A LabVIEW-Based Remote

Laboratory Experiments
for Control Engineering
Education
MILADIN STEFANOVIC,1 VLADIMIR CVIJETKOVIC,2 MILAN MATIJEVIC,1 VISNJA SIMIC2
1
Faculty of Mechanical Engineering, University of Kragujevac, Sestre Janjic 6, 34 000 Kragujevac, Serbia
2
Faculty of Science, University of Kragujevac, Radoja Domanovic´a 12, 34 000 Kragujevac, Serbia

Received 21 October 2008; accepted 2 February 2009

ABSTRACT: This paper deals with remote access to a real laboratory equipment using
contemporary computer and network technology for creating the environment that will enable
a remote user to perform the required laboratory exercises and control the laboratory
equipment. Architecture and characteristics of WebLab will be described with special attention
to the latest implemented laboratory experiment for control of the coupled water tanks (using
LabVIEW). This paper will also give results of researches among student population in order to
determine advantages and effects of using web laboratory in control engineering education.
ß 2009 Wiley Periodicals, Inc. Comput Appl Eng Educ; Published online in Wiley InterScience
(www.interscience.wiley.com); DOI 10.1002/cae.20334

Keywords: remote laboratory; automatic control; controller implementation; data

acquisition; e-learning

INTRODUCTION to solve the problems of tomorrow, with the tools from

yesterday.’’ The Web influences both the industry and
Much attention has been focused recently on modern education because it enables supervision and tele-
control education in Engineering. A leading idea to all operation of devices (cost reduction).
educators was given in [1]: ‘‘Educators must have an Recently, Information and Communication Tech-
open attitude towards new technologies. They should nologies (ICTs) have changed the conception of the
sensibly incorporate new technological development teaching process both in the classroom and in the
to avoid the risk of teaching the students of today, how theoretical teaching approaches. Traditional content-
oriented teaching approaches are being replaced by
student-oriented ones.
Correspondence to M. Stefanovic ([email protected]). The WorldWide Web has provided an opportunity
ß 2009 Wiley Periodicals Inc. for design and analysis of control systems through the

1
2 STEFANOVIC ET AL.

Internet. An increasing number of web-based software real laboratory equipment. This paper will also
packages have been developed to enhance the teach- present results of surveys that show effectiveness of
ing and design of control systems. Today one of the implementation of web laboratories in control engi-
most popular applications in control systems is web- neering education.
based educational environments and laboratories [2].
Different designs of web-based control labs are
presented in educational practice all over the world ARHITECTURE OF WEB LABORATORY
[3
8]. The web enables more flexible delivery of
teaching materials, distance education, new visual- The main goal and significance of the WebLab system
ization possibilities, interactivity, and cost reduction. [12] is to provide remote access to laboratory
The underlying fundamental promise of Internet- equipment and to be as universal as possible for
based laboratory approaches lies in students’ abilities various kinds of laboratory work and experiments. In
to connect to a computer-controlled laboratory setup this paper architecture, and a laboratory exercise of
of interest at anytime from anywhere, thus sharing web laboratory (WebLab) developed on the Univer-
existing limited resources in a more efficient manner sity of Kragujevac, will be presented. WebLab system
than is possible with the traditional on-site laboratory adds value to e-learning system and enables remote
approach [9]. laboratory work, which is not possible with traditional
The idea of having a remote web-based labo- e-learning systems. The basic precondition for the
ratory corresponds to attempt to overcome different experiment to be available remotely, through the
constraints and may be the next step in the remote WebLab, is the use of the programmable equipment
distance learning [10,11]. Remote web-based labo- that manages all the aspects of the experiment.
ratories may also allow researchers in different Experiments can be performed completely unattended
locations to carry out researches and design work on the side of experimental equipment, with full and
co-operatively and remotely at the same time. exclusive control from the remote user. Basic config-
The integration of the Internet into education is uration [13
16] of the WebLab system consists of the
most commonly achieved through the following web server and one or more PC acquisition servers
methodologies: that control the experiments. Even simpler config-
uration is possible with only one PC computer that
*
Developing a course website to centrally house serves both roles; the one of the web server and
various online functions and facilities course acquisition servers that control the experiment. Such
management. simple configuration is limited to cases with small
*
Creating a remote laboratory where multimedia number of experiments where one acquisition server
animation or simulations are provided to replace can control all experiments. In general, when there are
physical experiments. many experiments, more than one acquisition server is
*
Developing a web-based laboratory that enables needed. Acquisition servers are PC computers with
students to set up parameters and undertake one or more data acquisition systems (DASs) for
experiments from remote location. measurements of the physical quantities and control
of the experimental equipment. Besides DAS, acquis-
The main disadvantage of distance and e-learning ition servers can have also other kind of measurement/
approaches is absence of laboratory work. Multimedia control equipments attached, such as digital oscillo-
animations or simulations cannot bridge this gap. The scopes for very fast measurements and for laboratory
only possible solution to ensure practical work as a exercises devoted to learning how to use digital
part of concept of distance learning is implementation oscilloscopes or other kind of programmable equip-
of web laboratories that consist of remotely controlled ments. Data acquisition servers are connected to web
experiments with video feed-back. Besides, web server by computer network. Such configuration of
laboratories with remote control of experiments could the WebLab system enables acquisition servers to be
be used as a very useful educational tool in classical placed in any physical location which is provided with
or blended (mixture of classical and e-learning Internet access for connection with web server. The
approaches) learning environments. In this paper the actual architecture of the WebLab system is distrib-
architecture of web lab will be presented with uted, as it is physically situated in different buildings
experiment ‘‘Coupled water tanks’’ that was devel- of the institutions [17,18] that develop the WebLab
oped using LabVIEW. Presented solution enables system.
users to design their own controllers using a controller Users of the WebLab access the system by means
template, to upload them, and remotely control the of the web browsers as the only required tool on the
 LABVIEW-BASED REMOTE LABORATORY 3

