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

net/publication/303718978

3D Visualization of Mechatronic Systems using OpenGL

Article · September 2015

CITATIONS READS

0 3,857

3 authors, including:

Jakub Čerkala Anna Jadlovska

Develogics Technical University of Košice
42 PUBLICATIONS 2,188 CITATIONS 50 PUBLICATIONS 278 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jakub Čerkala on 01 June 2016.

The user has requested enhancement of the downloaded file.

 3D Visualization of Mechatronic
Sytems using OpenGL
Jakub ČERKALA, Anna JADLOVSKÁ, Štefan SZŰRÖS

Department of Cybernetics and Artificial Intelligence, Faculty of Electrical

Engineering and Informatics, Technical University of Kosice, Slovak Republic

[email protected], [email protected], [email protected]

Abstract — The main goal of this article is to show how to visualize dynamical processes in 3D
environment on the example of nonlinear mathematical pendulum and mobile robot with
differential chassis. Using Matlab or Simulink, by the solution of the model’s differential equations,
which describe its movement properties, it is possible to obtain data for visualization. The article
presents the necessary steps to create an own C# Forms application that is based on OpenGL for
visualization of dynamical system.

Keywords — nonlinear mathematical pendulum, mobile robot, Matlab/Simulink, visualization, 3D

virtual space

I. INTRODUCTION
The computer technology is evolving fast nowadays and it helps us to simplify our daily life.
This development includes also the computer graphics technology, which is extensively used in
various fields, such as entertainment or science. In entertainment, the computer graphics is used
to create video games or the special effects in movies, but in fields of science, it is mainly used
to visualize data. Depending on data, the visualization can by static (image, chart) or dynamic
(animation) and it can be done in 2D or 3D space. This article focuses on visualization of
dynamical systems presented on the example of mathematical pendulum and mobile robot with
differential chassis in three dimensional space. The first step in visualization is the selection of
data source. If we don’t have data measured on real system, we can calculate an approximation
data using mathematical model of selected dynamic system based on physical laws and
mathematics. Moreover, we need 3D models that are used for the data visualization as an
animation. There are several approaches to visualize data, for example by using programs like
Matlab/Simulink [1], [2], which offers a complete solution for dynamical visualization. With
Matlab/Simulink, we can calculate visualization data using simulation models of the chosen
dynamical systems and with the Simulink 3D Animation toolbox [3] we can create complete
visualization with 3D models. There is always an option to create a specialized application
which can visualize dynamical systems that is using a graphical library in chosen programming
language. In this case, the data can be obtained by using external computation in programs like
Matlab/Simulink or by an internal calculation, for example using the Runge – Kutta 4th order
method. This approach requires more programming but it gives the programmer more freedom.

II. APPLICATION FOR VISUALIZING DYNAMICAL PHYSICAL PROCESSES

The presented application that enables user to visualize various mathematical pendulums at
once in 3D space or a mobile robot moving in plane is a result of a bachelor thesis [4]. To
program this application, the OpenTK [5], a graphical library which enables the usage of
OpenGL [6] in C# programming language was used. Application can visualize data obtained by
external calculation, the simulation models of mathematical pendulum and mobile robot were
created in Matlab/Simulink. In case of pendulum, there is an additional option to use an internal
real-time calculation during visualization programed in a C# based on the 4th order Runge –
Kutta numerical method. In this application’s 3D environment it is necessary to implement a
camera through which the visualization can be observed. Last but not least, the 3D models
which are used to data representation have to be programmed as well or imported form an
external editor. The first main task is to obtain mathematical models of systems that will be
visualized.

351
 Electrical Engineering and Informatics VI
Proceedings of the Faculty of Electrical Engineering and Informatics of the Technical University of Košice

Fig. 1 Visualization of mathematical pendulum and mobile robot in created application

III. MATHEMATICAL MODELING OF NONLINEAR MATHEMATICAL PENDULUM

Simple or Mathematical pendulum is a bob with mass m which is attached to a fixed pivot point
by a rigid weightless rod of length l as shown in Fig. 2. The bob of the pendulum can swing
around the pivot point along a circular arc which radius is l. Friction and air resistance are
neglected. To calculate the motion of mathematical pendulum which is the angular deflection of
the pendulum, we need to solve the motion equation of pendulum by using Newton’s second
law [7], [8].

Fig. 2 The mathematical model of the mathematical pendulum

The sum of the forces which are acting on the object is equal with the mass of the object
multiplied by the acceleration of the object. The mathematical equation of this law can be
written as
F  ma. (1)
To solve the mathematical pendulum equation of motion we need to calculate the sum of forces
F which are acting on the bob and the bob’s acceleration a. As shown on Fig. 2, the only force
that is acting on the pendulum’s bob is the gravitational force Fg. Because the bob is moving
along the circular arc, we can divide the gravitational force Fg into two components. The first
component of the gravitational force Fcos is parallel to the rod of the pendulum and it doesn’t
affect the pendulum movement. The second component of the gravitational force Fsin is
perpendicular to the pendulum’s rod so it affects the pendulum’s movement [7] and therefore it
can be assumed as F in (1). To calculate the bob’s acceleration a, the second derivation of the
arc’s length s on which the pendulum’s bob moves has to be used. The arc’s length s can be
calculated by multiplication of the pendulum l length with an angular lean , the angular
acceleration of the bob is obtained as [7]
d 2 t  a
 . (2)
dt 2 l
If we assume that Fsin  mg sin  t  , then we can use the Newton’s second law (1) to get the
following 2nd order differential equations for motion of the mathematical pendulum [8]
d 2 t  g
 sin  t   0 . (3)
dt 2 l

