Editorial Board • Prof. Dr. Eng. Ioan NAFORNITA, Editor-in-chief • Prof. Dr. Eng. Virgil TIPONUT • Prof. Dr. Eng. Alexandru ISAR • Prof. Dr. Eng. Dorina ISAR • Prof. Dr. Eng. Traian JURCA • Prof. Dr. Eng. Aldo DE SABATA • Prof. Dr. Eng. Florin ALEXA • Prof. Dr. Eng. Radu VASIU • Lecturer Dr. Eng. Maria KOVACI, Scientific Secretary • Associate Prof. Dr. Eng. Corina NAFORNITA, Scientific Secretary Scientific Board • Prof. Dr. Eng. Monica BORDA, Technical University of Cluj-Napoca, Romania • Prof. Dr. Eng. Aldo DE SABATA, Politehnica University of Timisoara, Romania • Prof. Dr. Eng. Karen EGUIAZARIAN, Tampere University of Technology, Institute of Signal Processing, Finland • Prof. Dr. Eng. Liviu GORAS, Technical University Gheorghe Asachi, Iasi, Romania • Prof. Dr. Eng. Alexandru ISAR, Politehnica University of Timisoara, Romania • Prof. Dr. Eng. Michel JEZEQUEL, TELECOM Bretagne, Brest, France • Prof. Dr. Eng. Traian JURCA, Politehnica University of Timisoara, Romania • Prof. Dr. Eng. Ioan NAFORNITA, Politehnica University of Timisoara, Romania • Prof. Dr. Eng. Mohamed NAJIM, ENSEIRB Bordeaux, France • Prof. Dr. Eng. Emil PETRIU, SITE, University of Ottawa, Canada • Prof. Dr. Eng. Andre QUINQUIS, Ministère de la Défense, Paris, France
56
Embed
Editorial Board - Politehnica University of Timișoara · 2015-09-10 · Buletinul Ştiinţific al Universităţii Politehnica Timişoara TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Editorial Board
• Prof. Dr. Eng. Ioan NAFORNITA, Editor-in-chief
• Prof. Dr. Eng. Virgil TIPONUT
• Prof. Dr. Eng. Alexandru ISAR
• Prof. Dr. Eng. Dorina ISAR
• Prof. Dr. Eng. Traian JURCA
• Prof. Dr. Eng. Aldo DE SABATA
• Prof. Dr. Eng. Florin ALEXA
• Prof. Dr. Eng. Radu VASIU
• Lecturer Dr. Eng. Maria KOVACI, Scientific Secretary
• Associate Prof. Dr. Eng. Corina NAFORNITA, Scientific
Secretary
Scientific Board
• Prof. Dr. Eng. Monica BORDA, Technical University of
Cluj-Napoca, Romania
• Prof. Dr. Eng. Aldo DE SABATA, Politehnica University
of Timisoara, Romania
• Prof. Dr. Eng. Karen EGUIAZARIAN, Tampere
University of Technology, Institute of Signal Processing,
Finland
• Prof. Dr. Eng. Liviu GORAS, Technical University
Gheorghe Asachi, Iasi, Romania
• Prof. Dr. Eng. Alexandru ISAR, Politehnica University
of Timisoara, Romania
• Prof. Dr. Eng. Michel JEZEQUEL, TELECOM Bretagne,
Brest, France
• Prof. Dr. Eng. Traian JURCA, Politehnica University of
Timisoara, Romania
• Prof. Dr. Eng. Ioan NAFORNITA, Politehnica University
of Timisoara, Romania
• Prof. Dr. Eng. Mohamed NAJIM, ENSEIRB Bordeaux,
France
• Prof. Dr. Eng. Emil PETRIU, SITE, University of
Ottawa, Canada
• Prof. Dr. Eng. Andre QUINQUIS, Ministère de la
Défense, Paris, France
• Prof. Dr. Eng. Maria Victoria RODELLAR BIARGE,
Polytechnic University of Madrid, Spain
• Prof. Dr. Eng. Alexandru SERBANESCU, Technical
Military Academy, Bucharest, Romania
• Prof. Dr. Eng. Virgil TIPONUT, Politehnica University
of Timisoara, Romania
• Prof. Dr. Eng. Radu VASIU, Politehnica University of
Timisoara, Romania
Advisory Board
• Prof. Dr. Eng. Ioan NAFORNITA, Politehnica University
of Timisoara, Romania
• Prof. Dr. Eng. Alexandru ISAR, Politehnica University of
Timisoara, Romania
• Prof. Dr. Eng. Radu VASIU, Politehnica University of
Timisoara, Romania
• Prof. Dr. Eng. Florin ALEXA, Politehnica University of
Timisoara, Romania
• Prof. Dr. Eng. Vladimir CRETU, Politehnica University
of Timisoara, Romania
Buletinul Ştiinţific al Universităţii Politehnica Timişoara
TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
Volume 60(74), Issue 1, 2015
CONTENTS
Cristina Stolojescu Crisan, Alexandru Isar:
"Optical Coherence Tomography Speckle Reduction in the Wavelets Domain".......... 3
Mihai Micea, Cristina Stangaciu, Vladimir Cretu:
"Analysis of Non-Preemptive Scheduling Techniques for HRT Systems"..................... 9
Valentin Stangaciu, Olivia Datcu, Mihai Micea, Vladimir Cretu:
"INVERTA – Specification of Real-Time Scheduling Algorithms"............................. 15
Cristian Cosariu, Alexandru Iovanovici, Lucian Prodan, Mircea Vladutiu:
"TACTICS: Adaptive Framework for Reactive Control of Road Traffic Systems"..... 21
Maria Kovaci, Horia Balta:
"Performance of Turbo Encoders with 64-QAM Modulators Interfacing Systems in
[14]. Micea, M.V.: HARETICK: A Real-Time Compact Kernel
for Critical Applications on Embedded Platforms. In:
Proceedings of the 7th International Conference on
Development and Application Systems, Suceava, 2004, p. 16.
14
Buletinul Ştiinţific al Universităţii Politehnica Timişoara
TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
Volume 60(74), Issue 1, 2015
INVERTA – Specification of Real-Time Scheduling
Algorithms
V. Stangaciu1 , O. Datcu
2, M. Micea
3, V. Cretu
4
1 Faculty of Automation and Computers, Dept. of Computer and Software Engineering
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected] 2 Faculty of Automation and Computers, Dept. of Computer and Software Engineering
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected] 3 Faculty of Automation and Computers, Dept. of Computer and Software Engineering Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected] 4 Faculty of Automation and Computers, Dept. of Computer and Software Engineering
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected]
Abstract – This paper describes how the scheduling
algorithms for real time applications can be specified
formally and the development of a simulator that
verifies if a set of tasks for a real time application can be
scheduled with an existing scheduling algorithm or with
an algorithm defined by the user. This simulator is part
of the integrated visual environment for designing and
analysing real-time applications called INVERTA.
Keywords: scheduling, real-time, simulator
I. INTRODUCTION
Embedded systems and digital signal processing
(DSP) systems are used in a variety of application
today. Such applications include: automotive control,
nuclear plant surveillance, flight control systems, and
industrial mechatronics. These systems usually run
hard real-time tasks, for which the violation of their
time requirements (deadlines), may have catastrophic
impacts, thus special task scheduling policies must be
used. This class of hard real time scheduling policies
must provide schedulability tests which state if a
certain set of tasks is feasible or not. If a set of task is
feasible with a certain algorithm there is a guarantee
that no deadline is missed. Thus, these algorithms
have been and still are, heavily analyzed [1, 2].
OPEN-HARTS (Operating Environment for Hard
Real-Time Systems) is a methodology that was
introduced recently for development and
implementation of hard real-time systems and
applications and is based on signals and tasks. This
system is represented by the interconnection of two
sub-systems: one for analysis of the task set called
INVERTA (Integrated Visual Environment for Real-
Time Application Analysis and Development) and
one for running the task set called HARETICK (Hard
Real-Time Compact Kernel).
The paper has the following structure which will
be further described: problem statement, theoretical
foundations, related work, proposed solution and
research methodology, implementation, experimental
results, contribution and conclusions.
II. PROBLEM STATEMENT
INVERTA allows the building, specification and
visual display of real-time applications, designed as a
set of tasks of different types, each task having a
characteristic set of parameters (including parameters
of time) and a set of control links with other tasks of
the application.
The INVERTA sub-system which is presented in
this paper, along with HARETICK (Hard Real-Time
Compact Kernel) sub-system is part of OPEN-
HARTS (Operating Environment for Hard Real-Time
Systems) system. The role of the INVERTA sub-
system is to take the running context of the current
application from the HARETICK module, to analyse
the application, to modify its parameters and to send
the modified application back to it.