side of the user. Such distributed configuration of the that is controlled by the DAS and digital oscilloscope.
WebLab system is presented in the Unified Modeling Names of those nodes are equal to names of the
Language (UML) node diagram in Figure 1 (see Refs. existing experiments in the WebLab system—Diodes
[19
21]). Each box in diagram presents one physical and Transistors, Analog systems, Electrical signal
piece of equipment. PC 1 is web server, while PC velocity, Steep plane, and Automatic control system.
2
PC 4 are acquisition servers that are connected The node named PC WEB browser represents
with web servers using Win Sockets. The structure of remote user of the WebLab system. Although it is only
the WebLab software is presented with components one node in the diagram, it represents all users of the
named Web user interface, Experiment and Program- WebLab system. Each experiment can be controlled
mable devices. These software components consist of by only one remote user at a time.
classes for each experiment. Web user interface All software components are implemented in
consists of classes that are used for communication Cþþ and C# programming languages using Micro-
with remote user. soft platform. Older DAS were programmed in Cþþ
Experiment component consists of classes that without .NET, while the new DAS and the web server
control each experiment in the system. These two were programmed in C# .NET. Web server was
components are physically located on the web server. implemented using Microsoft ASPX.
The third software component named Programmable
devices consists of classes that are located on
acquisition servers and that actually control program-
COUPLED WATER TANKS EXPERIMENT
mable devices—DASs and digital oscilloscope.
Nodes with names DAS and digital oscilloscope
Framework for Experiment
represent programmable equipment used in WebLab
system. Nodes on the lowest level that are connected The framework for this laboratory exercise has been
just on one side are specific experimental equipment designed using LabVIEW for creating flexible and

Figure 1 UML node diagram of the WebLab system.

4 STEFANOVIC ET AL.

scalable measurement and control applications. cds/view/p/lang/en/nid/14605). This experimental

Experiment with coupled water tanks was added as setup is common and useful in control engineering
a independent hardware/software module to the education [22].
existing WebLab structure in Figure 1. It is different
from the previous experiments (in this web labora-
tory) as it was fully developed using NI LabVIEW 8.0
Laboratory Model and Experimental
software system (http://www.ni.com/labview/). That
Apparatus
fact slightly changed the previous general structure
of the WebLab given in Figure 1. As it was developed Water level in each tank is measured by the hydro-
with LabVIEW software system, integrated Lab- static pressure using voltage measurement transducer.
VIEW web server was used on the acquisition server. Voltages on the output of transducers in each tank are
No web server programming was required with measured by DAS. Diagram with principal schematic
LabVIEW, as it directly supports web access to arrangement of the coupled water tanks is presented in
experiments. Once the experiment is developed in Figure 2.
LabVIEW on the local PC, it is immediately available Experimental setup for coupled water tanks is
through the web with simple LabVIEW setup presented in Figure 3. Control of such coupled tanks
procedure. When using LabVIEW for development system can be quite complex as the tanks are
of the remote experiments, web server and acquisition interconnected so that water from the upper tank goes
server run on the same PC. down in the corresponding lower tank and water
The role of the central web server is changed in pumps simultaneously pump water in the diagonal
that case, as it serves as the main and central web tanks.
location that directs the user to other web server As the control task for such a system of coupled
running on the acquisition server that controls the water tanks can be formulated in many different ways
selected experiment. In that way, even with the starting from the simplest version with only one pump
changed configuration, WebLab system remains fully working and controlling water level in only one
modular and compatible for different implementa- tank—corresponding lower tank—to cases where
tions with programming languages or LabVIEW both pumps are working and controlling water levels
software system. Experimental setup for coupled in more than one water tank, the goal of developing
water tanks consists of four water tanks arranged in such remote experiment is to make the universal
two levels and two water pumps controlled by DAS. environment in which any control task can be
Used DAS is NI USB 6009 (http://sine.ni.com/nips/ performed.

