SIMULATION OF A DIGITAL COMMUNICATION SYSTEM

Alpaslan Güngör1, Feza Arıkan2, Orhan Arıkan3

1
Türk Telekomünikasyon A.Ş., Aydınlıkevler, Ankara
+(90) 312 555 18 19, [email protected]
2
Department of Electrical and Electronical Engineering, Hacettepe University, Beytepe, Ankara
+(90) 312 297 70 95, [email protected]
2
Department of Electrical and Electronical Engineering, Bilkent University, Bilkent, Ankara
+(90) 312 290 12 57, [email protected]

ABSTRACT
In this paper, basic components of a digital source encoder-decoder uses Linear Predictive Coding
communication system are simulated by a computer (LPC). For the time being, there is no coding techniques
program. The simulation program is modular and for digital sources. In the channel encoder-decoder,
flexible to incorporate any future additions and Hamming, Bose-Chaudhuri-Hocquenghem (BCH) and
updates. The simulation program allows the user to Reed-Solomon (RS) error correction techniques are
choose from various channel models, transmitter and implemented. For the modulator/demodulator units,
receiver antenna systems, modulation and channel Binary Phase Shift Keying (BPSK), Coherent
coding techniques. A communication system is defined Frequency Shift Keying (CFSK) and Quadrature Phase
by various parameters including the source, coding, Shift Keying (QPSK) are implemented. The simulation
modulation, antenna systems. In order to facilitate the program allows the user to add other desired coding-
input of these parameters and follow the flow of the decoding techniques and modulation-demodulation
simulation, the Graphical User Interface (GUI) is techniques easily.
designed for convenience to the user. The input
parameters can both be entered from the GUI or from
prepared user files. The major contribution of this
simulation system to the existing communication
simulators is the addition of flexible antenna systems
both at the transmitting and receiving ends. With this
simulation program, the antenna arrays can be located
anywhere on Earth, on any platform and array elements
can be placed on the platform by any desired
orientation. The simulation program results are
compared with both theoretical computations and
commercial simulator results and excellent agreement Figure 1: A Generic Digital Communication System
is observed in both cases.
The antennas are transducers between free-space and
transmitter and receiver systems. The antennas used in
1. INTRODUCTION simulation program are narrowband antennas that are
In this study, basic components of a digital designed to communicate in VHF-UHF range. The
communication system are simulated with channel readily incorporated antennas can be listed as electric
models and noise by using flexible antenna systems dipole, magnetic dipole (loop), slot, rectangular
both at the transmitting and receiving ends. As shown in aperture, circular aperture, microstrip and horn
Figure 1, a digital communication system is made up of antennas. By using these antenna types, linear, planar
both analog and digital parts. The digital part consists of or desired configuration of the arrays can be constructed
digital source, source encoder-decoder, channel with the desired number of elements. The simulation
encoder-decoder and digital modulator-demodulator, program allows the user to position the antenna arrays
and the analog part consists of analog source, anywhere on Earth, on any platform. The antenna
transmitter and receiver antenna systems, channel elements can be placed on the platform independent of
models and noise models. The developed simulation each other with any desired orientation. Each antenna
program is capable of inputing both digital and analog can be fed by a different amplitude and phase. In the
sources. For analog inputs like speech signals, the simulation program, Additive White Gaussian Noise
(AWGN) channel and Rayleigh fading channel are
implemented. Due to a large number of various R m = r − rˆ ⋅ d m (3)
parameters to define the communication system, the
user can enter the parameters from previously prepared As a consequence, the phase difference between mth
files or from the screen by the help of a user friendly antenna and the reference antenna is equal to k0 rˆ ⋅ d m .
menu in Graphical User Interface (GUI). The simulation
program neglects the coupling between the antennas The total field of the array consists of M elements can
and near-Earth effects. These missing corrections can be be given by
easily incorporated to the existing program in the future. e − jk 0 r M
∑ Am F m (θ , φ )e 0 m e jβ m
jk rˆ ⋅ d (4)
E ( r ) = jk 0 Z 0
In order to test the performance of the simulation 4π r m =1
program, the results are compared with both theoretical
computations and commercial simulator outputs. It is In (4), Am and βm represent amplitude and phase of the
observed that excellent agreement is observed and the excitation in mth element. If all elements have same
simulator can be further developed to be used in both antenna pattern, F m (θ , φ ) = F (θ , φ ) , (4) can be given by
analytical and practical studies to test cases where
e − jk 0 r M
experimentation is expensive and new analytical F (θ , φ ) ∑ Am e 0 m e jβ m (5)
jk rˆ⋅d
E ( r ) = jk 0 Z 0
coding, modulation, diversity, array design techniques 4π r m =1
are to be implemented. In the first section, a brief (5) can be written as the product of the electric field of a
review of antenna array theory is presented. The single element at a selected reference point (usually the
simulation results are provided in second section. origin) and the array factor AF given by

