical of lecture material
and practical will
expe- is considered as configurablenonlinearapplication for testing vari-
rience
rience of
of students.
students. Understanding
Understanding of
of lecture
lecture material
material will
will ous
ous
The modeling,
modeling,
remote identification,
identification,
control framework nonlinear
for a control
control
robotic approaches.
approaches.
arm situated
be much
rience of easier
students. for them if
Understanding they haveof to conduct
lecture practical
material will ous modeling, identification, nonlinear control approaches.
be
be much easier
much easier forvarious
them if ifjust
theylearned
have to toresults.
conductThus, practical The
The remote
remote control
control framework
of framework
Technischefor for a robotic
aa robotic arm
arm situated
situated
exercises
be testingfor
much easier them
forvarious
them ifjust they have
theylearned conduct
have toresults.
conductThus, practical
the at
practical The the laboratory
remote control framework Universität
for robotic forarmeducation
situated
exercises
exercises testing
testing various just learned results. Thus, the
the at
at
and the
the laboratory
laboratory
research of
of
purposes Technische
Technische
is presented Universität
Universität
in for
for
Ruderman education
education
et al.
experimental
exercises testing equipment
various isjusthighly
learneduseful for educational
results. Thus, the at the laboratory of Technische Universität for education
experimental equipment is highly
highly useful for educational
educational and
and research
research purposes
purposes is
is presented
presented in
in Ruderman
Ruderman et
et al.
experimental
and research
experimental equipment
control
equipment purposes.is
is highlyRecentlyuseful
useful for
mobile robots
for educational and (2012).
and
(2012).
In
research
In
this
this
work
purposes
work
a
a
number
is presented
number
of
of
projects carried
in Ruderman
projects carried et al.
out
out
by
al.
by
and
and research
research
quadcopters control
control
(Pyrkin purposes.
purposes.
et al. Recently
Recently
(2014a,b); mobile
mobile
Bazylev et robots
robots
al. and
and
(2014)) (2012).
students
(2012). In
In this
using
this work
the
work a number
framework
a number areof
of projects
described.
projects carried
carried out
out by
by
and research (Pyrkin
quadcopters control purposes.
et al. Recently
(2014a,b); mobile
Bazylev et robots
al. and students using the framework are described.
(2014))
quadcopters (Pyrkin
have become(Pyrkin et al. (2014a,b);
popularetinal.conducting Bazylev et al.
various experimental (2014)) students
students using
using the
the framework
framework are
are described.
described.
quadcopters
have become popular (2014a,b);
in conducting
conducting Bazylev et al.
various experimental (2014))
experimental In this paper a practical part devoted to control of ma-
have
have become
studies. They popular
become popular in
can be constructed
in conducting various
by students or purchased In
various experimental In this
this paper aa practical part devoted to control of ma-
studies.
studies. They
They
ready-made. can
can
For be
be constructed
constructed
example, by
by
Lego Mindstormstudents
students or
or
NXT can be In
purchased
purchased this paper
nipulators
nipulators paperof the
of a practical
the practical
course
part
part devoted
course “Control
“Control devoted to
to control
Methods
Methods for
of
of ma-
for Robotic
control ma-
Robotic
studies. They
ready-made. can
For be constructed
example, Lego by students
Mindstorm or
NXT purchased
can be nipulators
Applications”,
nipulators of
of the
which
the course
is