Figure 2 Diagram of coupled water tanks.

 LABVIEW-BASED REMOTE LABORATORY 5

array of values. For any of the mentioned periodic

functions, frequency in Hz and amplitude in V can be
set. Manual mode is used for various purposes such as
testing of the experimental setup working condition,
setting the initial water levels, identification of the
system parameters, or demonstration of the system
operation. Water level for each tank is measured by
the DAS and represented by an indicator in the form
of vertical tank where the height of the blue colored
column represents water level in the real tank. On
the right side of each tank water level indicator is a
blue control for setting the water level with white
arrow that can be moved vertically. Positions of the
white arrows determine maximum water levels in
the upper tanks, and beside that, water levels that
are to be regulated and maintained in the lower
tanks. Water levels are regulated only in the automatic
mode of operation, and not in the manual mode. UI
in Figure 4 which is Front Panel (FP) of the main
Virtual Instrument (VI) can be accessed locally from
the same PC and also remotely (Internet) from any
Figure 3 Coupled water tanks experimental setup. PC by web browser using the LabVIEW built in web
server.
In order to achieve that goal, the design of the Remote access to any VI by web is very simple to
experiment with coupled water tanks is divided into setup, and web pages for VI can be customized in
two independent tasks: standard way like any other web page.
FP of the VI is the same in the web browser
*
Development of WEBLab environment. window as it is on the local computer. In the lower part
*
Development of controller for automatic control of the web page in Figure 4 is real-time live video
of tanks system. from the web cam following the experiment. Only one
user can control the VI, while others can observe the
Development of WEBLab Environment. The first experiment from the VI and web cam live video.
task was to develop working environment in
LabVIEW software system that will manage all the Development of the Controller. The second task of
functions for measurements on the experimental setup the software support for this experiment is the design
and for controlling the water pumps. Development of of the system controller template that can be used for
the complete user interface (with following functions: implementation of any type of controller.
different ways of driving the water pumps, indication The basic role of a controller is to enable the
of current water levels in tanks, drawing water level/ maintenance of liquid level in the given tanks using
time diagrams, saving measured data and output data two pumps to fill the tanks. Different types of
in files for online overview and analysis) is also controllers could be used to control the level of fluid
included in that task. The upper part of Figure 4 in tanks. Therefore, VI for this instrument is designed
presents the user interface (UI) for coupled water in such a way that the controller is not necessarily
tanks experiment. Before the start of the experiment, fixed, but it is possible to use a special controller
duration in seconds and sampling time interval in mS implementation VI, that is, sub which is a exchange-
should be set up by writing the values in cor- able module in the VI for the control of the
responding text boxes. Experiment can be performed experiment. Template controller is designed as a
in two working modes, manual and automatic, which separate sub VI that is connected with experiment
are selected with the switch Automatic/Manual. In the main VI, FP of which is given in Figure 4, using the
manual mode, pumps 1 and 2 can be controlled by strictly specified interface. Any sub VI serving as a
selection of the continuous manual control from min controller that has the same interface can be inserted
to max with the rotating knob or one of the periodic and connected with the main VI.
functions: rectangular, sinusoidal, triangular, saw VI that represents the controller for this experi-
tooth, and arbitrary function that is defined with ment must have a standardized interface that is
6 STEFANOVIC ET AL.

Figure 4 Front panel of the experiment main VI—water tanks experiment.

connected with the rest of the system. Any VI with controller can be reduced using differences in signals.
required interface can be dynamically linked with the In this case e(n) and e(n
1) signal errors are used and
rest of the system that provides a working environ- where e(n) ¼ desired level (n)
measured levels (n)
ment for the controller. Dynamic linking is a index (i) indicates tanks and i-1, 2, 3, 4. There are also
possibility to change module controllers without used u(n
1), previous control signal j ¼ 1, 2.
restarting development and execution environment— Controller generates two signals u(n) which control
LabVIEW. Previously implemented controller in the the pumps. This means that the controller should have
form of VI with required interface loads as a VI file 4 inputs e(n), 4 inputs e(n
1), 2 inputs u(n
1), a
from FP VI using control field for the selection of total of 10 inputs and two outputs u(n).
VI entitled ‘‘Controller path’’ in the top left of Interface for connection of VI with other VI is
Figure 4. The main task of the working environment is defined by analogy in definition of prototype of
to bring all the necessary signals for the work of functions in programming languages, C, for example.
controllers, as well as to accept signals generated by Prototype should determine the exact type of data for
the controller, which are used to control pumps. In the each input and output. Prototype or interface VI is
experiment with tanks, signals that are brought to the defined as ‘‘connector VI.’’ For each VI input, there
controller represent current value of liquids in all must be the appropriate input control, and for each
tanks, then the required values to be held in all the output there should be one indicator. These controls
tanks, the previous value of the liquid from the and indicators are necessary elements of the interface,
previous measurements, as well as previous values of and if there is no need for these controls they must be
control signal. Current value of the signal is marked created for the purpose of connecting with the rest of
with N, while the previous values in that case with the system. This is developed for the liquid-level
n
1. Number of signals that are brought to the controller, where the text boxes are used for input
 LABVIEW-BASED REMOTE LABORATORY 7

