Matlab Simulation For Com

Conference of Informatics and Management Sciences
ICTIC 2013
March, 25. - 29. 2013
Matlab Simulation for Comparison of Different QAM

Constellations and Mapping for a Channel with
Additive White Gausian Noise
Miroslav Makarov
Department of Communication Equipment and Technologies
Technical University of Gabrovo
Gabrovo, Bulgaria
miro.makarov@yahoo.com
Abstract This paper presents a Matlab based simulation that
compares different QAM (Quadrature Amplitude Modulation)
constellations. It also allows to compare two constellations of the
same shape, but with different mapping of the bits. The
simulation includes seven main steps. The criterium for the
comparison is the Bit Error Rate (BER). Seven different
constellations are chosen for study.
Keywords- Matlab simulation; QAM; constellations; BER;
mapping
I.
INTRODUCTION
The goal is to realize a simulation technique, that compares

different QAM (Quadrature Amplitude Modulation)
constellations and mapping schemes using Matlab. The chosen
criterium is the Bit Error Rate (BER). That allows the most
suitable QAM, constellation and mapping for the situation to
be selected.
II.
MAIN SIMULATION STEPS [6]
A. Loading the input data for the simulation

The simulation gives the opportunity to choose the input
signal between a real signal (audio signal) or an random
generated bit stream. They are loaded into Matlab workspace.
The use of real audio file is prefered, because it is convenient
to hear the changes, introduced by the channel (changes in
frequency band, introduction of noise and so on)[5].
B. Perform ADC (Analog to Digital Conversion)
This step involves the Analog to Digital Conversion if an
real audio signal is selected as input data for the simulation.
C. Chose the QAM to be compared
The constellations may be: square QAM constellation,
performed by the qammod function in Matlab and custom
QAM constellation performed by the genqammod function[6].
For the custom mode the constellation points must be entered
point by point. It is convenient this way because it allows to
compare two or more constellations, that have the same shape,
but use different mapping.
The 2nd International Conference

http://www.ictic.sk
D. Add noise to the signal

This step involves addition of additive white Gaussian
noise to the modulated signal. In order to plot the BER curves a
loop is used, where the loop variable is the Signal to Noise
Ratio (SNR).
E. Demodulate the received QAM signal
In this step the demodulation of the received signal is
performed and is made separately for every different value of
the SNR.
F. Plotting the BER curves
This step involves bitwise comparison of the input and the
received bit streams and plot the BER curves.
G. Digital to Analog Conversion (DAC)
This step is required only if further research of the changes,
introduced by the channel will be made.
III.
CHOSING THE QAM ORDER, CONSTELLATION AND

MAPPING
Seven different 16 QAM constellations and mapping are

chosen [1],[2]. They are shown on the figures below. They are
as follows:
Fig.1 shows 16 QAM square constellation with points (4, 2,
-2, -4) and mapped according DVB-T standard [3],[4].
-4, -6) and mapped according DVB-T standard [3],[4].
-1, -3) .
Fig.4 shows 16 QAM circular constellation with two
amplitudes (1 and 3).
Fig.5 shows 16 QAM circular constellation with 2
amplitudes (3 and 5).
Fig.6 shows 16 QAM circular constellation with 4
amplitudes (1, 3, 5 and 7).
SECTION
7. Information and Communication Technology
- 297 -

ICTIC 2013
March, 25. - 29. 2013
Figure 1. 16 QAM square constellation with points (4, 2, -2, -4) and mapped
according DVB-T standard
Figure 3. 16 QAM square constellation with points (3, 1, -1, -3)
Figure 2. 16 QAM square constellation with points (6, 4, -4, -6) and mapped
according DVB-T standard
Figure 4. 16 QAM circular constellation with two amplitudes (1 and 3)

http://www.ictic.sk
SECTION
- 298 -

ICTIC 2013
March, 25. - 29. 2013
DVB-T standard; 5 corresponds to fig.4; 6 corresponds to

fig. 5 and 7 corresponds to fig. 6.
Figure 5. 16 QAM circular constellation with 2 amplitudes (3 and 5)
Figure 7. BER curves for the 7 different 16 QAM constellations and/or

mapping
Figure 6. 16 QAM circular constellation with 4 amplitudes (1, 3, 5 and 7)
IV.
RESULTS OF THE SIMULATION
The results of the simulations are given on figure 7. It

represents the BER behavior of the different 16 QAM
constellations. The BER curves are as follows: 1- corresponds
to fig. 1; 2 corresponds to fig. 2; 3 and 4 correspond to
fig.3, where 3 is Matlab mapped and 4 is mapped according the

http://www.ictic.sk
It shows the BER curves in the range from BER = 10-1 to

BER = 10-4.5. It can be seen that the most suitable constellation
for a channel with additive white Gaussian noise is the square
constellation with point coordinates {3; 1; -1; -3}, given on
figure 3. The mapping method does not influence considerably
the BER curves. The most unsuitable constellation is the Nonuniform square constellation with a=4. The constellation is
given on figure 2 and the BER curve is on figure 7, marked
with a 2. It can be seen from the results, that the non-uniform
square constellations (fig.1 and fig. 2) have worse BER
performance than the uniform constellation. The BER
deteriorates with the increase of the coefficient a. The
coefficient a corresponds to the module of the coordinates of
the inner points (i.e. for constellation points 4;2;-2;-4 a=2,
and for constellation points 6; 4; -4; -6 a=4) . The case of the
circular constellation with only 2 amplitudes is the opposite:
the constellation from figure 5 (amplitudes 3 and 5) has better
BER performance than the constellation from figure 4
(amplitudes 1 and 3). The constellation, given on figure 6 has
almost the same BER performance like the constellation on
figure 2. The BER performance of that constellation versus the
increase of the coefficient a will be further researched.
SECTION
- 299 -

ICTIC 2013
March, 25. - 29. 2013
V.
advisor Assoc. Prof. Nikolay Tashkov for the helpful

discussion.
CONCLUSIONS
In this paper a Matlab based simulation model is

proposed. The obtained BER vs. SNR curves are for seven
different constellations and/or mapping, all 16 QAM. From
the so plotted BER curves and the constellations it can be
clearly seen which one is the most suitable for a given case,
according to transmitted power, SNR and desired BER. The
so proposed Matlab simulation gives good opportunity this
comparison to be made.
REFERENCES
[1]
[2]
[3]
[4]
[5]
ACKNOWLEDGEMENT
The author wishes to thank the reviewers, whose
constructive comments and suggestions improved the
presentation of this paper. Thanks also to the authors PhD

http://www.ictic.sk
[6]
Webb, W., L. Hanzo. Modern quadrature amplitude modulation.

Principles and applications for fixed and wireless communications,
IEEE PRESS, 1994.
Proakis, J.G. Digital communication. IEEE Press London, 1988.
Konov, K. Digital radio and television broadcast, DIOS, Sofia, 2011.
DVB-T Standard ETSI EN 300 744 V1.6.1 (2009-01)
Tashkov, N., M. Makarov. BER Research for QAM Modulation.
International scientific conference UNITECH11, 18-19 November
2011, Gabrovo, Bulgaria, pp. 322-326.
MathWorks, www.mathworks.com, available: March, 2013
SECTION
- 300 -

Matlab Simulation For Com

Uploaded by

Copyright:

Available Formats

Matlab Simulation For Com

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Matlab Simulation For Com

Uploaded by

Copyright:

Available Formats

Conference of Informatics and Management Sciences

March, 25. - 29. 2013