course “Control
being
“Controltaught Methods
at
Methodsthe for Robotic
Department
for Robotic
ready-made.
used to construct
ready-made. For
For example,
simple Lego
example, mobileMindstorm
Lego robots equipped
Mindstorm NXT
NXT can can be
with
be Applications”,
Applications”, which
which is
is being taught at the Department
used
used to construct
to
actuators, construct
sensors and simple
simpleencoders mobile
mobile robots
(seerobots
Bobtsov equipped
equipped with
with
et al. (2011);
of
of
Control Systems
Applications”,
Control Systemswhichand is being
and being taught
Informatics
taughtof
Informatics of
at the
the Department
atITMO
ITMO
University,
Department
University,
used to
actuators, construct
sensors simple
and encoders
encoders mobile (see robots
Bobtsov equipped
etIn with
al.Bazylev of
is
of Control
presented.
Control Systems
It
Systems and
consists
and Informatics
of three
Informatics
(2011); is presented. It consists of three exercises assumed of
exercises
of ITMO
ITMO University,
assumed to be
University,
actuators,
Kolyubin et
actuators, sensors and
al. (2012);
sensors and Pyrkin et
encoders (see
al.Bobtsov
(see (2013a)).et
Bobtsov et al.
al. (2011);
(2011); is presented. It consists of three exercises assumed to
to be
Kolyubin
Kolyubin
and Pyrkin et
et al.
al. (2012);
(2012);
(2013) Pyrkin
the Pyrkin
Bioloidet et al.
al.
kit (2013a)).
(2013a)).
is used toIn In Bazylev
Bazylev
construct
carried
is
carried
out
presented.
out
by
byIt students
consists
students
at
of
at
the
three
the
laboratory.
exercises
laboratory.
They
assumed
They to be
concern
concernbe
Kolyubin
and Pyrkinet al. (2012);
(2013) the Pyrkin
Bioloid et al.
kit (2013a)).
is used to In Bazylev
construct carried
such
carried out
robotic
out by
by students
problems
students asat
at the
path
the laboratory.
and They
trajectory
laboratory. They concern
planning,
concern
and
and Pyrkin
Pyrkin (2013)
an experimental (2013) setupthe
the forBioloid
for
Bioloidtesting kit
kitofofis used
used to
control
is control construct
approaches
construct such
toapproaches such robotic
robotic problems
problems as
as path identification
and trajectory planning,
an
an experimental
for experimental
biped robots.setup setup for
In Bazylev testing
testing
et al.of control
(2014) approaches
models of a such
forward
forward
and inverse
robotic
and problems
inverse as path
kinematics,
kinematics,
and
and trajectory
path identification
trajectory of
planning,
of sinusoidal
planning,
sinusoidal
an
for experimental
biped robots. setup
In for
Bazylev testing
et al.of control
(2014) approaches
models of a forward
signal
forward and
and inverse
parameters,
inverse kinematics,
tracking
kinematics, system identification
design,
identification etc.of
of sinusoidal
sinusoidal
for
for biped
quadcopter robots.
and
biped robots. In
2-DOF Bazylev
In Bazylevrobotic et al.
arm (2014)
are models
investigated of
and a
a signal parameters, tracking system design, etc.
quadcopter
quadcopter and 2-DOF
and
combined together. 2-DOF roboticetarm
robotic
al. are
arm (2014)
are
models of
investigated
investigated and
and
signal
signal
The paper
parameters,
parameters,
is organized
tracking
tracking system
system design,
as follows. design,
The brief
etc.
etc.overview of
quadcopter and 2-DOF robotic arm are investigated and The paper is organized as follows. The brief overview
combined
combined together.
together. The paper
the course
paper and is organized
used equipmentas follows. The
is presentedbrief
The briefin Section of
overview of
2.
combined together.
Use of various applications to motivate students, pro- The the
the course
course