ISBN 978-80-553-2178-3 c 2015 FEI TUKE 352

 Electrical Engineering and Informatics VI
Proceedings of the Faculty of Electrical Engineering and Informatics of the Technical University of Košice

The force acting on the pendulum have always the opposite direction as the angular lean of the
pendulum, that’s why the pendulum is always returning to its vertical position [7]. In case of
damped pendulum, the damping force can be included as [9]

Fd  k t  , (4)
where k is the damping constant and is the pendulum’s angular velocity that is equal to the
first time derivation of the pendulum’s angular lean [7]. The 2nd order motion differential
equation of the mathematical pendulum with damping force can be written as two 1st order
differential equation for easier numerical implementation as
d t 
  t  , (5)
dt
d t  g
  sin  t   k t  (6)
dt l

IV. MATHEMATICAL MODELING OF MOBILE ROBOT WITH DIFFERENTIAL CHASSIS

A rather different model to visualize is the mobile robot with differential chassis [10]. The
mobile robots have two differentially driven wheels ant it moves in plane with defined in 0GSS.
The robot’s position can be represented using its Centre of Gravity (CoG) coordinates x(t), y(t)
and orientation angle φ(t). Also, for better visualization it is required to include wheel angular
positions θR(t), θL(t) to achieve the wheel rotation illusion. The position of the robot with wheel
radius r and wheel to CoG distance b is depicted on Fig. 3.

Fig. 3 The position definition of mobile robot in global plane

The mathematical model of the mobile robot usually consists of the kinematic model, that
converts robot’s velocities into position and the dynamic model, which represents the influences
of mass, inertia or friction on resulting velocities. In this case, we use only the kinematics
model defined for angular velocities  R (t ) ,  L (t ) that can be written as [10]
   r  cos  (t ) r  cos  (t ) 
 x (t )   
   2 2
  r  sin  (t ) r  sin  (t ) 
 y (t )   
d  2 2   R (t ) 
  r r   . (7)
dt 
 (t )      L (t ) 
   2b 2b 
 (t )   1 0 
 R   
 L (t )   0 1 

V. OBTAINING THE DATA FOR VISUALIZATION

For both models, the pendulum or the mobile robot, it is possible to calculate data for
visualization using Matlab/Simulink using implemented simulation models. To get data in
programing and simulation environment Matlab, we have to create a function that implement
mathematical model of pendulum (5), (6) and using the ode45 solver function, which uses the
4th Runge-Kutta numerical method in Matlab we can calculate the data for visualization. The

ISBN 978-80-553-2178-3 c 2015 FEI TUKE 353

 Electrical Engineering and Informatics VI
Proceedings of the Faculty of Electrical Engineering and Informatics of the Technical University of Košice

simulation result of the mathematical pendulum without damping force is shown on Fig. 4 and
the Fig. 5 shows the obtained data of the mathematical pendulum with damping force.

Fig. 4 The pendulum’s angular deflection without damping force simulated in Matlab

Fig. 5 The pendulum’s angular deflection with damping force simulated in Matlab

The mathematical model of the pendulum (6) can be implemented also as Simulink model. The
simulation results are the same as in previous approach, however the Simulink environment
enables user to change the parameters or simulation properties with minimum input. The
obtained plots that describe the pendulum motion are clear.
Simulation model of mobile robot with differential chassis implemented in Simulink based
on its kinematics (7) can be tested for non-equal input wheel velocities ωR ≠ ωL and the
simulation experiment shown on Fig. 6 presents the movement results.

Fig. 6 Mobile robot moving on circular trajectory

The movement of the robot is usually shown in plane. But there is no clear information about
the velocities of the robot, its orientation or number of preformed loops. It is unpractical to
independently plot all five position variables [x(t), y(t), φ(t), θR(t), θL(t)]T in term of time,
however the visualization can show them at once in user-friendly form.

VI. VISUALIZATION
To be able to visualize the movement of the models, it is necessary to create a 3D environment
that can import and interpret the data. The key features of the visualization are the camera,
model and the animation loop.

ISBN 978-80-553-2178-3 c 2015 FEI TUKE 354

 Electrical Engineering and Informatics VI
Proceedings of the Faculty of Electrical Engineering and Informatics of the Technical University of Košice

1) Camera
The cameras in a three-dimensional computer graphics play an important role. They are used to
observe the visualization from a freely chosen position and they are used for the projection of
the scene. The projection output image of from the perspective of the camera view. In presented
application, a basic camera was created. This camera has unrestricted movement in 3D
environment and can observe the 3D environment from every perspective.