Most scheduling simulators do not offer the
possibility to simulate a customized real time
scheduling algorithm. This is a drawback because
users that propose new algorithms cannot test them to
see if they are feasible or not. Another disadvantage
of some of the existing scheduling simulators is that
they are not optimized to work for high number of
tasks.
III. THEORETICAL FOUNDATIONS
A real time system is defined by J.S. Ostroff as:
“A real-time system (RTS) is any system in which the
time at which the output is produced is significant.
This is usually because the input corresponds to some
movement in the physical world, and the output has to
relate to that same movement. The lag from input time
to output time must be sufficiently small for
acceptable timeliness.” [3]
Real time system can be divided in the following:
critical RTS (not meeting the deadline can result in a
catastrophe), strict RTS (not meeting the deadline
results in a wrong behaviour of the system), and soft
15
RTS (not meeting the deadline results in the loss of
the system’s value and of the quality provided by the
system).
Tasks scheduling refers to finding reliable
solutions for the processor’s assignment, for each
tasks, in a way in which there is no overlapping in
their execution while the system operates.[4]
Taking into consideration if they admit or not
interruptions, the scheduling algorithms can be
classified as follows: preemptive (the execution of a
task can be interrupted by a task with a higher
priority) and non-preemptive (the execution of a task
cannot be interrupted).
Off-line non-preemptive scheduling techniques
provide solutions to hard real-time constraints and
predictability, which are important demands in critical
applications. On the other hand, these scheduling
techniques do not provide flexibility, as online
scheduling techniques like the ones that rely on task
prioritization (RM, EDF, LLF and others).
A scheduler is the part of a system that deals with
the operation of scheduling a task set. In order to find
a valid schedule for a task set the scheduler executes a
schedulability test. The scheduler can be preemetive if
the execution of a task can be interrupted by another
task and non-preemtive if no interruption is allowed.
Fig. 1 presents a real-time scheduler [5]. As it can
been seen in Fig. 1, the scheduling algorithm needs
the task set and the resource management protocol to
apply the schedulability test, for a given system
architecture, and give an answer if the task set can be
scheduled or not.
Fig. 1. Real-time scheduler
IV. RELATED WORK
Liu and Layland [6] showed that RM is the best
fixed priority algorithm to be used in a uniprocessor
system. They proved that a task set that is not
schedulable by RM it cannot be scheduled by any
other fixed priority scheme. They were the first
authors who provided a necessity condition for a set
of n periodic tasks under RM, based on the processor
utilization factor U (1) and an upper bound bn (2),
both defined below:
∑=
=n
i i
i
T
CU
1
, (1)
where, Ci represents the computation time of task
i and Ti represents the period of the same task i.
)12( /1 −= n
n nb (2)
The condition is that if the processor utilization
factor is greater than bn, then the set of tasks is not
schedulable by RM. This condition was improved by
Bini in [7] where the Hyperbolic Bound (HB)
improves the acceptance ratio by a factor of √2 for
large n, compared with the Liu and Layland test.
According to HB method, a set of periodic tasks is
schedulable by RM if condition (3) is satisfied:
Cn
i
iU1
2)1(=
≤+ (3)
In [8] a sufficiency test is provided for the same
RM algorithm. The task set is proven to be
schedulable if the utilization factor is smaller or equal
to:
)12( −≤ nnU (4)
The first formulation of the Rate Monotonic
Analysis was done by Lehoczky in [9]. The goal of
the article was to present an exact characterization of
the ability of the rate monotonic algorithm to meet the
deadlines for a set of period tasks. The article also
includes a stochastic analysis of the performance of
the algorithm when the task sets are generated
randomly. Manabe and Aoyagi improved this article
in [10] by reducing the number of points where the
time demand has to be checked. Another
improvement was done by Bini and Buttazzo [11],
who proposed a way to trade complexity versus
accuracy of the RM feasibility tests.
In [5] Chen presents an overview of the existing
real-time scheduling tool-kits. These tools are useful
for real-time system designers and programmers to
verify if a task set is schedulable with a scheduling
algorithm. Chen divides these scheduling tool-kits
based on their functionality in the following
categories: simulators, simulation languages and
frameworks.
A drawback of the simulators is that they have all
the functionality predefined and the user cannot add
new code. Among the developed simulators there are:
GAST [12], DET/SAT/SIM, PERTS SAT,
DTRESS/PERTSSim, AFTER, Brux, CAISARTS,
and Scheduler 1-2-3.
A simulation language called STRESS was
proposed in [13]. Although STRESS is a good tool to
evaluate scheduling algorithms and can be used to
design new algorithms, the cost of a context switch is
considered to be zero, a task can only start on a tick of
the system clock and resources are limited to
semaphores. Asserts (A Software Simulation
Environment for Real-Time Systems) [14] is another
simulation language which is focused on distributed
and heterogeneous systems. The user can define
nonstandard systems by specifying the task body in
pseudo-code.
Frameworks take into consideration the user
requirements and the possibility of extension. A
framework is able to generate, compile and the run
code based on the user specification of a simulation
environment, scheduler, resource management
16
protocols, and task set. A framework of the Oregon
State University, which is implemented in C++ was
presented in Chen’s study from [12]. Another
framework that targets failure analysis and
hierarchical scheduling was described by Matthew
Francis Storch in [15].
Cheddar [16] is another framework, which was
implemented in Ada language, and allows the user to
check if a real time application meets its temporal
constraints. The purpose for creating this framework
was mainly educational. This framework can connect
to other tools such as editors, design tool and
simulations, easily because the data sent to the
framework and received by the framework is in XML
format.
V. PROBLEM STATEMENT
This paper defines a meta-language for the
INVERTA environment, which has the ability to
model numerous schedulers (executives). The
simulation will be based on scripts that will be
translated into simulation parameters and interpreted
by the simulation engine.
The general architecture of the simulator
described in this paper is presented in Fig 2. The
simulator was developed as a plugin for INVERTA
application. As it can been seen in the figure, the
simulator plugin receives as an input a configuration
for a task set and an XML file in which the scheduling
algorithm is specified. INVERTA environment is
used to describe the configuration of the task set. The
XML specification file is generated by the Formal
Specification plugin from INVERTA. This plugin
offers a User Interface where the scheduling
algorithm can be defined in an XML format.
Fig. 3 illustrates the structure of the XML file
used for describing the scheduling algorithm. The
XML file is composed of five tags. The first one is the
ScheduleName, in which the name for the scheduling
algorithm is entered. The second tag, Acronym,
identifies the acronym used for the algorithm. The
value from this tag is optional. The next tag,
DeclarePriority, describes the type of the scheduling
algorithm: static, dynamic or special. The forth tag,
DeclarePreemtiveBehavior, specifies if the algorithm
is preemptive or non-preemptive. The condition for
priority assignment is defined in the last tag, called
PriorityAssignement.
In order to evaluate the expression that defines
the priority assignment for a scheduling algorithm, the
expression is first split into atoms, which are stored in
a list of atoms. An atom can be an operator, a numeric
constant or a task parameter. Based on the literature
review, a set of task parameters were identified:
Task Set
Configuration Scheduling Plugin
XML Specification
of Scheduling
Algorithm
Formal
Specification
Plugin
Scheduling output
Fig. 2. General architecture of Task Simulator
• The name of the scheduling algorithm SchedulerName
• Expression used for assigning of priorities PriorityAssignement
Fig. 3. XML Specification file structure
− T[i] - The task relative period
− D[i] - The task relative deadline
− C[i] - The task computation time
− P[i] - The task priority
− S[i] - The task start time inside current period
− d[i] - The task absolute deadline
− s[i] - The task absolute start time
In the next step, the expression is transformed in
Reversed Polish Notation. From this notation the
binary evaluation tree was constructed. The result of
the expression is obtained from the in-order traversal
of the tree. The above steps are presented in Fig. 4.