Sections 3–5
is organized
and
and are used
used
as follows.
equipment
equipment
devoted to is
is presented
presented
description of in
in
overview
the Section
Section
practical
of
2.
2.
Use
Use
moteofof various
various
control applications
applications
science and improve to
to motivate
motivate students,
practicalstudents,
experience pro-
pro- the
is Sections course 3–5and are used equipment
devoted to is presented
description of in
the Section
practical 2.
Use
mote of various
control applications
science and to
improve motivate
practical students,
experience pro-is Sections
exercises
Sections 3–5
on
3–5 are
control
are devoted
of to
manipulators.
devoted to description
descriptionThe of the
reactive
of the practical
response
practical
mote control
quite common science
between and improve
scientists andpractical
professors experience
across the is
is exercises on control of manipulators.
mote
quite
quite
control science
common between and improveand
scientists practical
professors experience
across the exercises
of on
on control
the students
exercises control of
passed the roboticThe
of manipulators.
manipulators. The
The
reactive
reactive
course
reactive
response
response
is given
response in
world.common
quite In Astrom
common between
betweenet al. scientists
(2005) and
scientists the professors
and problem of
professors across
across the
bicycle
the of
of the
the
Section students
students
6. The passed
passed
conclusion the
the
of robotic
robotic
the paper course
course
and is
is
furthergiven
givenworkin
in
world.
world.
dynamics In
In Astrom
Astrom
modeling et
et al.
al.
is addressed. (2005)
(2005) the
the
The problem
problem
presented of
of bicycle
bicycle
results of
are directionthe
Section 6. students
6.are
The passed
conclusion the
of therobotic
the paper
paper course is
and further given
further workworkin
world.
dynamics In Astrom
modeling et
is al.
addressed.(2005) the
The problem
presented of bicycle
results are Section
Section 6. The
The conclusion
presented
conclusion in of
Section
of the 7.
paper and
and further work
dynamics modeling
used in education
dynamics modeling to isattract
addressed. The
students. presented
At results
results are
the department of direction are presented in Section 7.
used
used in education
in education
Mechanical to isattract
to
Engineering
addressed.
attract The At
students.
students.
of Eindhoven
presented
At the department
the department
University
are
of Tech-of direction
of direction are are presented
presented in in Section
Section 7. 7.
used in education
Mechanical to
Engineering attract of students.
Eindhoven At the department
University of of
Tech-
Mechanical Engineering
nology the liquid-based of Eindhoven
emulator University
of manufacturing of Tech-
pro- 2. COURSE OVERVIEW
Mechanical Engineering of Eindhoven University of Tech- 2. COURSE
2. COURSE OVERVIEW
nology
nology thedesigned
the
cesses was
nology the
liquid-based
liquid-based
liquid-based
emulator
emulator
(see details
emulator
of manufacturing
of
in Starkov
of
manufacturing
et al. (2012))
manufacturing
pro-
pro-
to
pro- 2. COURSE OVERVIEW OVERVIEW
cesses
cesses was
was designed
designed (see
(see details
details in
in Starkov
Starkov et
et al.
al. (2012))
(2012)) to
to
cesses was designed (see details in Starkov et al. (2012)) to The education course “Control Methods
The education course “Control
Methods
for Robotic Ap-
for
 This paper is supported by Government of Russian Federation The education
plications”
The education course
is included
course “Control Methods
in the curriculum
“Control Methods of Robotic
for
for Robotic
the master’s
Robotic
Ap-
Ap-
Ap-