M
AF = ∑ Ame (6)
jk 0 rˆ ⋅ d m
1. BRIEF REVIEW OF ANTENNA ARRAYS e jβ m
m =1
A generic antenna array consists of M non-uniformly
oriented elements is shown in Figure 2. The array factor is a function of the number of
elements, their geometrical arrangement, their relative
magnitudes, and their relative phases [3]. The radiated
power from an antenna system is proportional to the
square of the elecric field and hence

P (θ ,φ ) ∝ F (θ ,φ ) AF
2 2
(7)

This relation expresses the concept of pattern

multiplication which states that the radiation pattern of
an array is the product of the pattern function of the
Figure 2: A generic M-element array.
individual radiating element with the array pattern
function. In an antenna array, the directivity is mainly
The far-zone electric field radiated by a reference
contributed by the array factor.
antenna located at the origin is given by
In the developed simulation program, the user can
design any type of array from the GUI menus both at
e − jk 0 r (1)
E 0 (r ) = jk0 Z 0 F (θ , φ ) transmitting and receiving ends. Also, for convenience
4π r to the user, two special arrays, namely Broadside array
where F(θ, φ ) electric field pattern of the antenna, [1]. and Ordinary End-Fire array, can be chosen from the
For each antenna shown in Figure 2, the electric field GUI menu. Broadside array pattern has its maximum
radiation is given by radiation at an angle which is perpendicular to the
antenna platform. It is obtained when all the
e − jk0 Rm transmitting antennas are in phase. Ordinary End-Fire
E m ( r ) = jk 0 Z 0 F m (θ , φ ) (2) array has its maximum radiation in the direction af the
4π R m
antenna platform. It is obtained when there is a phase
For far zone where | r | >>dm , where m = 1,2, . . .,M, the difference depends on the wave number and the
rays from all of the antennas in the array are essentially distance from the reference antenna as β=-k0d or β=k0d.
parallel. Thus the distance from mth antenna to the far- Equation (5) is based on the assumption that the all
field point of interest can be written as Rm ≈ r in the antennas in the array have same orientation. But in this
amplitude factor. However, this approximation can not study, it is posibble that each antenna has different
be used in the exponential function and the distance Rm orientation. Hence, the far-zone electric field of the
is given by array is based on (4).
 2. THE DEVELOPED SIMULATION the open circuit voltage induced on the nth receiver
PROGRAM antenna can be given as
The detailed description of the simulation program is
Vocn = h effasn ⋅ E asn (10)
provided in [3 ]. The simulated system can input both
digital and analog sources. For analog inputs, two
examples of speech signals are readily loaded. The where heffasn is the effective kength of the nth antenna
source encoder-decoder uses Linear Predictive Coding with respect to nth antenna coordinate system, [1]. Vocn
(LPC). For the digital sources, the program enables the is a [1,L] matrix. L is the number of the samples in Vocn.
user to enter any desired bit sequence or the input can In Figure 1, the noise is added to the electric field
be chosen from the menu. Currently, there is no coding propagated in free space. In the simulation program, the
techniques for digital sources. In the channel encoder- noise is added at the input of the demodulator and it is
decoder, Hamming, Bose-Chaudhuri-Hocquenghem assumed to incorporate the channel noise, antenna
(BCH) and Reed-Solomon (RS) error correction noise, cosmic noise and the receiver noise. The open
techniques; and for the modulator/demodulator units, circuit voltage with added noise is given by
Binary Phase Shift Keying (BPSK), Coherent
Frequency Shift Keying (CFSK) and Quadrature Phase Vocng = Vocn + vn (11)
Shift Keying (QPSK) are implemented. The user can
choose from electric dipole, magnetic dipole (loop), Thus, the signal to noise ratio (SNR) is defined as the
slot, rectangular aperture, circular aperture, microstrip ratio of the signal and noise powers at the demodulator
and horn antennas in designing both transmitter and input. The power of the open circuit voltage induced on
receiver arrays. By using these antenna types, linear, the antenna is given as
planar or desired configuration of the arrays can be
constructed with the desired number of elements.
AWGN channel and Rayleigh fading channel are also S=
[V T
ocn .Vocn ] (12)
implemented. In order to facilitate the input antenna L
positions and computation of incoming electric field, The noise variance σ 2 from the user defined SNR value
various coordinate systems are defined in [3]. The total is given as
field of the array consists of M elements in Earth S (13)
σ2 =
coordinate system can be given by 10 ( SNR ) / 10