17
Expression
Parser
Expression String
Reversed Polish
Notation
Expression Tree
Expression Tree
Evaluation
(Inorder)
Fig. 4. Expression evaluation steps
The list of atoms is iterated in order to verify each
atom. If an atom is a number, it is added to the
Reversed Polish Notation list. If the atom is an
operator and the stack is empty the atom is pushed on
the stack. If the stack is not empty, the precedence of
the current atom is compared with the precedence of
the atom from the top of the stack, and a specific
action is performed based on the precedence. If the
atom is a start of parenthesis character the atom is
pushed on the stack. On the other hand, if the end of
parenthesis is encountered the content of the stack
until the start of parenthesis is stored in the output
RPN list. The pseudo-code used to specify the RPN
list construction algorithm is very similar with C
programing language. The reserved words are written
in bold and the main operations are listed in italic
style:
− isNumber – returns true if an atom is a number
and false otherwise
− isOperator – returns true if an atom is an
operator and false otherwise
− isStartParan – returns true if the atom is a start of
parenthesis character and false otherwise
− isStopParan – returns true if the atom is a stop of
parenthesis character and false otherwise
− isStackEmpty – returns true if the sack is empty
and false otherwise
− Push – adds an element to the stack
− Pop – removes the element from the top of the
stack
− Peek – returns the element from the top of the
stack
− Precedence – returns the precedence of the
operator given as a parameter
− AddRPNList – adds an element to the Reversed
Polish Notation list
Reversed Polish Notation construction algorithm
1: foreach (Atom in AtomList) do 2: if isNumber(Atom) do 3: 4: 5: 6: 7: 8: 9:
10: 11: 12: 13: 14: 15: 16: 17: 18: 19: 20: 21:
AddRPNList(Atom)
Push(Atom) else if isStartParan(Peek()) do
Push(Atom)
else if isOperator(Atom) do if isStackEmpty() do
else if Precedence(Atom) > Precedence(Peek()) do Push(Atom)
else while (!isStackEmpty() && !isStartParan(Peek()) && Precedence(Atom) < Precedence(Peek())) do
TempAtom = Pop()
end do Push(Atom)
end if else if isStartParan(Atom) do
Push(Atom)
22: 23:
else if isStopParan(Atom) do
24: 25: 26:
while (!isStackEmpty() && !isStartParan(Peek())) do
TempAtom = Pop()
end do
27: 28:
Pop(Atom) end if while (!isStackEmpty()) do
29:
30:
TempAtom = Pop()
AddRPNList(TempAtom)
AddRPNList(TempAtom)
AddRPNList(TempAtom)
31: 32:
end do end foreach
Fig. 5 Reversed Polish Notation Construction
Algorithm
VI. EXPERIMENTAL RESULTS
The output of the Scheduling PlugIn from
INVERTA for the task set defined in Fig. 6 and
scheduled with Rate Monotonic Non-Preemptive, a
static algorithm, is presented in Fig. 8. Fig. 7 presents
the XML file that specifies the Rate Monotonic Non-
Preemptive algorithm.
Fig. 6 Task set scheduled with RM algorithm
18
Fig. 7 XML specification for RM algorithm
Fig. 8 RM scheduling example
The output of the Scheduling PlugIn from
INVERTA for the task set defined in Fig. 9 and
planned with MLFNP - Minimum Laxity First Non-
Preemptive, a dynamic algorithm, is presented in Fig.
11. The task set from Fig. 9 was taken from the
example that was treated in [1] for MLFNP algorithm.
Fig. 10 presents the XML file that specifies the
MLNFNP algorithm.
Fig. 9 Task set scheduled with MLFNP algorithm
Fig. 10 XML specification for MLFNP algorithm
Fig. 11 MLFNP scheduling example
VII. CONCLUSION
The development of real-time systems remains a
very important research domain because of the
complexity of the problems which characterize these
systems. Task scheduling is one of the most important
problems from real-time systems and without which
the function of the system would be unfeasible. This
fact is supported by the tremendous number of
research papers from this domain which treat different
types of scheduling algorithms. INVERTA
environment is intended to help users define real-time
applications in a visual user friendly environment,
analyse these applications from the feasibility point of
view and simulate existing and custom defined
scheduling algorithms.
ACKNOWLEDGMENT
This work was partially supported by the strategic
grant POSDRU/159/1.5/S/137070 (2014) of the
Ministry of National Education, Romania, co-
financed by the European Social Fund – Investing in
People, within the Sectoral Operational Programme
Human Resources Development 2007-2013.
19
REFERENCES
[1] S. Baruah, M. Bertogna, and G. Buttazzo, "A
Review of Selected Results on
Uniprocessors," in Multiprocessor
Scheduling for Real-Time Systems, ed:
Springer International Publishing, 2015,
ISBN: 978-3-319-08695-8, pp. 29-33.
[2] Yan Feng Zhai and Feng Xiang Zhang, "A
Review of Sufficient Schedulability Analysis
for Fixed Priority Scheduling Systems,"
Applied Mechanics and Materials, vol. 741,
no. 1, pp. 856-859 2015.
[3] J. S. Ostroff, "Formal methods for the
specification and design of real-time safety
critical systems," J. Syst. Softw., vol. 18, no.
1, pp. 33-60, 1992.
[4] M. V. Micea, "Proiectarea si implementarea
sistemelor timp-real pentru aplicatii critice de
achizitie si prelucrare numerica de semnal,"
PhD, Politehnica Timisoara, 2004.
[5] J. Chen, "Extensions to Fixed Priority with
Preemption Threshold and Reservation-
Based Scheduling," PhD, University of
Waterloo, 2005.
[6] C. L. Liu and J. W. Layland, "Scheduling
Algorithms for Multiprogramming in a Hard-
Real-Time Environment," J. ACM, vol. 20,
no. 1, pp. 46-61, 1973.
[7] E. Bini, G. C. Buttazzo, and G. M. Buttazzo,
"Rate monotonic analysis: the hyperbolic
bound," Computers, IEEE Transactions on,
vol. 52, no. 7, pp. 933-942, 2003.
[8] R. Devillers, Jo, #235, and l. Goossens, "Liu
and Layland's schedulability test revisited,"
Inf. Process. Lett., vol. 73, no. 5-6, pp. 157-
161, 2000.
[9] J. Lehoczky, S. Lui, and Y. Ding, "The rate
monotonic scheduling algorithm: exact
characterization and average case behavior,"
in Real Time Systems Symposium, 1989.,
Proceedings., 1989, pp. 166-171.
[10] Y. Manabe and S. Aoyagi, "A Feasibility
Decision Algorithm for Rate Monotonic
andDeadline Monotonic Scheduling," Real-
Time Syst., vol. 14, no. 2, pp. 171-181, 1998.
[11] E. Bini and G. C. Buttazzo, "Schedulability
analysis of periodic fixed priority systems,"
Computers, IEEE Transactions on, vol. 53,
no. 11, pp. 1462-1473, 2004.
[12] J. Johnson, "The Impact of Application and
Architecture Properties of Real-Time
Multiprocessor Scheduling," PhD, CTH
Department of Computer Engineering,
Computer Architecture Laboratory (CAL),
MicroMultiProcessor Group, 1997.
[13] N. C. Audsley, A. Burns, M. F. Richardson,
and A. J. Wellings, "STRESS: a simulator
for hard real-time systems," Softw. Pract.
Exper., vol. 24, no. 6, pp. 543-564, 1994.
[14] K. Ghose, S. Aggarwal, P. Vasek, S.
Chandra, A. Raghav, A. Ghosh, and D. R.
Vogel, "ASSERTS: a toolkit for real-time
software design, development and
evaluation," in Real-Time Systems, 1997.
Proceedings., Ninth Euromicro Workshop
on, 1997, pp. 224-232.
[15] M. F. Storch, "A framework for the
simulation of complex real-time systems,"
University of Illinois at Urbana-Champaign,
1997.
[16] F. Singhoff, J. Legrand, L. Nana, L. Marc,
and #233, "Cheddar: a flexible real time
scheduling framework," presented at the
Proceedings of the 2004 annual ACM
SIGAda international conference on Ada:
The engineering of correct and reliable
software for real-time & distributed
systems using Ada and related technologies,
Atlanta, Georgia, USA, 2004.
20
Buletinul Ştiinţific al Universităţii Politehnica Timişoara
TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
Volume 60(74), Issue 1, 2015
TACTICS: Adaptive Framework for Reactive Control of
Road Traffic Systems
Cristian Cosariu1, Alexandru Iovanovici
2, Lucian Prodan
3, Mircea Vladutiu
4
1 Faculty of Automation and Computer Engineering, Computer Engineering and Information Technology Dept.
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected] 2 Faculty of Automation and Computer Engineering, Computer Engineering and Information Technology Dept.
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected] 3 Faculty of Automation and Computer Engineering, Computer Engineering and Information Technology Dept. Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected] 4 Faculty of Automation and Computer Engineering, Computer Engineering and Information Technology Dept.
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected]
Abstract – This paper proposes an adaptive traffic
framework used to respond to continuous traffic
changes in a network with control points in key
intersections as derived trough complex network
analysis. The main actuators of this framework are the
intelligent traffic ligths which run the entire adaption
algorithm without affecting the current deployed
infrastructure. We illustrate the proposed solution
through a case study conducted over the city of
Timisoara, Romania. Our algorithm was tested using the
VISSIM simulator and results show improvements in
reducing waiting times and queue lengths over the
currently deployed solution based on fixed time plans.