 This paper is supported by Government of Russian Federation plications”
plications” is
is included
included in
in the
the curriculum
curriculum of
of the
the master’s
master’s
This paper is 2014/190
(GOSZADANIE
 This paper is
supported(project
supported
by Government
by 2118)).
Government
of Russian Federation
of Russian Federation degree program
plications” is at the
included Department
in the of
curriculumControlof Systems
the and
master’s
(GOSZADANIE
(GOSZADANIE 2014/190 2014/190 (project
(project 2118)). degree
degree program at the Department of Control Systems and
(GOSZADANIE 2014/190 (project 2118)).
2118)).
degree program at the Department of Control Systems and
program at the Department of Control Systems and
Copyright
2405-8963 ©© 2016,
2016 IFAC 22 Hosting by Elsevier Ltd. All rights reserved.
IFAC (International Federation of Automatic Control)
Copyright
Copyright
Peer review© 2016
©under IFAC
2016 responsibility
IFAC of International Federation of 22
22 Control.
Automatic
Copyright © 2016 IFAC 22
10.1016/j.ifacol.2016.07.147
 2016 IFAC ACE
June 1-3, 2016. Bratislava, Slovakia Oleg I. Borisov et al. / IFAC-PapersOnLine 49-6 (2016) 022–027 23

3. TRAJECTORY PLANNING USING SPLINE

FUNCTIONS

The first exercise is devoted to a common trajectory

planning problem (see Fu et al. (1988); Spong et al.
(2006)). The robot used in this task is Kawasaki FS06N
(see Fig. 1). The input data for students is the following
• Denavit-Hartenberg parameters of the robot;
• Cartesian coordinates of the four reference points
(they are assumed to be starting, departing, arriving
and final ones);
• time moments assigned to the given reference points.
The entire trajectory from starting point to the final one
is separated by the given reference points into three parts.
Students are supposed to solve the inverse kinematic task
for the each reference point specified by Cartesian coordi-
Fig. 1. Students at work nates. Time should be parametrized and converted to the
relative time τi assigned to the each part of the trajectory
Informatics of ITMO University. Students learn this course (i is a number of the trajectory part, for this case it looks
for one semester. This academic discipline requires skills as i = {1, 2, 3}) such that 0 ≤ τi ≤ 1. Then it needs to take
and knowledge relevant to automation control, design of into account the constraints related with exact values of all
digital systems, programming and computer networks. the generalized coordinates (obtained from the solution of
Students are supposed to study all these areas during the inverse kinematics), velocities and accelerations of the each
passed bachelor program. joint corresponding to the reference points with respect to
the relative time τi . Then the such well-known approach
This course is comprised of theoretical and practical parts. as the spline interpolation is used to calculate all inter-
The first one is traditionally represented by a series of mediate values of the generalized coordinates between the
lectures on advanced control for robots. The second part given reference points. So, changing of all the joint relative
provides involvement of students in work with the robotic positions are described by low degree polynomials (spline
equipment at the laboratory of the department. Carry- functions) with unknown coefficients. Taking their first
ing out tasks they use various mobile and manipulating
robots. There are even a series of the exercises devoted
to controller design for a robotic boat described in Wang
et al. (2015b,a). In this paper three scenarios of laboratory
tasks on control design for manipulators are demonstrated.
There are various robots used in these exercises:
• Kawasaki FS06N;
• Mitsubishi MELFA RV-3SDB;
• KUKA youBot.
The exercises presented in the paper are focused on path
and trajectory planning problems, automatic code gener-
ation for a manipulator and design of tracking system for
moving target.
Kawasaki FS06N is a standard 6-DOF industrial robot
equipped with the SCHUNK 3-finger gripping hand (see
Fig. 1). Mitsubishi MELFA RV-3SDB is an also 6-DOF
manipulator, which is a part of the Festo Robot Vision Cell
(see Fig 2). The latter represents the educational setup
for simulation of such industrial operations as welding,
assembly, sorting, etc. KUKA YouBot is a mobile platform
with four omni-directional wheels, 5-DOF robotic arm
and 2-finger gripper (see Fig. 5). These robots allow to
practically verify theoretical approaches learned at the
course “Control Methods for Robotic Applications”. Feed-
back used in these robots can be complexed with various
external sensors. So, a regular USB camera is used below
to design tracking system using the KUKA youBot. The
latter also allows to use ROS (Robot Operating System;
see its overview in Quigley et al. (2009)) to provide, for
example, a such important for mobile robots function as Fig. 2. The Festo Robot Vision Cell equipped with the
SLAM (Simultaneous Localization and Mapping). robot Mitsubishi MELFA RV-3SDB

23
2016 IFAC ACE
24
June 1-3, 2016. Bratislava, Slovakia Oleg I. Borisov et al. / IFAC-PapersOnLine 49-6 (2016) 022–027

Fig. 3. The experiment of drawing a hypotrochoid. From left to right: the input image, arc approximation, reference
points defining, picture drawn by the robot

Fig. 4. The experiment of drawing a portrait of Alexander Pushkin. From left to right: the input image, arc
approximation, reference points defining, picture drawn by the robot

and second derivatives the polynomials for velocity and ac- Next, there is a kind of challenge for students at the
celeration respectively can be obtained. The last step is to third step. The derived coordinates can be already used
form the matrix equation including all the polynomials and to program the robot using point-to-point movement. But
defined constraints and to calculate unknown coefficients the size of such input sequence is significant. Since the
of the polynomials. As result, the polynomials for the each robot recalculates its position and configuration at each
joint describing its generalized coordinate, velocity and achieved reference point, the velocity in this case will be
acceleration with known coefficients are obtained and can substantially decreased. In addition, if the reference path
be used to program the robot to move through the given is considerably sophisticated, the size of the code may even
points. exceed capabilities of the robot controller.