Fig. 7 The movement possibilities of the created camera

The possibilities of camera movement are depicted on Fig. 7, the flying camera have
unrestricted movement in the x, y axis and can rotate on axis y and x. When rotating around the
axis z, the only restriction is applied - the camera cannot turn more than 90 degrees up and
down. These restrictions of the rotation around the axis z are important to prevent the camera to
flip upside down.

2) 3D models
3D models are used to show the data during visualization. Three-dimensional models are
created by polygons which are defined by vertices. The combination of polygons will create a
network of polygons, a mesh, which can contain different types of polygons, but the most
common polygon is a triangle - the triangles can be used to create any other existing types of
polygons. For more complex models it is always better to work with more polygons to achieve
a better model accuracy.

Fig. 8 The process of creating a 3D model with polygons

For visualization of the mathematical pendulum, we had to create a 3D model of the pendulum.
The 3D model of a mathematical pendulum consists of a static mount, rod and weight. The rod
connects the static mount and weight of the pendulum. The pendulum’s rod is dynamic - its size
is changed according to the length of the pendulum used in simulation. The rod and the weight
are connected and they can move around a pivot point (see Fig. 9). This model was created by
primitives like a cylinder and circle. The 3D models of mobile robot consist of box that
represents the robot’s base and two cylinders as wheels.

Fig. 9 3D model of the mathematical pendulum and mobile robot

Both 3D models are very simple approximations of real systems, because they are created
programmatically, but they serve only as demonstration models. To improve the model quality,
the best approach is to implement the model loader.

ISBN 978-80-553-2178-3 c 2015 FEI TUKE 355

 Electrical Engineering and Informatics VI
Proceedings of the Faculty of Electrical Engineering and Informatics of the Technical University of Košice

3) Animation of dynamic process

To create an animation of dynamic process we have to calculate projection images and display
them continuously with a time delay, because this creates an illusion of movement. We can
create these images easily by displaying the calculated data in the actual moments of time. Data
for visualization were calculated by a fixed time step. To get a smooth visualization, the step
time, used also as delay between two images has to be chosen reasonably - not too wide, but not
too small. To animate the 3D models it is necessary to move or rotate them or their parts
according to data.
In case of pendulum, the 3D model only moving parts are the rod and the weight. We can
visualize the mathematical pendulum movement if we rotate these parts around the pendulum’s
pivot point (see Fig. 10). The angle of rotation depends on the calculated lean. To visualize the
robot with rotating wheels, it is necessary to use rotation and translation transformations in
specific order – firstly rotate wheels, than rotate the whole robot.

Fig. 10 Movement of the pendulum’s 3D model

Fig. 11 Movement of the mobile robot’s 3D model

VII. CONCLUSION
In this article, we have presented the example how to create a 3D dynamical visualization of the
mathematical pendulum with or without dampening and the mobile robot. With obtained
mathematical models, the simulation models were programmed to get simulation data in
Matlab/Simulink environment. An application with 3D environment based on OpenGL with
free look flying camera was created to visualize obtained data by 3D model of the pendulum.

ACKNOWLEDGMENT
This work has been supported by the Research and Development Operational Program for
project: University Science Park Technicom for innovative applications with knowledge
technology support, ITMS code 26220220182, co-financed by the ERDF (70%) and by grant
KEGA - 001TUKE-4/2015 (30%).

REFERENCES

[1] Matlab – official wepage [online]. [cit. 2015-06-05]. <http://www.mathworks.com/products/matlab/>

[2] Simulink – official webpage [online]. [cit. 2015-06-05]. <http://www.mathworks.com/products/simulink/>
[3] Simulink 3D animation toolbox – official webpage [online]. [cit. 2015-06-05].
<http://www.mathworks.com/products/3d-animation/>
[4] Štefan Szűrös, Návrh a programovanie 3D prostredia na báze OpenGL v jazyku C# - Bachelor thesis (Supervisor:
doc. Ing. Anna Jadlovská, PhD.). Košice. FEI, 2015. 56 s.
[5] OpenTK - official webpage [online]. [cit. 2015-06-05].< http://www.opentk.com/>
[6] OpenGL – official webpage [online]. [cit. 2015-06-05]. <https://www.opengl.org/>
[7] Edward R. Scheinerman, Invitation to dynamical systems, Prentice Hall College Div, 1995. ISBN 0-13-185000-8
[8] Richard H. Enns, George C. McGuire, Nonlinear Physics with Mathematica for Scientists and Engineers. Boston:
Birkhäuser, 2001. ISBN 0-8176-4223-4.
[9] Douglas D. Novaes, “Perturbed damped pendulum: finding periodic solutions via averaging method,” Revista
Brasileira de Ensino de Física, v. 35, n. 1, 1314 (2013), published
[10] Čerkala, J. a Jadlovská, A. Dynamics with Friction in Mobile Robot Simulink Model. In: Technical Computing
Bratislava 2014 Proceedings, pp. 65–81., Lap Lambert Academic Publishing, 2015. ISBN 978-3-659-40792-5.

ISBN 978-80-553-2178-3 c 2015 FEI TUKE 356

View publication stats