Keywords: traffic control framework, intelligent
transportation systems, complex network analysis,
urban topology, road traffic quality
I. INTRODUCTION
Congestion and its side effects are real problems that
concern any urban transport system. Intelligent
Transportation Systems (ITS) gather the most
significant work done in this direction in order to
improve urban transportation operations.
Large and complex systems are still being
developed and deployed all over the world. A large
number of them use a centralized control scheme to
coordinate traffic movement based on the input read
from pavement installed sensors, cameras, video
surveillance, on-car devices and the list could
continue [2]. But, all these control systems require a
framework to guide the integration of all used smart
devices into a real intelligent system.
Based on the data acquisition methods traffic
systems can be static or real time. The real time
control ones respond to traffic changes by processing
the recorded data as they read it. A further analysis
reveals that real time traffic systems are reactive or
proactive [2]. In the proactive approach, traffic
control system is adapting its operations based on the
data estimated to be on a certain moment of time.
Reactive systems respond to traffic changes with a
certain delay, caused by the read time needed to
determine actual traffic conditions. Proactive systems
were deployed in the early stages of ITS development,
but do not seem to have a general solution and
continue to motivate the research in this direction.
While algorithms trying to forecast traffic conditions
are still being developed [3], reactive methodologies
are already implemented by systems like, SCATS,
SCOOTS, UTOPIA, MOTION or BALANCE [2].
Instead of trying to forecast traffic conditions,
another solution is to react quickly and adapt to traffic
changes as they occur. Minimizing the reaction time
of a system to adapt to traffic changes where reactive
systems still have to be improved. The most used
traffic actuator by the reactive systems remains the
traffic signal [4]. From changing phase order to
modifying cycle length and switching between
different timing plans to find the right phase order are
just few of the currently used solutions [5]. Reactive
systems are systems whose role is to maintain an
ongoing interaction with their environment rather than
produce some final value upon termination. Typical
examples of reactive systems are Air traffic control
system, programs controlling mechanical devices such
as a train, a plane, or ongoing processes such as a
nuclear reactor.
TACTICS is the adaptive traffic framework
envisioned to respond to continuous traffic changes in
a network that implements the three layered
formalism proposed in [6]. The main actuators of this
framework are the intelligent traffic lights which run
the adaptive green time algorithm. The hardware
deployment is done without affecting the current
infrastructure. A new hardware that uses only video
camera detection and communication module will be
used, without the need of installing pavement sensors
where they are not already installed. The proposed
workflow was partially tested as described in [6],
21
using the VISSIM [19] simulator. Improvements were
obtained in terms of reducing waiting times and queue
lengths over the currently deployed solution based on
fixed time plans.
This paper proposes a framework for developing
a reactive traffic control system based on the adaption
of green time values for traffic signals without
modifying the cycle length or phase order. As there is
no general solution found yet, we define an approach
where traffic lights are the only active system
components that self adapt and communicate each
other in a distributed manner. We cover the
exchanged message definition required by TACTICS,
in order to change green times to control traffic
movements in a traffic network.
II. STATE OF THE ART
Much work was carried in the area of intelligent
transportation systems. From a theoretical point of
view most of the traffic theory was based on the
background of ideal fluids, at most taking into
consideration the compression properties [7]. All
these approaches have major problems when applied
to real-life traffic, or otherwise stated: real road traffic
is neither an ideal fluid nor it behaves like one.
In the last years, the mathematical models for
road traffic simulation have been improved. Most of
the classical models, inspired by gas or fluid behavior
in pipes give non-realistic results in modern traffic
situations and are considered inappropriate [8], but in
the last decade we witness a refactoring of these
models and implementation in simulation tools [9].
Responsible for this effect is the nonlinear and chaotic
character of the systems that describe road traffic, the
so-called: ”butterfly effect” [8]. The slightest changes
in traffic conditions on a road upstream the point of
observation induces effects and current models are not
able to give accurate “what-if” simulations.
For these systems, primary data is represented by
the number of vehicles passing on a road segment
over a given time period (possibly also the
distribution by categories: cars, trucks, bicycles,
pedestrians etc) and the average speed on that given
segment of road at any given time of day and any
given day of week [misra2011global]. Additional data
can be represented by the average acceleration and
deceleration when entering and exiting the road and
even the statistical distribution of the weight of the
vehicles and the number of traffic incidents/accidents.
The problem of improving the capacity of the
existing transportation infrastructure was previously
addressed from applying the mathematical models
presented above to the evolution of control rules to
improve system structure and reduce the complexity
of city topology [11]. In [9, 10] we can see solutions
designed for identifying the critical areas in an
existing topology or to predict problems in a proposed
one and to perform the simulation and validation
(finding the maximum traffic capability) of any
particular intersections or road segments. But these
approaches require a framework for the
implementation of the proposed methodologies.
An adaptive traffic control framework is
addressed in [12] and it is used in case of an
emergency large scale evacuation. The authors use a
methodology based on a model reference adaptive
control (MRAC) framework to serve their scope.
The field of Cyber-Physical Systems (CPS)
emerged in 2006, integrates the fields of computation
and controlling of physical entities. Opposed to
traditional embedded systems, CPS is typically
designed as a network of interacting elements with
physical input and output instead of as standalone
devices. The notion is closely related to concepts of
sensor networks. Complex, distributed and dynamic
systems like the ones providing air and road traffic
control and smart cities have been discussed in the
CPS community, concluding the need for an inter-
disciplinary combination of diverse engineering
fields. Several goals and requirements in large-scale
CPS have been identified so far, concurrency, real-
time capability, distributed control, self-adaption, self-
organization, reliability and fault tolerance [13].
Classical engineered solutions focus on
centralized approaches relying on global information,
but they lack the dynamic dependencies, which make
them easy to understand and manage. Centralized
approaches, however, assume that collecting data and
its processing meet real-time requirements. In large
and complex systems, this period of collecting and
processing data is longer than entities can wait for a
response. Traffic in large road networks is one
example of a situation where centralized optimization
is almost impossible: continuously collecting dynamic
traffic information from all roads, optimizing traffic
flows takes too long to be practically deployed in real
world networks. New approaches must at least self-
adapt to changing demand and loads in the network to
route vehicles to their destinations [13].
Self-organization implies previously described
self-adaption and also explores new strategies to reach
other objectives. Physical environments and
conditions may change frequently, requiring methods
that detect changes without external request or
modification. As a main desiderate for any system is a
high reliability and an increased fault tolerance. CPS
brings together specific engineering methods and
computer science research on embedded systems,
scheduling and distributed algorithms, emphasizing
the mapping of processes and physical features. A
good example of CPS domain is the control of vehicle
flows with the goal of reducing congestion and travel
times in a road network.
III. PROPOSED SOLUTION
A. TACTICS Framework
In [6] the authors propose a three layered traffic
system control stack, from which they have described
the methodology that runs at the first layer. Briefly,
22
their method consists in several steps that use an
adaptive mechanism to modify green time values to
improve local conditions for a single intersection.
A1. Deployment
In this context, we consider each intersection as part
of a higher complexity structure, a network in which
intersections communicate to each other to find a
global traffic optimum. Because we cannot decouple
local intersection’s behavior from the entire network,
we propose to interconnect the ones identified as the
central loading points in terms of traffic load. In [14]
the authors proposed the methodology for selecting
key nodes that will act in master-slave configuration
to reach correlated decisions using a communication
mechanism over the network. Complex Network
Analysis is used over the entire network and mark
nodes with highest betweenness [15] as master nodes.
Traffic data collection falls outside the scope of this
paper and according to [6] it is a layer 1 specific
operation. Selecting key nodes in the traffic network
is an operation specific for layer 2 and is directly
related to the proposed framework; because it selects
the nodes that will constitute the so called Intersection
Control Unit, see Fig 1.
Using the three layered optimization stack we
define the communication procedure and the specific
messages that define the upper layer of the stack. This
third and last step is responsible for the system’s
response and adaption to continuous traffic changes.
Each node uniquely identified by a traffic light will be
dynamically controlled to act as a traffic officer.
Fig. 1. Traffic network for a city using TACTICS understanding
Our proposed framework defines the physical
implementation of the three layered stack proposed in
[6]. The first layer runs local adaption mechanisms
that change green time values at intersection level
based on the detected traffic flow. But, running this
algorithm on each intersection is not an optimal
solution because of the high number of intersections
in a city. The layout of this framework can use the
algorithm described in [16] to deploy the system in a
real world situation. Because local intersection’s
behavior must be seen as part of a traffic network,
central loading points in terms of traffic load must be
selected. STiLO methodology [14] identifies “hot
points” and selects the relevant to work in master-
slave configuration to reach correlated decisions.