4. PATH PLANNING BASED ON BITMAP IMAGE The solution of this issue can be found by students in arc
approximation using software functionality of movement
The second practical task is focused on programming of along an arc, specified with only three points. This makes
complex flat motion paths, which can be specified by the robot recalculate its parameters only three times
an input image file. The robot Mitsubishi RV-3SDB is instead of multiple times depending on how many points
used to conduct this scenario (see Fig. 2). The purpose is make up this arc. Such solution allows to increase the
automatic code generation to move the end-effector along velocity of movement and decrease the code size. The task,
some counters represented on the input bitmap image. The which students have to solve, is to separate such sequences
algorithm is comprised of the three steps: edge detection of points, that make up corresponding arcs with some δ-
(in case of a colorful image), tracing, arc approximation. region.

If the input image is colorful, the first step is edge In accordance to a set of the algorithms described above
detection. It is carried out by the algorithm described in the robot can draw pictures with a pen grasped by its
Canny (1986). As result, a binarized image of the original end-effector given the corresponding input image files. The
one should be obtained. On the other hand, the desired experiments of drawing a hypotrochoid and portrait of
contours can already be specified using black and white Alexander Pushkin are conducted. The results with all the
colors (for example, this can be done manually). steps are shown in Fig. 3 and Fig. 4.

On the second step it needs to form an array with all coor- 5. TRACKING SYSTEM FOR MOVING TARGET
dinates of the neighbor black points of the binarized image.
It is carried out by the Moore-Neighbor tracing algorithm Design of adaptive systems with delays is one of the
(see Terrier (1999); Weisstein (2005)). The result is the scientific interests at the Department of Control Systems
sequence of the point coordinates. and Informatics (Pyrkin et al. (2010, 2013b)). A number

24
 2016 IFAC ACE
June 1-3, 2016. Bratislava, Slovakia Oleg I. Borisov et al. / IFAC-PapersOnLine 49-6 (2016) 022–027 25