Figure 5 Simple controller with VI inputs and outputs (controller is ad hoc student attempt to
make intelligent on
off controller for the correct filling of all tanks 4—the idea is to show the
simplicity of the implementation of arbitrary controller).

control, while the numeric indicators are used as Figure 7 presents the water level/time diagram for
indicators. Figure 5 depicts a BD for one version of a lower left tank which gets water from pump 1 directly,
simple controller. Icons of input controls are shown on and which level is maintained by PID controller with
the left side, while the indicators are on the right side. FP in Figure 6.
Previous value of n
1 is marked as ‘‘old.’’
Rectangles are features that can assure input of
code in syntax very similar to C programming
EDUCATIONAL GOALS AND TASKS OF
language.
LABORATORY EXERCISE
While designing controller, attention is devoted
strictly to controller, while all technical details
Educational Goals
concerning DAS, timing, experiment duration, con-
nections with experimental setup, and other specifics Laboratory experiences, which imitate the complexity
are handled by the main VI. of real life practices, are essential elements in
Figure 6 presents the FP of the sub VI that engineering education. In practical sessions, students
implements PID controller. In the upper part of learn not only by listening, like in theoretical courses,
Figure 6, above T are controls and indicators that but also through ‘‘learning-by-doing.’’ When students
serve as interface connections, while T and controls interact with the laboratory plants they have the
below serve for setting the constants for the controller. opportunity to verify what happens when they modify
Controller in Figure 6 has two independent sections and manipulate the experiment. Several studies
for each pump. As all constants for the controller for have shown that laboratory practices have achieved
pump 2 are 0, that controller is turned off. positive influence on learning skills, on well-prepared
8 STEFANOVIC ET AL.

Figure 6 Front panel for implemented PID controller.

professionals, especially in scientific and techno- laboratory equipment before doing the actual experi-
logical fields. Some authors [11] illustrate that ment, and review the experiment procedures.
students retain 25% of what they hear, 45% of what For engineers of different profiles is very
they hear and see, and 70% if they use the ‘‘learning- important to understand how system of coupled water
by-doing’’ methodology. The construction of fully tanks works, how to apply measurement techniques
furnished scientific laboratories is very expensive and appropriate hardware components, and how to
because specialized and sophisticated equipment is control the problem through the writing of control
needed. This cost is sometimes prohibited for many software. Problems of coupled water tanks, regula-
institutions. The few existing laboratory equipment tion, and flow levels are frequent in the water supply
has to be shared among researchers and students systems and sewage, biosystems (some problems of
enrolled in different programs with adjusted schedule medical physiology), as well as in many systems of
and different knowledge levels. Instructional staff processing industries. For instance, in process indus-
spends much time of every lab session describing the try (paper production industry, petrochemical indus-
equipment lab, demonstrating the correct operation of try, etc.) problems of fluid transport, storage and fluid
retention in tanks, and transfusion into the tanks are a
common case. Then, the problems of chemical
treatment or fluid mixing in the tanks are also very
frequent. In all these cases, the level of fluid in the
tanks must be controlled, as well as the flow between
the tanks. Often tanks are coupled so their levels are in
mutual interaction, and that must be mutually
controlled. Control of levels and flow into tanks is
one of the central issues of process/chemical engi-
neering systems.
Laboratory setup presented in Figure 3 allows
wide specter of experiments with various levels of
Figure 7 Water level/time diagram of PID controller. complexity.
 LABVIEW-BASED REMOTE LABORATORY 9