TACTICS implements the characteristics of a
cyber-physical system to create a fault tolerant
framework for the adaptive control of traffic
movements. This system consists in several
customized Intersection Controller Units; each of
them handles an entire intersection, covering all the
signal controllers in that physical location. For each
direction a Queue Detector (QD) is installed to
determine the queue length for that specific direction.
Their results act as input for each Signal Controller
(SC) which is responsible for the new green time
changes. All the SCs in the intersection are
interconnected (Wireless or not) creating the so called
Intersection Controller Unit (ICU), see Fig 2. This is
responsible for the behavior and the adaption of the
entire intersection to traffic changes. Any city, or
large portions of it, can be reduced to several
independent ICUs which are all interconnected, but
with no centralized control center. On each of these
units, STiLO methodology is applied to define if it is
running in a master or a slave configuration.
Fig 2. Intersection Controller Unit (ICU)
Fig 3 shows the working flow diagram for each
ICU. Literature gives different solutions for real
traffic data gathering [17], such as license plate
recognition to roadside sensors that log in real time
traffic data. Each QD reads the queue length using of-
the-shelf car detectors and classification tools.
Otherwise, a hardware module capable of estimating
the length and dynamics of a queue must be
implemented and used for queue detection. Data
collected is feed into the Traffic Data Acquisition
System which creates the modified Origin Destination
table and the traffic/flow matrix of the intersection.
The literature gives us different solutions for real
traffic data gathering [4, 17, 18], ranging from license
plate recognition to roadside sensors that log in real
time traffic data. For our proposed framework we
have decided to use the video data collection
mechanism, mainly for its ease of deployment.
Using the formulas described in [6] these
structures provide input for the Adjustment
Mechanism working at the SC level. These
computations lead to the new set of green times. The
new computed values along with the parameters and
messages are ready to be sent to the interconnected
intersections via Communication Controller. The
Feedback Controller also receives these values and it
decides to wait or not for an external response. The
Communication Controller is responsible for sending
the messages to the interconnected intersections and
also receiving the corresponding responses. These are
parsed and sent to the Feedback Controller which will
23
take them into consideration or not before setting the
new green times in the ICU.
One can see that the Communication Controller
could be missing and in this case the adjustment
works only at intersection level. This happens if the
intersection that is being optimized is isolated and it
works as standalone or if the communication is
offline. This framework uses no redundancy since it
can work offline without any centralized control. If
the master nodes are to implement the hardware
redundancy it will be a cost increase in order to
protect of a failure that is not a real threat to the
system, since each signal controller can take the role
of ICU. Several solutions are to be further studied,
like the need of a failure detection module can be
implemented to monitor the state of ICU.
Fig 3. Functional block diagram of an ICU of TACTICS
TACTICS implements the three layered
optimization stack in [6], the communication
procedure and the specific messages that are defined
so that the system responds and adapts to continuous
traffic changes. Each node uniquely identified by a
traffic light is dynamically controlled to act as a
virtual traffic officer. For this framework to be
operational, the network topology will have to be
defined at deployment time. A procedure for a new
node insertion, corresponding to a new traffic signal
installation is needed to be defined. Using this
mechanism, each node is capable of positioning itself
into the network, by knowing his neighbors and it is
able to find its role. STiLO must be run for the new
deployed node to determine its role in the network.
The adaptive green time mechanism is the core of
this algorithm, because it is determines and sends the
new green times to the traffic signals operating in
intersections. The dynamic of each traffic light-
controlled intersection is defined using a set of only
three parameters and new green time values are
derived based on their values. These are, green time
value, meaning the time which allows traffic to flow
through an intersection, traffic flow, representing the
number of vehicles passing on a specific direction and
cycle length, which is the timeframe between two
consecutive green times.
Several steps are performed for changing traffic
signal timings. First step is to determine whether a
local intersection has a problem in managing passing
traffic flow through it. Next step is to determine if it is
possible to make changes locally or not, based on the
input values read. If the intersection can respond to
traffic changes by changing its own green time values
then it will determine the changing coefficient that
will be sent to the interconnected ones. In case the
current intersection is identified using STiLO as
master than it communicates to the slaves the changes
made on the impacted directions. It also notifies the
other interconnected masters about the changes. The
greenTimeIncrease and the coefficient_level are
computed and sent to the connected intersection. The
response is expected during the same cycle in to know
if changes are accepted or not. The algorithm starts
over and reads traffic data after each cycle is over.
Depending on the desired goal, different sets of
parameters can be selected as input data; similar to
vehicle to infrastructure, V2I, or infrastructure to
vehicle, I2V, which use physics parameters (speed,
acceleration). These cover the behavior of any
intersection and provide all the information needed to
assess new timing plans. Due to reduced number of
operations this will need low computational power. In
a real-world system, measuring and collecting data
traffic values still represents a challenge.
A.2. Adapting Green Time Values
The adaptive green time mechanism is the core of this
algorithm, because it is responsible for effectively
determine and send the new green times to the traffic
signals operating in intersections. We start by defining
the dynamic of each traffic light-controlled
intersection using a set of only three parameters and
we will derive new traffic signals based on their
values. These are, green time value (Gt), meaning the
time which allows traffic to flow through an
intersection, traffic flow (td), representing the number
of vehicles passing on a specific direction and cycle
length (Cl), which is the timeframe between two
consecutive green times.
Several steps must be performed in order to
change traffic signal timings. First is to determine if
the local intersection has a problem in managing
passing traffic flow. Next is to determine if it is
possible for it to make changes locally, based on the
input values read and it will compute the changing
coefficient that will be sending to the interconnected
ones. If the current intersection was identified by the
algorithm as a master than it will communicate to the
slaves the changes made on the impacted directions
and also will notify the other interconnected masters
about the changes. As the results are sent, a response
is expected during the same cycle in order to know if
changes were made or not. The algorithm restarts and
reads traffic data on each cycle.
B. Inter Traffic Signal Communication
As for reading and computing new green times the
methodology was described earlier, it is the
communication part that we will detail in this part.
We define two types of messages: requests and
reports, to be exchanged between master and slave
intersections. Their format is defined in Fig 4 and has
24
a minimal format in order to be easily implemented
regarding the transmission method used (TCP/IP,
Bluetooth etc).
Message ID Message Type Source Target Payload
Fig 4. Message format used by TACTICS
Based on the resulting coefficient values and on
the adaptive green time methodology, six Message
IDs are defined: REQ_INC_LOW, REQ_INC_HIGH,
REQ_DEC_LOW, REQ_DEC_HIGH, REP_YES,
REP_NO and an optional ACK can also be used, but
this depends on each intersection load.
REQ_INC_LOW and REQ_INC_HIGH each
correspond to a request for increasing the green time
value with low or high coefficient as described in
[14]. The same applies for REQ_DEC_LOW and
REQ_DEC_HIGH where they represent a request for
decreasing the green time values. REP_YES and
REP_NO are the reports sent by the slave intersection
as an answer to each of the before mentioned requests.
A bidirectional communication is proposed to
exchange information using a simple request-reply
report, where each intersection notifies the
interconnected one about the changes that is going to
perform. Each intersection will also take into
consideration the incoming requests if its local
conditions permit it. When the other intersection
acknowledges the message, it means that the
information will be used for the next timing
adjustments and a negative answer means the
information cannot be used because of the already
calculated green times. Time aspect is important
because there is no synchronization of traffic signals.
The main target of the proposed framework is to
assure the environment for traffic optimization
process in order to ensure a continuous traffic flow
between key intersections inside an urban traffic
network. Each intersection is seen either as a
standalone entity or part of a complex network
described by three parameters: green times, traffic
flow and cycle lengths. By correlating intersections
and interconnecting nodes to operate in synergy,
faster flow will be achieved at network level.
Several cases are identified: one is when the
green time of the slave intersections overlaps the
master green time value and the second is the case
when the response from the slave is received during
maser's green time. In the first case the request from
the master is not reaching the slave in the current
cycle which means no response from the slave. This is
the specific case in which the master will adapt its
green time without any change from the slave. The
adaption from the slave will take place in the next
cycle following the response to master.
Each semaphore has its own working time: cycle
length, number of phases, changing order and the list
could continue. Because of this aspect, rules must be
described, so the communication between the
intersections is optimal and also to avoid unnecessary
overhead inside ICU. All computations are done
during the first red time period after a cycle is
completed. In this interval, the new green times and
coefficient levels are determined based on each
specific methodology. All other requests coming from
slave intersections in the next period will be taken
into account only in the next cycle.
Another rule is that no answer is kept more than
one cycle. When the request from the master is not
reaching the slave, because of a larger cycle length
and in this case, the master is always changing its
values and sending new requests until it gets a
response. If the communication is lost, each
intersection acts as master without sending any
message. Statistically, acting as master an intersection
could improve locally for short time and because any
congestion is limited in time it could cover the time
needed to pass that situation.