is represented by the desired rotation angle of the robot

link. The server tracking system can be implemented in
a graphical programming software environment for conve-
nience of design and debugging. The control commands
from the server controller model to the program launched
at the robot are transmitted via TCP/IP.
The problem of target tracking is the main task of this
lesson. The control system must ensure alignment and
retention of the observation line between the center of
the tracking device and the moving target with some
precision. The tracking device in this task is comprised of
two SISO subsystems. Their outputs represent an elevation
and azimuth of the observation line and must tend to
the elevation and azimuth of the target, respectively. The
control signals are based on the errors between the current
and real values of these angles.
A feedback in this system is provided with a USB web-
camera. It is placed on the last link of the robotic arm.
The image of the target displayed on the liquid-crystal
panel is taken as the bright moving circle (the equipment
is shown in Fig. 5).
Students can use various ways of image processing in order
to get the coordinates of the target. The most common
way is selection of the object basing on its color. The
image is binarized with some specified brightness values
for the each RGB channel to extract the object from
the background. The total binary mask is obtained by
logical multiplication between masks for the each channel.
Further, the coordinates of the center of the selected
image segment are determined with respect to the upper
left corner of the frame. The X-axis within the frame
is considered to be directed to the right and the Y -
axis is directed downward. This algorithm could be easily
implemented in various programming languages. Other
approaches suggested by students are welcome as well.
Fig. 5. KUKA youBot used for the tracking system
A distance to the target and direction of its movement
are unmeasurable. Thus, the problem can be formulated
of modern approaches were given to students during the
as stabilization of the moving object center at the center
lecture courses. This theory forms a basis for conducting
of the camera frame. In this case, the error signals are
the practical exercises of application of learned results to
differences between the coordinates of the frame and
robotic systems.
object centers along X- and Y -axes, respectively, in pixels
The purpose of this task is to design a tracking control of the camera frame. Since tracking is performed biaxially,
system for the robotic manipulator KUKA youBot. Work the two last links of the robot are used to orientate the
with the equipment is assumed to be carried out at the camera. The last link rotates around the vertical axis of
lab of the department. Students usually work in pairs due the local coordinate system, while the penultimate link
to limited number of the robots. The control system must rotates around the horizontal one.
run on the two computers simultaneously using the client-
Consider the problem of an unknown signal tracking for a
server model. The client part is launched on the internal
linear system with an input delay
computer of the robot. It collects data from the sensors,
sends it to the server, executes control commands given
from the server. The server part runs on the separate per- ẋ(t) = Ax(t) + Bu(t − h), (1)
sonal computer and receives data from the client, processes y(t) = Cx(t), (2)
it through the control system and sends the corresponding e(t) = g(t) − y(t), (3)
control signals to the client. Connection between them
is provided via Ethernet channel. In accordance to this
where x ∈ Rn is the state vector, u ∈ R is the input
scheme students are able to work on the separate comput-
ers simultaneously and solve distinct tasks in cooperation signal, h is the known constant delay, y ∈ R is the output
to achieve the formulated goal. of the system, g ∈ R is the reference signal as the desired
output of the system, e ∈ R is the error of reference signal
The client control system can be coded by the C language tracking, An×n is the state matrix, Bn×1 is the matrix of
using the motor control libraries provided by the robot the control inputs, C1×n is the matrix of the output. For
manufacturer. The reference signal for the each actuator the control signal u(t − h) = 0 holds for t < h.

25
2016 IFAC ACE
26
June 1-3, 2016. Bratislava, Slovakia Oleg I. Borisov et al. / IFAC-PapersOnLine 49-6 (2016) 022–027

20 where the coefficients

κuj = µ̂uj cos(ω̂jg h) − ν̂ju sin(ω̂jg h), (8)
ζju = µ̂uj sin(ω̂jg h) + ν̂ju cos(ω̂jg h). (9)
Y (t) = g2 (t), pixels

10
During the testing experiments of the adaptive controller
students can see that a large value of the time delay
0 does not degrade performance of the system. Increasing
a number of sinusoids in the system also does not notice-
ably affect performance of the proposed control approach.
−10 Performance of the PD controller heavily decreases at the
same conditions and makes it inappropriate for use.