General goal is obvious: use one of the control (development of software, database development,
techniques to control and regulate height levels in supervision functions, acquisition, storage, presenta-
coupled tanks/or ensure that controlled levels follow tion and visualization of data etc.).
set trajectories. The general objectives of this The methodology of problem-solving approach
laboratory experiment are to ensure that students: (problem-based learning) based on experimental
apparatus in a wide range of fundamental interdisci-
(1) Get familiar with the system of feedback plinary theoretical content includes:
control—general ideas from the theory of
systems, components, and systems terminol- A. Concepts of modeling and system analysis:
ogy. *
System modeling.
(2) Gain experience in calibration of sensors, *
Identification of the system.
performance and implementation of labora- *
Linearization of the model.
tory experiment, and preparation of appro- *
Conversion of the model (presentation of the
priate software. model in complex, frequent and time
(3) Test acquired knowledge in the field of domain, etc.).
modeling and process identification. *
The study of stability and performance of
(4) Analyze static and dynamic features of other systems based on the model.
processes and systems with feedback to make B. Algorithm management in a broader sense
quality assessments and evaluate the main (with their associated functions):
parameters of the control system. *
Control of open and closed feedback.
(5) Demonstrate performance and limits of the *
PID control procedures and adjustments of
system with open feedback (manual control) PID regulators.
when there is a need for change in the set of *
P control with the associated compensators.
values or fix the effects of disturbance. *
Synthesis of controllers in the frequency
(6) Demonstrate unstable work with the system domain.
with feedback, as the result of influence of *
Feedback by the state, controller synthesis
clear time delays in the system. based on the model in space conditions and
(7) Demonstrate influence of changes of P concept of optimal control (LQR controller,
(proportional) and the dynamic and static etc.).
characteristics of the system. *
Implementation of observer or estimators of
(8) Demonstrate influence of the introduction condition (Kalman’s filter, etc.).
of integral regulation (transformation of P *
Synthesis of controller by method of setting
regulators in PI) in terms of elimination errors pole.
of stations state system of control system with *
Predictive control (MPC algorithms).
feedback. *
Fuzzy control, etc.
(9) Gain experience in the synthesis and the C. Restriction factors in the real functioning of the
setting up of various control algorithms (the system:
usual methods: PI þ feedforward control, PID *
The presence of measurement noise.
algorithms, the method of setting pole, MPC *
Influence of disorders on functioning of the
algorithms, feedback by the state, etc.). system.
(10) Gain experience in the implementation of *
Effects of no-modeled dynamics (or ambi-
algorithms to manage through the writing of guity of the system).
the appropriate software to work the system *
Effects of saturation of actuators (wind-up).
in real time.
(11) Gain experience in the application process Using this remotely control laboratory experi-
of computers (PLC, HMI, etc.) and accom- ments, students could cover number of contents
panied software for the implementation of above. Using manual mode students could test
specific control strategies. experimental setup working condition, set the initial
water levels, and perform the identification of the
Considering the fact that experiment could be system’s parameters or demonstration of the system
controlled via computer network, some other possible operation. Students could implement and test
educational tasks in the field of design, development, adequate control algorithm (using template for
and implementation of systems for supervision, development of different controller); compare dif-
control, and data acquisition are also possible ferent control algorithms or different methods of
10 STEFANOVIC ET AL.

adjustment of the same control algorithm (P, PI, PID, maximal height of fluid level in the tank;
fuzzy); and suggest a specification of the rational different for each student).
technical demands according to the functionality of *
Structural block diagram of the system in
the systems. They also could perform analysis of continual and digital version.
comparable simulation and experimental results and *
Check of stability of linearized model (using
make conclusion in which criteria the experimental MatLab and Simulink, present linear and
simulation confirm the theoretical consideration nonlinear model (continual and discrete).
(taking into account different factors of functioning
of real system: measurement noise, disturbance such The same, step signal is set on all models and
evaporation, heating, etc.). level of the fluid has been monitored. Validity of
linearized model (its correspondence to nonlinear
model in different points of static characteristics) is
monitored. Using MatLab, students should give
Example of Project Task
alternatives and characteristics of the plant model
As a part of different courses this project task has been (presentations of poles and zeros, frequency charac-
given to the students of Faculty of Mechanical teristics of the plant).
Engineering. Control strategy should be achieved by
controlling/monitoring the level of liquids in the lower 2. Identification of the plant: Nonlinear Simulink
tanks, irrespective of changes in the position of valves model is the basis for further simulation and
that allow filling all four tanks. There are many for experiment and parameter identification.
possibilities for creating different control algorithms, Identification of the plant gives opportunity to
which later can be tested experimentally by changing examine effects of influence of model, input,
the parameters of the system and the introduction of sampling time, identification method on the
disorder. Synthesis of control algorithm is always quality if a model is developed (ARX, for
based on preexistent information on the model of instance). Procedure of identification of model
controlled process and the available signals. Identi- using MATLAB and verification should be
fication of the model is a special mission, as well as described.
skills acquisition and signal processing. 3. Synthesis control algorithm: Selection of struc-
Application of ICT technologies for the imple- ture and adjustment of parameters of PID
mentation of control and associated functions (mon- control algorithm. In this step, continual and
itoring, diagnostics, visualization, etc.) presents digital PID algorithm should be defined.
especially field for the exercise. The task requirements 4. The analysis of the system: Analyze dynamic
are possible for other configuration laboratory features of the system with feedback using
models. programming packages Matlab and its tools
ltview and sisotool (to give the parameters and
Project Task. The plant is a tank filled with fluid analyze the dynamic response in selected
(water). Output parameter is the level of fluid in the working point, the frequency characteristics of
tank. The level of fluid is influenced by input flow and the system, poles and zero pole, etc.).
output flow. It is free to select control variable and 5. The results of the simulation: Simulation of
disturbance. The task is to present solution for function of control system with real aspects of
regulation of the level of the flow in the tank, the system (real plant is not linear). Test the
simulation of results on computer, and verification validity of the control algorithm, digital and
of suggested solution. continual and present the following diagrams:
(a) the response (disturbance ¼ 0), (b) the
1. The structure of the system: Define all compo- response by the disorder (the reference level is
nents of system, describe it, and present a constant). Eventually: (c) Improve simulation
mathematical model of the system. Students through animation, (d) examine sensitivity of
should present: the system to change the parameters OU (draw
*
Functional schema and description of work. the appropriate indexes performance and ana-
*
Detailed and realistic (nonlinear) mathemat- lyze their change), (e) on presented plant
ical model of the plant in a broader sense. implement fuzzy control and compare it with
*
Static characteristic of the plant. PID control, (f) implement some of the
*
Linearization of mathematical model of the advanced control algorithm by choice.
plant near working point (defined with x% of 6. Conclusion.
 LABVIEW-BASED REMOTE LABORATORY 11