IV. CASE STUDY
The case study follows the changes made in the
system before the framework implementation and
after. An indicator of the improvements in the
network will be the time a queue is decreased, with no
adaption and using the proposed adaptive framework
control system. The proposed methodology finds the
optimal traffic balance for all directions in a single
intersection and communicates its results with the
interconnected ones in order to achieve a more
balanced network. But, continuous recalculation will
naturally lead to a point in time when adapting green
times is not possible anymore.
The proposed working model was evaluated
using the VISSIM simulator, a microscopic
simulation tool that provides conditions for testing
different traffic scenarios in a realistic manner. With
VISSIM, the urban network was defined around the
central part of Timisoara city and it simulated several
groups of traffic lights working using TACTICS
framework configuration.
Results present several traffic controlled
intersections, subject to the adaptive traffic signal
control, all in central area of Timisoara. Using
VISSIM, specific queue counters were set on each
direction to monitor traffic flow. These counters
record traffic data passing through during simulation
time. Two parameters are of specific interest: average
queue and maximum queue length. One central
intersection adapts its green time phases dynamically,
according to the described methodology. Traffic
values are injected into the urban network using
VISSIM specific traffic data zone generators. During
simulation, green times were adapted with five and
ten time units, increasing green time for the directions
heading north and decreasing south heading direction.
To determine the impact over one of the studied
intersections, traffic conditions were measured on all
four exits, recording values before and after adaption
of green times. The results show improvements at
local intersection level for the intersection that adapts
signal timings. Compared with the initial value, there
25
are moments in time when the improvements reach
almost 40% percent for the Average Queue Length,
see Fig 5 and Fig 6. This parameter describes a more
dynamic intersection, with shorter waiting times.
Meanwhile, the Maximum Queue Length parameter
shows an interest aspect when it reduces the pick the
value, fact that is caused by the progressive response
to the increasing traffic conditions.
Fig 5. Queue Length for one intersection VISSIM simulation results
Fig 6. Maximum Queue Length for one intersection VISSIM
simulation results
V. CONCLUSIONS AND FUTURE WORK
In this paper we proposed and tested in simulation an
adaptive traffic control framework, designed to
respond to dynamic changes in traffic conditions by
using intelligent traffic signaling. We described our
approach to be an efficient one in terms of new
hardware required and communication overhead
needed. Because it requires only a new module per
intersection and it uses current infrastructure without
any additional pavement installed sensors.
TACTICS is designed to interact with already
installed traffic monitoring ITS technologies and
proposes a self adapting methodology, without any
centralized control using a low message overhead for
each intersection due to its small number of
exchanged messages. The results presented in the case
study, show also low message overhead which makes
this framework an energy efficient one.
The cost for the new hardware installed in each
intersection is estimated to be around 12.000 Euros
based on our calculation. This certifies that this
solution is a low cost one compared to the costs of
installing an intelligent solution for an intersection,
which usually reach 30.000 - 40.000 Euros.
ACKNOWLEDGMENT
This work was partially supported by the strategic
grant POSDRU/159/1.5/S/137070 (2014) of the
Ministry of National Education, Romania, co-
financed by the European Social Fund – Investing in
People, within the Sectoral Operational Programme
Human Resources Development 2007-2013
REFERENCES
[1] K. Fehon, „Adaptive Traffic Signals, Are we missing the
boat?,” in ITE District 6, Annual Meeting, Sacramento, 2004.
[2] A. Stevanovic, „Review of Adaptive Traffic Control Principles
and Deployments in Larger Cities,” in International Scientific
Conference on Mobility and Transport, Munich, 2009. [3] O. Juhlin, „Traffic behaviour as social interaction-implications
for the design of artificial drivers,” in Proceedings of 6th World
Congress on Intelligent Transport Systems (ITS), Toronto, 1999.
[4] A. Stevanovic, „Adaptive Traffic Control Systems: Domestic
and Foreign State of Practice A Synthesis of Highway Practice –
Advanced Transportation Concepts”. [5] Warberg, Andreas and Larsen, Jesper and Jorgensen, Rene
Munk, "Green wave traffic optimization-a survey", Informatics and
Mathematical Modelling (2008).
[6] C. Cosariu, L. Prodan and M. Vladutiu, „Toward traffic
movement optimization using adaptive inter-traffic signaling,” in IEEE 14th International Symposium on Computational Intelligence
and Informatics (CINTI), Budapest, 2013.
[7] Papageorgiou, Markos and Diakaki, Christina and Dinopoulou,
Vaya and Kotsialos, Apostolos and Wang, Yibing, "Review of road
traffic control strategies", Proceedings of the IEEE (2003), 2043--
2067. [8] Daganzo, Carlos F, "Requiem for second-order fluid
approximations of traffic flow", Transportation Research Part B:
Methodological (1995), 277--286.
[9] Aw, A and Rascle, Michel, "Resurrection of" second order"
models of traffic flow", SIAM journal on applied mathematics
(2000), 916--938. [10] Bernot, Marc and Caselles, Vicent and Morel, Jean-Michel,
"Optimal transportation networks: models and theory", Springer
Verlag (2009).
[11] Montana, David J. and Czerwinski, Steven, "Evolving Control
Laws for a Network of Traffic Signals", MIT Press (1996), 333--
338.
[12] Zhou, Binbin and Cao, Jiannong and Zeng, Xiaoqin and Wu,
Hejun, "Adaptive traffic light control in wireless sensor network-
based intelligent transportation system" (2010), 1--5. [13] Senge, S. and Wedde, H.F., "Bee-Inpired Road Traffic Control
as an Example of Swarm Intelligence in Cyber-Physical Systems"
(2012), 258-265.
[14] Iovanovici, Alexandru and Cosariu Cristian and Prodan,
Lucian and Vladutiu, Mircea, "A Hierachical approach in
Deploying Traffic Light based on Complex Network Analysis" (2014), 232--237.
[15] Rami Puzis and Yaniv Altshuler and Yuval Elovici and
Shlomo Bekhor, "Augmented betweenness centrality for
environmentally-aware traffic monitoring in transportation
networks".
[16] Iovanovici, Alexandru and Topirceanu, Alexandru and
Cosariu, Cristian and Udrescu, Mihai and Prodan, Lucian and
Vladutiu, Mircea, "Heuristic Optimization of Wireless Sensor
Networks using Social Network Analysis" (2014). [17] Iovanovici, Alexandru and Prodan, Lucian and Vladutiu,
Mircea, "Collaborative environment for road traffic monitoring"
(2013), 232--237.
[18] Kevin Fehon, PE and Principal, DKS, "Adaptive Traffic
Buletinul Ştiinţific al Universităţii Politehnica Timişoara
TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
Volume 60(74), Issue 1, 2015
Performance of Turbo Encoders with 64-QAM
Modulators Interfacing Systems in Fading Environment
Maria Kovaci1 Horia Balta
1,2
1 Faculty of Electronics and Telecommunications, Communications Dept.
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail: [email protected] 2 Valahia University of Targoviste, 2 Avenue King Carol I, 130024, Romania, e-mail: [email protected]
Abstract – This paper presents a study on the
interfacing between the turbo encoder and modulator.
The binary allocation of the bits from a turbo coded
symbol towards the modulator symbol can be done in
several ways. This study shows the performance of the
allocation modes taking into account the quadrature
amplitude modulation with 64 points and the Rice
fluctuating transmission channel. The simulations
presented show that the performance of the entire
transmission system, measured in coding gain may be
influenced by up to 1 dB by a suitable choice of the
standards. Under its different variants, QAM is used
in digital cable television or wireless and cellular
technology applications. The 64-QAM is a good
compromise between spectral efficiency (6 bit/s/Hz)
and performance of bit/frame error rate (B/FER)
versus signal to noise ratio (SNR) [1]. 64-QAM gives
a symbol error rate of 10-6
for a SNR of about 19 dB
for uncoded system in non-fluctuating channel (i.e.,
Additive White Gaussian Noise channel – AWGN
channel) and, practically, it cannot be used in fading
channel. However, using a turbo code, a BER of 10-10
can be obtained at a SNR of 9 dB for the AWGN
channel and at SNR of 13 dB for the pure fluctuant
channel (Rayleigh channel). Obviously, the
advantages are the spectral efficiency and the
simplicity of the implementation. For these reasons,
the square 64-QAM is the most frequently digital
modulation encountered in applications. For example,
in LTE is specified that such modulation techniques
with Gray allocation can be used to minimize the
BER [2].