6. DISCUSSION
−20
−30 −20 −10 0 10 20 30 The feedback from the students passed this practical
X(t) = g1 (t), pixels course is discussed in this section. Most of the learners
we canvassed highlight positive effects of these exercises.
Fig. 6. A trajectory of the target
On the one hand, interest of students is stimulated by this
Changing of the elevation and azimuth of the moving course. The experimental study is noted to be significantly
object can be approximately described by sinusoidal func- more challenging and entertaining than just computer
tions. This is a rather strong restriction on applicability of simulation of an abstract mathematical model. This aspect
the such tracking system. On the other hand, the problems allows to attract students to the course of robotics, which
of tracking and compensation are solved uniformly. It is a new, developing and leading technical trend in control
means that the control system can be used to stabilize science.
surveillance systems located on moving objects, including
vessels, for which this type of the disturbance is common. On the other hand, accumulation of experience and prac-
tical skills is provided in the course. As the students are
In general the biased multisinusoidal signal represents the able to operate real industrial robots within the tasks, they
reference signal σ g feel themselves confident of their future engineer activity.
l2 At the same time, the theoretical material of the lectures
g(t) = σ g + µgj sin(ωjg t) + νjg cos(ωjg t), (4) becomes more intelligible for students.
j=1
Thus, this practical course devoted to control design for
consisting of l2 harmonics with unknown amplitudes µgj robots is aimed to provide students with interesting and
and νjg , and frequencies ωjg , j = 1, l2 is the harmonic useful educational activities in the department laboratory,
number. attract their attention to the young applied science, help
them to learn practical skills and assimilate the contents
For this exercise the reference signals are chosen in the of the lectures.
form π 
g1 (t) = −11 cos t + 14 cos(t), (5) 7. CONCLUSION
π5 
g2 (t) = −18 cos t . (6) The preparation of highly qualified personnel of any ap-
5 plied science is inconceivable without development of prac-
where g1 (t) is a coordinate of the target on the horizontal tical skills. It is obvious, that they should be cultivated
axis, g2 (t) is a coordinate on the vertical axis. This on a par with theoretical knowledge. Thus, the student
trajectory is shown in Fig. 6. exercises on control of manipulators aimed to improve
At the first step students have to design and tune the their competency are demonstrated in this paper. They
PD controller, which is a common solution of such tasks. represent a practical part of the course “Control Meth-
The control system should include a block of the simulated ods for Robotic Applications”, which is being taught at
input delay. It is easy to note during the experiment that the Department of Control Systems and Informatics of
efficiency of the PD controller falls dramatically with the ITMO University. Students of the department appreciate
increased input delay. It is the reason of necessity to design opportunity to see how various robotic systems really
the sophisticated control system. work, how controller parameters affect plant behavior, how
this behavior can be improved, etc. Besides the practi-
Next, students can apply the method based on identifica- cal exercises, students include some robotic applications
tion of target motion parameters. Following the results of into their bachelor and master theses, since it looks more
Pyrkin et al. (2014c) use the control u(t) in the form demonstrative than just simulation. Design of new tasks
l2
 and their adaptation to the laboratory equipment can be
u(t) = σ̂ + u
µ̂uj sin(ω̂jg (t + h)) + ν̂ju cos(ω̂jg (t + h)) highlighted as the direction of further work.
j=1
l2
 REFERENCES
u
= σ̂ + κuj sin(ω̂jg t) + ζju cos(ω̂jg t), (7) Astrom, K.J., Klein, R.E., and Lennartsson, A. (2005).
j=1 Bicycle dynamics and control: adapted bicycles for edu-

26
 2016 IFAC ACE
June 1-3, 2016. Bratislava, Slovakia Oleg I. Borisov et al. / IFAC-PapersOnLine 49-6 (2016) 022–027 27