 e − jk0re M
 After obtaining the Vocng, the signals are sent to the
E e (r e , t ) =  jk 0 Z 0
4πre
∑A m F em (θ vsm , φvsm )e
j ( k0 rˆe •d evm + β m ) jw t
S l (t )e c (8) demodulator, channel decoder, and source decoder. The
 m =1 
signals can be futher be processed for a variety of
In (8), re is the magnitude of the distance vector, re , applications such as tracking, direction finding, space-
directed from the origin of the source antenna system to time coding, spatial and polarization diversity. In order
origin of the receiver antenna system with respect to to compare the simulator performance with theoretical
Earth coordinate system, r̂e is the unit vector in the computations and commercial simulator results, a bit
sequence is generated. Bit error rate (BER) is a standard
radiation direction, devm is the distance vector from the performance measure and in the following figures BER
origin of the transmitter array to mth antenna with for BPSK, QPSK and CFSK for AWGN channel and
respect to Earth coordinate system, Fem(θvsm, φvsm ) is Rayleigh fading channel are presented. The simulation
electric field pattern vector of the mth transmitter is performed with one half wave dipole in VHF band,
antenna with respect to Earth coordinate system, Am and both at transmitting and receiving ends. The distance
βm reperesent amplitude and phase excitation of the mth between the transmitter and receiver is 371 m. In case
element, Sl(t) is the baseband representation of the of AWGN channel, bit error rates for BPSK, QPSK and
modulated signal, wc is the carrier frequency. The CFSK are shown in Figure 3, Figure 4 and Figure 5,
receiver array consists of N elements. The incident respectively. The results for each modulation technique
electric field to the nth antenna (n = 1,2, . . .,N) with are compared for theoretical computations in [4] and
respect to Earth coordinate system can be given as denoted by * in figures. As shown Figure 3, Figure 4
and Figure5, the simulation produces BER results for
E asne = E e ( r e , t )e
− jk 0 d ane • rˆe (9) 100 iterations is not sufficient to match the theoretical
results. For 1000 iterations, the simulation results and
Easne is the noiseless incident electric field to the nth
theoretical computations are in accordance. As a
antenna with respect to Earth coordinate system. After
consequence, for this scenario, in order to duplicate the
the conversion of the electric field from the Earth
theoretical finding, the simulation program needs at
coordinate system to the nth antenna coordinate system,
least 1000 iterations. In Figure 6, BER for BPSK,
QPSK and CFSK modulations are provided for 1000 results for each modulation technique are compared for
iterations for AWGN channel theoretical results in [4]. For 1000 iterations the
simulated BER results agree with those of the
theoretical computations. The simulation results are also
compared with a commercial simulator in [3]. There is
very good agreement in the received power obtained
from [5] and the power obtained from the simulation
program.

Figure 3: BER results for BPSK in AWGN channel

Figure 7: BER results for Rayleigh fading channel

Figure 4: BER results for QPSK in AWGN channel 3. CONCLUSION

In this study, basic components of a digital
communication system are simulated complete with
channel and noise models. The developed simulation
program allows the user to design antenna arrays both
at the transmitting and receiving ends. The simulation
program is modular and flexible to incorporate any
future additions and updates. In order to facilitate the
input of the long list of necessary parameters for
Figure 5: BER results for CFSK in AWGN channel simulation and follow the flow of the simulation, a GUI
based menu is designed for convenience to the user.
The input parameters can both be entered from the GUI
or from previously prepared user input files.
The performance of the simulation program is tested
by obtaining the bit error rates for AWGN channel and
Rayleigh fading channels, for BPSK, QPSK and CFSK
modulation techniques. When the results are compared
with theoretical computations, a very good agreement is
observed for 1000 iterations. Also, the received power
Figure 6: BER results for AWGN channel obtained from a commercial simulator agrees very well
with the simulator results.
In case of Rayleigh fading channel, open circuit voltage
induced on the nth receiver antenna can be given as 4. REFERENCES
[1] Collin, R.E., Antennas and Radiowave Propagation,
Vocng (t ) = α n e − jβ n Vocn (t ) + v n (t ) (14) McGraw-Hill, New York, 1985.
[2]Balanis, C.A., Antenna Theory, Analysis and
Design, John Wiley & Sons, Inc., New York,1997.
where α n e − jβ is attenuation factor and phase shift for
n
[3]Güngör, A., Simulation of Digital Communication System,
nth path, vn(t) is white Gaussian noise. The amplitude α MS Thesis (In Turkish), Hacettepe University, Beytepe,
has Rayleigh distribution, βn has uniform distribution. Ankara, 2005.
From the assumption that the channel is slowly fading, [4]Haykin, S., Communications Systems, John Wiley & Sons
the phase shift can be recovered from the receiving Inc., New York.
signal. For Rayleigh fading channel, bit error rates for [5] ADS 2002 (Advanced Design System),
BPSK, QPSK and CFSK are given in Figure 7. The http://eesof.tm.agilent.com.