Of course, there are also disadvantages. One of them
is that constellations with QAM modulations Gray
allocation does not protect equally all the bits of the
modulator symbol. Neither the 64-QAM modulation
constellations is no exception to this. The problem
arising is to find the binary allocation variant between
the coded symbol and the modulator symbol which
optimizes the performance. Our previous studies have
been dedicated to this question for QAM
constellations [3], [4], [5], in AWGN channel. In the
present paper we study the turbo coded bit allocation
for the 64-QAM constellations in Rice fading
environment. A similar study, for 16-QAM was done
in [6]. In this study we used both the double binary
turbo code (DBTC) of the DVB-RCS2 standard [7]
and the single binary turbo code (SBTC) of the LTE
standard [2].
The Rice channel to which we referred above is a
model for the real channels in which the received
signal is a mixture between the direct wave (Line of
Sight– LOS) which is propagated directly from the
transmitter to the receiver and the waves reflected by
different objects.
In this paper, as in [5], we have analysed three
locations for the placement of the information and
parity bits generated by turbo coding in the symbol
modulator. In the first case the information bit was
placed in the best protected position, followed by two
parity bits placed in less protected positions. In the
second case the information bit is placed on the
middle position, so that in the better and less protected
positions are placed the parity bits. Finally, in the
third case, the information bit appears on the poorly
protected position. The results of simulations show a
completely different behaviour in the performance of
B/FER vs SNR of these allocation variants.
The structure of this work is organized as
follows. In Section II are presented the turbo encoders
used in this paper (single binary - SBTE and double
binary - DBTE) in order to identify the bits to be
allocated in the symbol modulator. Section III briefly
describes the square 64-QAM with the same aim to
identify positions from the modulator symbol that will
be filled by turbo encoded bits, nominated previously.
Section IV is dedicated to presenting allocation
alternatives. Section V shows the simulation results
and Section VI concludes the paper.
27
Fig. 1. The scheme of the SBTE.
II. THE TURBO ENCODER
The direct coupling between the turbo encoder and the
modulator supposes the representation of the turbo
coded block under a periodical structure form, with a
period equal to the modulator symbol length. The
structure of a turbo coded block is influenced by the
structure of the turbo encoder and puncturing matrix.
This section describes the SBTE specified in [2] and
the DBTE specified in [7], configured for the coding
rate 1/3 and 2/3, respectively.
A. Single binary turbo encoder
Fig. 1 shows the structure of a SBTE. Input sequence
u is encoded directly by the convolutional encoder C1
and via interleaver (π) by the encoder C0. Depending
on the requirements, the outputs of the two
convolutional encoders are punctured to obtain higher
coding rate. It follows redundant sequences x0 and x1,
which, along with the original information sequence
u=x2 form SBTE's output. In the absence of
puncturation, the (natural) encoding rate of SBTE's is
1/3. At this rate, the turbo coded block size is 3×NS
where NS is the length of interleaving. In other words,
one turbo coded block consists of NS symbols of the
form xj=(jjj
xxx012
,, ), with j from 0 to NS-1.
B. Double binary turbo encoder
Fig. 2 shows the scheme of a DBTE. Unlike SBTE, a
DBTE generates a four-bit symbols xj=( jjjjxxxx 0123 ,,, )
at its natural rate 1/2. In this case the size of a turbo
coded block is 4×ND where ND is the length of inter-
symbol interleaving. Note that DBTE performs both
the inter-symbol interleaving (information symbols
are interleaved) and the intra-symbol interleaving (the
bits from information symbol are interleaved).
Fig. 2. The scheme of the DBTE.
Because the modulator symbol for 64-QAM contains
6 bits, three for each carrier, for compatibility, we
chose to use the coding rate 2/3. To obtain the coding rate 2/3 for DBTE. we have used
the punctured matrix:
=
10
01pdM , (1)
which also applies to sequences x1 and x0. The
structure of a turbo coded block is of the form:
... j
x3 , 13
+jx , ...
... jx2 , 1
2+j
x , ...
... j
x1 , , ...
... , 10
+jx , ...
with j from 0 to ( ) 12 −DN .
Thus, in both cases (SBTE with coding rate 1/3 and
DBTE with coding rate 2/3) we have obtained a
periodic structure of the data block of 3 or 2×3 bits. These triplets of bits will form the modulator symbol for 64-QAM, symbol of 6 bits, as shown in the next
section.
III. THE SQUARED 64-QAM
The constellations for 64-QAM square modulation is
presented in Fig. 3. A signal modulated using squared
[3] R. Watson, “Understanding the IEEE 802.11ac Wi‐Fi Standard – Preparing for the next gen of WLAN”, http://www.merunetworks.com/collateral/white-papers/wp-ieee-
Buletinul Ştiinţific al Universităţii Politehnica Timişoara
TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
Volume 60(74), Issue 1, 2015
Digital Rights Management - Creative Commons
Perspective
Cristina Vasilescu1, Mihai Onița
2
1 Faculty of Communication Sciences, Communication, Public Relations and Digital Media Str. Traian Lalescu Nr. 2a 300223 Timisoara, Romania, e-mail [email protected] 2 Faculty of Electronics and Telecommunications, Communications Dept.
Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail [email protected]
Abstract - This paper is addressed to an area with a
significant development in recent years: Digital Rights
Management (DRM). These data copyright can be
applied to several types of digital materials as images,
audio recordings, videos, and text. To be more specific,
we present in the paper, Creative Commons (CC)
technology, as an alternative to classical DRM. We
bring in discussion layers and types of a CC license, and
we include a study case of most popular platforms under
CC license. We make some recommendations and
extract some conclusions.
Keywords: DRM, Creative Commons, Public License,
CC platform, video, audio, text
I. INTRODUCTION
According to the Romanian Copyright Office,
Copyright is a legal term that it recognizes rights of
creators of literary, scientific or any work of
intellectual creation. Digital Rights Management
(DRM) is an intellectual property right that the
authors have over their creations. By creation,
researchers refer to any material: photos, audio
recordings, videos, written materials (text), etc. These
rights represent a method of protection recognized by
law, and they apply to everyone, regardless of status,
education, race or religion [1]. The Romanian law, for
example, gives the author the right to authorize or
prohibit (quoted from the Law) [1]:
• Reproduction of work, distribution of work;
• Commercialization of copies with author approval;
• Renting work, loan work;
• Public communication of the creation directly or
indirectly;
• Broadcasting the work;
• Cable retransmission of the work;
• Making derivative works;
These are the rights (patrimonial rights) that the law
recognizes the author. Of course, there are some
exceptions, but not major. Copyrights apply to
published materials and unpublished materials,
finished or unfinished. The material is recognized and
protected by the simple fact of its implementation,
even if it was not brought to the public attention [2].
Digital Rights Management is connected with systems
that restrict access to the digital media space. It is a
technology used by content providers to control the
usage and distribution of images, digital music, video
or files [3]. DRM fights against illegal modification,
copying, viewing or distribution/distributing of digital
media materials. Some of the copyright holders argue
that DRM handles large losses due to illegal
distribution of copyrighted material.
The DRM system is designed to adjust the
dissemination of digital information for following
types of digital materials: video, music, audio,
electronic books, software, video games. The
technology associated with DRM is intended to
provide the seller control over digital content or
devices after they have been entrusted to the buyer.
Content owners may use different types of DRM to
protect their intellectual property [4]:
• Restrictive Licensing Agreement controls access to
digital materials, copyright, public areas, etc.;
• Encryption (Encryption);
• Scrambling control online information access and
reproduction (e.g. backup copies for personal use);
• Digital signatures - provides secure content and
allows secure transactions;
• Fingerprint/watermarking incorporating information
about ownership to facilitate tracking and
monitoring the use, copying and distribution [5].
II. ALTERNATIVES
Open licenses are those materials considered to be
implicit protected by law and provide access to the
work that can be reused and redistributed [4]. Creative
Commons is a global non-governmental organization
dedicated to supporting a free and open Internet,
enriched through free knowledge and creative
resources so that people everywhere can use them,
distribute and develop [6].
39
Fig. 1. Layers of a CC license [6]
All Creative Commons product licenses have
common features. Any license helps creators (referred
here as licensors) retain their copyright while
allowing others to copy, distribute or use their
contents. Licenses incorporate an innovative design
with a structure composed of three layers: Legal
Code, Human Readable, and Machine Readable
(Fig.1). This organization has four types of items that
may constitute the type of license required [7]:
Attribution: people using the material must give credit
to the author.
Noncommercial: Individuals are not allowed to
distribute, modify or re-use the material if the purpose
is a commercial advantage or monetary compensation.
No derivatives: The material can be distributed, but
must be kept in original form without modification.
Share Alike: The adapted or modified material should
be distributed under the same Creative Commons
license
Fig. 2 reveal the possible combination of CC licenses:
Fig. 2. Types of CC licenses [4]
Attribution CC BY - this type of license allows others
to share, remix, modify/add to the original work as
long as credit is given for the original work. This type
of license is one of the most convenient services of
this kind offered by Creative Commons (CC).