cation and research. Control Systems Magazine, IEEE, Automatic Control, 12110–12115.
25(4), 26–47. Quigley, M., Conley, K., Gerkey, B., Faust, J., Foote, T.,
Bazylev, D. and Pyrkin, A. (2013). Stabilization of biped Leibs, J., Wheeler, R., and Ng, A.Y. (2009). Ros: an
robot standing on nonstationary plane. In Methods open-source robot operating system. In ICRA workshop
and Models in Automation and Robotics (MMAR), 2013 on open source software, volume 3, 5.
18th International Conference on, 459–463. IEEE. Ruderman, M., Hoffmann, F., and Bertram, T. (2012). A
Bazylev, D., Zimenko, K., Margun, A., Bobtsov, A., matlab-based framework for the remote control of a 6dof
and Kremlev, A. (2014). Adaptive control system for robotic arm for education and research in control theory.
quadrotor equiped with robotic arm. In Methods and In Advances in Control Education, volume 9, 366–371.
Models in Automation and Robotics (MMAR), 2014 Spong, M.W., Block, D.J., and Åström, K.J. (2001). The
19th International Conference On, 705–710. IEEE. mechatronics control kit for education and research.
Bobtsov, A., Pyrkin, A., Kolyubin, S., Chepinskiy, S., In Control Applications, 2001.(CCA’01). Proceedings of
Shavetov, S., Kapitanyuk, Y., Kapitonov, A., Titov, the 2001 IEEE International Conference on, 105–110.
A., Surov, M., and Bardov, V. (2011). Using of lego IEEE.
mindstorms nxt technology for teaching of basics of Spong, M.W., Hutchinson, S., and Vidyasagar, M. (2006).
adaptive control theory. In Actes 18th IFAC World Robot modeling and control. John Wiley & Sons.
Congress, Milan (Italie), 9818–9823. Starkov, K.K., Kamp, H., Pogromsky, A.Y., and Adan, I.J.
Bobtsov, A.A., Pyrkin, A.A., and Kolyubin, S.A. (2009). (2012). Design and implementation of a water-based
Adaptive stabilization of a reaction wheel pendulum on emulator of manufacturing processes. In Advances in
moving lego platform. In Control Applications,(CCA) Control Education, volume 9, 396–401.
& Intelligent Control,(ISIC), 2009 IEEE, 1218–1223. Terrier, V. (1999). Two-dimensional cellular automata
IEEE. recognizer. Theoretical computer science, 218(2), 325–
Canny, J. (1986). A computational approach to edge 346.
detection. Pattern Analysis and Machine Intelligence, Wang, J., Borisov, O., Gromov, V., Pyrkin, A., and
IEEE Transactions on, (6), 679–698. Bobtsov, A. (2015a). Adaptive controller implementa-
Fu, K.S., Gonzalez, R., and Lee, C.G. (1988). Robotics: tion for surface robotic vessel. In Control Conference
Control Sensing. Vis. Tata McGraw-Hill Education. (CCC), 2015 34th Chinese, 3230–3235. IEEE.
Kolyubin, S., Bobtsov, A., Pyrkin, A., Kapitonov, A., Wang, J., Pyrkin, A.A., Bobtsov, A.A., Borisov, O.I., Gro-
Shavetov, S., Khovanskiy, A., Krasnov, A., Vlasov, mov, V.S., Kolyubin, S.A., and Vlasov, S.M. (2015b).
S., and Feskov, A. (2012). Lego mindstorms nxt for Output control algorithms of dynamic positioning
students’ research projects in control field. In Advances and disturbance rejection for robotic vessel. IFAC-
in Control Education, volume 9, 102–106. PapersOnLine, 48(11), 295–300.
Pyrkin, A., Bobtsov, A., Kolyubin, S., Borisov, O., Gro- Weisstein, E.W. (2005). Moore neighborhood.
mov, V., and Aranovskiy, S. (2014a). Output controller From MathWorld–A Wolfram Web Resource.
for quadcopters with wind disturbance cancellation. In http://mathworld. wolfram. com/MooreNeighborhood.
Control Applications (CCA), 2014 IEEE Conference on, html.
166–170. IEEE.
Pyrkin, A., Bobtsov, A., Kolyubin, S., Borisov, O., Gro-
mov, V.S., et al. (2014b). Output controller for quad-
copters based on mathematical model decomposition. In
Control and Automation (MED), 2014 22nd Mediter-
ranean Conference of, 1281–1286. IEEE.
Pyrkin, A., Bobtsov, A., Kolyubin, S., Surov, M., Shave-
tov, S., Borisov, O., and Gromov, V. (2013a). Simple
output stabilization approach for robotic systems. In
Manufacturing Modelling, Management, and Control,
volume 7, 1873–1878.
Pyrkin, A., Bobtsov, A., Kolyubin, S., Vedyakov, A.,
Borisov, O., Gromov, V., Margun, A., and Bazylev, D.
(2013b). Fast compensation of unknown multiharmonic
disturbance for nonlinear plant with input delay. In
Adaptation and Learning in Control and Signal Process-
ing, volume 11, 546–551.
Pyrkin, A., Smyshlyaev, A., Bekiaris-Liberis, N., and
Krstic, M. (2010). Rejection of sinusoidal disturbance of
unknown frequency for linear system with input delay.
In American Control Conference (ACC), 2010, 5688–
5693. IEEE.
Pyrkin, A.A., Bobtsov, A.A., Nikiforov, V.O., Vedyakov,
A.A., Kolyubin, S.A., and Borisov, O.I. (2014c). Out-
put control approach for delayed linear systems with
adaptive rejection of multiharmonic disturbance. In The
19th World Congress of the International Federation of

27
View publication stats