7. Reading (and annex): Behavior and efficiency Table 1 Number of Students Included in the Survey
of controllers could be traced directly by the No. of students
visual apparatus, then through the level indica-
Year/semester number Control group WL
tors in the tank shown in Figure 4 and over
diagrams that write change of the liquid level in 2007/I 28 28
tanks during the period. In addition, during the 2007/II 26 26
experiment all measurements of liquids level 2008/I 34 34
in tanks and control voltages are saved in six
files, four files for the levels of each of the four survey there were 11 sub-grades. Results show that
tanks, and two files for the control signals of students who had access to web laboratory (they could
each pump. The contents of these files with perform laboratory exercises on their own, repeat
the numerical values can be used for precise them, and analyze the results) have better scores and
analysis of behavior controllers after the better fulfillment of educational goals compared with
execution of the experiment. Reference values other group (Table 2).
(desired) of level of liquids in the reservoirs In addition the students from ‘‘web lab’’ group
and time are also recorded. had better final grades compared with other group
(it is important to emphasize that both groups were
equal in the beginning) (Fig. 8). Another issue that
EVALUATION OF USAGE AND was raised was usability of web laboratory. According
EDUCATIONAL RESULTS to counter of access to exercise, the experiment
(coupled water tanks) was started 455 per month.
There are two main issues that should be answered: Figure 9 presents the number of experiments per day
for a period of 1 month.
(1) Evaluation of web laboratory in educational According to previous questionnaire and surveys
process. that web laboratory and remote control experiments
(2) Evaluation of usability of web laboratory could improve quality of education and contribute to
(number of access and preferred GUI— better fulfillment of educational goals, students like
graphic user interfaces) for students. to perform online experiments (68 students made
455 accesses to remote experiments).
Implementation and introduction of web labo-
ratories in educational practice in Faculty of Mechan-
ical Engineering, University of Kragujevac, started in CONCLUSION
the first semester 2007. Experiments supported by
web laboratory are part of different courses (automatic WebLab system was designed and implemented to
control, CIM, etc.). For the reason of this survey, enable remote access to instruments and experiments
students were separated in two groups (groups with for academic education in the domain of sciences and
the same number and quality of students). One group engineering. It is a modular and distributed system
was labeled as ‘‘Traditional’’ (they did not have web that can be easily expanded by adding new hardware
lab support) other group was labeled as ‘‘Web Lab’’ and software resources for new experiments. In
(they had access to web lab). Survey covered three addition to benefits of remote access and 24 h/7 day
semesters: 2007/I (the first semester in 2007), 2007/II availability, students can use it at their own learning
(the second semester in 2007), and 2008/I (the first pace. Sharing of specific and expensive equipment
semester in 2008). The number of students included among various institutions is made much easier, as it
experiment is presented in Table 1. can be used from any place with secure and controlled
As it was stated in Educational Goals Section, access.
the important educational goals were recognized. The There is little availability of real physical systems
most important educational goals are listed in the or laboratories to perform the experiments in control
table. On the final exam students are evaluated courses [23]. WebLab University of Kragujevac has
according to fulfillment of educational goals (for the laboratory experiment in control courses that could be
purpose of this survey). Each of the stated outcome remotely controlled (Gantry Crane—implementation
was evaluated according to the answers to the specific based on C# and Coupled tanks—implementation
group of questions (there were 11 groups of questions based on LabVIEW). WebLab system at University
in order to evaluate fulfillment of educational goals). of Kragujevac is a combination of two different
Students received just one grade but for the purpose of implementation approaches that are transparent to the
12 STEFANOVIC ET AL.