Attribution NoDerivs CC BY ND - allows
redistribution (for commercial or non-commercial
purposes)with the condition that the content is not
altered.
Attribution-NonCommercial Share Alike CC BY NC
SA – allows others to remix, add or remove parts
from the non-commercial material with the condition
to recognize the source and to license the new content
respecting the same terms.
Attribution-Share Alike CC BY SA - offers the
opportunity to remix, modify or add to the content
(even commercial usage). The procedure has to be as
described above at other licenses. CC BY SA is often
compared to open source software licenses. Any
derivation from the original work will carry the same
license. This type of license is used by Wikipedia and
is recommended for Wikipedia materials that allow
improvements or additions or may be used in similar
projects.
Attribution Non-Commercial CC BY NC refers to
non-commercial materials that can be remixed,
modified, updated without the need for additional
licenses for the resulted content.
Attribution Non Commercial No Derivs CC BY NC
ND is the most restrictive of all licenses, allowing
others only to download and share the content as it is,
with the condition that they acknowledge the source,
without being able to make changes or use for
commercial purposes [7].
III. CASE STUDY - CC LICENSED PLATFORMS
There are a series of platforms, online applications
that have collections of images, music, videos and
documents that can be reused under certain
restrictions related to copyrights. These can be
divided into four categories, namely: an online
database of images, an online database for audio-
video materials, an online database of texts, and
online database for multimedia searching applications.
In the current study, we have identified those under
Creative Commons (CC), cataloged with Alexa
ranking and briefly described them.
Table 1
Application Domain Alexa
Rank
Flickr Images 130
Google images Images 2.587.437
Pixabay Images 1.040
Fotopedia Images 169.411
Open clipart Images 18.964
Instagram Images 34
Kepguru Images 265.400
Gorgraph Images 57.189
Creativity 103 Images 349.467
Deviant Art Images 160
Jamendo Audio- video 20.233
ccMixter Audio 62.954
Free sound Audio 12.868
Sound cloud Audio 176
Tribe of noise Audio 1.436.736
Europeana Audio- video 53.996
Youtube Audio- video 3
Blip tv Audio- video 14.975
Vimeo Audio- video 172
40
Wisdom Commons Text 447.235
Travellers point Text 39.525
Intra text Text 336.002
Creative Commons General
content 3899
Internet Archive General
content 234
Freebase General
content 1.740.431
Wikipedia
Commons
General
content 207
A. Images
Flickr, www.flickr.com is a site that hosts photos and
videos. It enjoys great popularity among bloggers that
store a lot of pictures for later use distributing them. It
can also be used to a mobile phone or with a computer
[8].
Google image, https://images.google.com is a search
and storage platform for images that allows users to
search the Web for image content. Keywords for the
image search are based on the image's file name.
When an image is sought, it displays a thumbnail.
When the user accesses the image, it is displayed in a
box on the website belongs to. The user can close the
image and can browse the web, or view the full image
in various sizes [9].
Pixabay, http://pixabay.com is a site that provides
access to a database of high-quality images under free
licenses. The images can be distributed and used
without any restriction because they are shown under
Creative Commons CCO dedicated to the public
domain. Images can be copied, modified, distributed,
and even used for commercial purposes without the
need for permission or without having to pay for
them. There is still the possibility that what is found
in these pictures to be under the protection of
trademarks or because of private rights [15].
Fotopedia, http://www.fotopedia.com was created by
five former Apple employees and represents a
database for images of photographers and authors
who have entered a form of cooperation. The
collaborators names have attached a hyperlink directly
related to their personal website where you can find
the entire gallery with high-quality pictures on various
topics from around the world. Unfortunately, in July
of 2014, Fotopedia management announced its
cessation asking users to store their data in personal
computers because if they did not, they would lose all
materials stored on the company server.
Open clipart, https://openclipart.org is a digital media
community that can store vector clip creations under a
free license. The project started in early 2004 by the
Inkscape developers desiring to collect specimens of
flags from around the world. It had a positive
development therefore objectives were extended to
generic clipart.
Instagram, www.instagram.com is a fun and different
way to share life with friends through a series of
images. It was created from the desire to allow the
sharing of life events through images as close to the
time they occur. The application was named from a
combination of two words: instant and telegram.
Kepguru, http://kepguru.hu is an online application,
launched in Hungary that became very popular. To
upload images is required an email address, username,
password, and the users consent to the rules imposed
by developers.
Gorgraph, www.geograph.org.uk at the moment of
launching had the main goal to collect, publish,
organize and archive the information or images
representative of Great Britain, Ireland and the Isle of
Man. Through this website was created access to a
geographic database freely available to the public. All
photographic observations are registered under a
Creative Commons Attribution-Share Alike license
granting those who access the site, rights to use the
materials for any purpose, as long as credit is given to
the copyright holder and that derivative works are
used under the same license.
Creativity 103, http://creativity103.com is a source of
photographic materials that has all sorts patterns and
textures, unusual and abstract; all available for free
under a Creative Commons licenses. It was released
in 2001 due to the lack of sites for people who wanted
to use textures and backgrounds in their projects. The
platform currently contains more than 2500 files, 6GB
of free photos. The downloads are designed to be used
directly in the drawings, as layer textures or as a
source of inspiration and ideas for further
development.
DeviantArt, www.deviantart.com is described in
Chapter IV.
B. Audio-video
Jamendo, www.jamendo.com is a music website and
an open community of music authors. It is an
economic model that allows free music downloads for
Internet users while providing revenue opportunities
for artists through commercial usage [11]. The name
"Jamendo" comes from the fusion of two musical
terms, i.e., "jam session" and "crescendo".
ccMixter, http://ccmixter.org is a website that offers
remixed music under Creative Commons. It provides
the possibility to download and listen to any type
music anywhere, anytime and with anyone. Some
songs may have certain restrictions, depending on the
applied licenses. The site supports popular formats
like MP3, WMV, OGG and others. Those who wish
to upload audio material on this site are advised to
archive their materials before sending them.
Free sound, www.freesound.org aims to create a
database of audio snippets, samples, and records
provided with Creative Commons licenses that allow
reuse. It provides new ways to access materials by
41
browsing using keywords; uploading and
downloading tons to and from the database under the
same Creative Common License; also offers the
ability to interact with other sound artists.
Sound cloud, https://soundcloud.com is the largest
social music platform in the world, where any user
can create sounds and can share them. Recording and
uploading sounds on this platform allow users to share
easily either privately with friends or on public blogs,
websites, and social networks. Also, sound creators
can use the platform to receive detailed statistics and
feedbacks from SoundCloud community. It can be
easily accessed via smartphone applications for
iPhone and Android.
Tribe of noise, www.tribeofnoise.com is an ever-
growing community that has at this moment 25,000
artists from 185 countries. It connects amateur
musicians with professionals from the media and
enterprises worldwide that need to provide music with
all rights included. Independent artists can preserve
their rights and at the same time, can take advantage
of the best collective business deals.
Europeana, www.europeana.eu is an Internet portal
that acts as an interface for books, paintings, films, art
objects and archival records that have been digitized
in Europe. These stored data on a single Internet
address allow users to explore Europe's cultural and
scientific heritage from early prehistory and until
today [12].
YouTube, www.youtube.com is a platform that allows
a large number of people to discover, watch and share
videos. It provides a forum for people to connect,
inform, but also to inspire others. You can find
videos, TV clips, music videos, and other content
such as video blogging, short original videos, and
educational videos. The access to this content is free
and can be made by any device as long as there is an
Internet connection [13].
BlipTv, www.blip.tv belongs to Studios Maker. It
develops, manufactures and distributes the best web
[17] Plagiarism.org, Definitions and types of plagiarism,
http://www.plagiarism.org, Accessed September 2014
[18] N. Helberger, N. Dufft Digital Rights Management and
Consumer Acceptability, A Multi-Disciplinary Discussion of
Consumer Concerns and Expectations, State-of-the-Art Report,
INDICARE project
44
Buletinul Ştiinţific al Universităţii Politehnica Timişoara
TRANSACTIONS on ELECTRONICS and COMMUNICATIONS
Volume 60(74), Issue 1, 2015
The detection of moving objects in video by background
subtraction using Dempster-Shafer theory
Oana Munteanu12
, Thierry Bouwmans2, El-Hadi Zahzah
2, Radu Vasiu
1
1 Faculty of Electronics and Telecommunications, Multimedia Dept. Bd. V. Parvan 2, 300223 Timisoara, Romania, e-mail: [email protected], [email protected] 2 Mathematics, Image and Applications Laboratory, University of La Rochelle