Table 2 Results of Survey: Fulfillment of Educational Goals (5
10 Grade Scale)

Number of students—88
Educational goals Traditional Using WebLab

(1) Get familiar with the system of feedback control—general ideas from the theory of 8 9
systems, components, and systems terminology
(2) Gain experience in calibration of sensors, performance and implementation of laboratory 6 10
experiment, and preparation of appropriate software
(3) Test acquired knowledge in the field of modeling and process identification 7 9
(4) Analyze static and dynamic features of processes and systems with feedback to make 8 9
quality assessments and evaluate the main parameters of the control system
(5) Demonstrate performance and limits of the system with open feedback (manual control) 7 8
when there is a need for change in the set of values or fix the effects of disturbance
(6) Demonstrate unstable work with the system with feedback, as the result of influence of 6 8
clear time delays in the system
(7) Demonstrate influence of changes of P (proportional) and the dynamic and static 7 8
characteristics of the system
(8) Demonstrate influence of the introduction of integral regulation (transformation of 7 8
P regulators in PI) in terms of elimination errors of stations state system of control
system with feedback
(9) Gain experience in the synthesis and the setting up of various control algorithms (the 6 8
usual methods: PI þ feedforward control, PID algorithms, the method of setting pole,
MPC algorithms, feedback by the state, etc.)
(10) Gain experience in the implementation of algorithms to manage through the writing of 7 8
the appropriate software to work the system in real time
(11) Gain experience in the application process of computers (PLC, HMI, etc.) and 7 10
accompanied software for the implementation of specific control strategies

user of the WebLab. Implementation based on measurement and control applications, except for
programming languages and Microsoft platform is some specific tasks that could require customized
quite satisfactory but takes much longer time for software. Some other solutions have been developed
development compared to development with Lab- using the platform Matlab/Simulink [23,24]. The
VIEW system that enables rapid development and scalability and the ease in designing a user interface
excellent visual design using number of ready-to- provided with LabVIEW, as well as its multi-platform
use controls and indicators not present elsewhere. implementation, make this package well suited to
LabVIEW is in general better and much more develop didactic tools. Moreover, the possibility to
effective alternative to programming when developing compile stand-alone virtual instruments, which can be

Figure 8 Average grades of students in different groups.

 LABVIEW-BASED REMOTE LABORATORY 13

Figure 9 Number of experiments per day (totally 455 per month).

distributed freely to the students, is a big advantage. educational outcomes in the field of control engineer-
Students are provided with stand-alone VIs, which ing. The main 11 educational goals were recognized
enable them to experiment locally or remotely the and success of different students’ groups was
behaviors of the controllers they have studied in class, measured according to these goals. Fulfillment of
such as the PID controller. educational goals and average grades of students show
The most important characteristic of the Lab- that usage of web laboratories produce better results
VIEW software support for experiment with Coupled and contribute to better control engineering education.
water tanks is the design of the universal controller The main reasons for these advantages are the facts
template for implementation of any kind of controller that web laboratories change character of courses,
while the designer is free from all implementation make laboratory exercises easier (giving unlimited
details concerning the experimental setup. All imple- opportunity for students to repeat experiments using
mentation details are handled by the main experiment different parameters), and give open access (from
VI which uses controller based on template as sub VI. anyplace, anytime).
Users can design their own controllers using con-
troller template, and remotely control real laboratory REFERENCES
equipment. This feature enables easy changing and
comparing different controllers, variation of different [1] R. Šafaric, M. Truntic, D. Hercog, and G. Pacnik,
parameters, validation of studied control algorithms, Control and robotics remote laboratory for engineering
and changing module controllers without restart education, Int J Online Eng 1 (2005), 1
8.
software environment. [2] Q. Yu, B. Chen, and H. H. Cheng, Web-based control
Once the experiment is developed in LabVIEW, system design and analysis, IEEE Control Syst Mag
with simple setup it is available on the web for remote 24 (2005), 45
57.
users. [3] Automated Internet Measurement Laboratory, Rensse-
Web laboratory has been used in educational laer Polytechnic Institute: http://nina.ecse.rpi.edu/shur/
remote/.
process in University of Kragujevac for years. During
[4] Bugscope, University of Illinois at Urbana: http://
the year 2007 and 2008, different experiments
bugscope.beckman.uiuc.edu/.
controlled over Internet were used in education, [5] Control Systems Lab, The University of Tennessee
especially control engineering education. The survey at Chattanooga: http://chem.engr.utc.edu/Webres/
on different groups of students, during that time, Stations/controlslab.html.
shows that students who had access to web laboratory [6] e-Lab, City University of New York, http://www.
(and possibility to perform laboratory exercises on missiontechnology.com/nsfrobot/.
their own with full support of web lab) have better [7] i-Lab, MIT, http://ilab.mit.edu/ServiceBroker/.
14 STEFANOVIC ET AL.

[8] Web-Lab, University of Kragujevac, http://www.cqm. urement and control via Internet, University of
co.yu/weblab/. Kragujevac, Kragujevac, Serbia, 2007.
[9] F. Barrero, S. Toral, and S. Gallardo, eDSPLab: [18] M. Matijević, V. Cvjetković, V. Ranković, and M.
Remote laboratory for experiments on DSP applica- Stefanović, Internet-mediated process control labora-
tions, Internet Res 18 (2008), 79
92. tory, University of Kragujevac, Kragujevac, Serbia, 2007.
[10] K. Forinash and R. Wisman, Building real laboratories [19] J. Hua and A. Ganz, Web enabled remote laboratory
on the Internet, Int J Cont Eng Educ Lifelong Learn 15 (R-Lab) framework, 33rd ASEE/IEEE Frontiers in
(2005), 56
66. Education Conference, Boulder, CO, November 5
8,
[11] S. Gallardo, F. Barrero, M. R. Marťˇinez-Torres, S. L. 2003, pp T2C-8
T2C-13.
Toral, and M. J. Duran, Addressing learner satisfaction [20] J. Hardison, D. Zych, J. A. del Alamo, V. J. Harward,
outcomes in Electronic Instrumentation and Measure- S. R. Lerman, S. M. Wang, K. Yehia, and C.
ment Laboratory Course Organization, IEEE Trans Varadharajan, The microelectronics WebLab 6.0—An
Educ 50 (2007), 129
136. implementation using web services and the iLab
[12] C. C. Ko, B. M. Chen, and J. Chen, Creating WEB shared architecture, Exploring Innovation in Education
based laboratories, Springer, New York, 2004. and Research
[13] V. M. Cvjetkovic and Y. S. Sulema, Remote laboratory [21] F. Davoli, G. Spano, S. Vignola, and S. Zappatore,
for supporting e-studies in electronics, Int J Online Eng LABNET: Towards remote laboratories with unified
2 (2006). http://www.online-journals.org/index.php/ access, IEEE Trans Instrum Meas 55 (2006),
i-joe/issue/view/34 1551
1558.
[14] V. M. Cvjetkovic, On line experiments with linear [22] B. Voglauer, R. Garcia, and H. P. Jörgl, Improvements
analog systems from the first to nth order, REV 2006, of a three-tank-system operated in real time with
Remote Engineering and Virtual Instruments, Maribor, Matlab in a PLC-Profibus-Network, 16th IFAC World
Slovenia, June 29
30, 2006. Congress, Prague, 2005.
[15] V. M. Cvjetkovic, D. Stevanovic, and M. Matijevic, [23] L. M. Jimenez, R. Puerto, O. Reinoso, R. P. Neco, and
Remote system for development, implementation and C. Fernandez, Remote control laboratory using Matlab
testing of control algorithms, Int J Online Eng 3 (2007). and Simulink, IEEE International Symposium on
http://www.online-journals.org/index.php/i-joe/issue/ Industrial Electronics, 2007, ISIE2007, June 4
7,
view/38 2007, pp 2963
2966.
[16] V. M. Cvjetković, M. Matijević, and M. Stefanović, [24] S. Bogosyan, M. Gokasan, A. Turan, and R. W. Wies,
Remotely controlled experiment for gantry crane, REV Development of remotely accessible Matlab/Simulink
2007, Porto, Portugal, June 25
27, 2007. based electrical drive experiments, IEEE International
[17] M. Stefanović, M. Matijević, M. Ravlić, and V. Symposium on Industrial Electronics, 2007, ISIE2007,
Cvjetković, Laboratory experimental setups for meas- June 4
7, 2007, pp 2984
2989.

BIOGRAPHIES

M. Stefanovic received his PhD from the M. Matijevic received his PhD from the
Department of Production and Industrial Department of Automatic Control, Faculty of
Engineering, Faculty of Mechanical Engi- Mechanical Engineering, University of Kra-
neering, University of Kragujevac, Serbia. gujevac, Serbia. He is currently an associate
He is currently an assistant professor in the professor in the Department of Automatic
Department of Industrial Engineering. His Control and Applied Mechanics. His current
current research interest includes web serv- research interest includes digital control,
ices, information systems, and CIM systems. computer-added measurement and control.
He is a member of International Federation
for Information Processing—Council TC3—Education.
V. Simic is a PhD student in the Faculty of
Science, University of Kragujevac, Serbia.
V. Cvijetkovic received his PhD from the She is a research assistant in the Department
Faculty of Science, University of Kraguje- of Physics and Information Technology,
vac, Serbia. He is employed at Faculty of Faculty of Science. Her current research
Science, Department of Physics as an interest includes computer science, artificial
assistant professor. He teaches subjects like intelligence, and expert systems.
informatics for the students of physics:
Object-oriented programming, Information
systems, Web programming, Data acquisi-
tion, Expert systems. His scientific interest
and field of research include informatics support for physics, expert
systems for spectra analysis, data acquisition, measurement and
control, web laboratories.