This action might not be possible to undo. Are you sure you want to continue?
DEPARTMENT OF TECHNOLOGY AND BUILT ENVIRONMENT
Development of a MATLAB Simulation
Environment for VehicletoVehicle and
Infrastructure Communication Based on
IEEE 802.11p
Samaneh Shooshtary
Vienna, December 2008
Master’s Thesis in Telecommunications
Examiner: Prof. Claes Beckman
Supervisors: Univ.Prof. Dipl.Ing. Dr.Ing. Christoph F. Mecklenbr¨auker
Dipl.Ing. Alexander Paier
INSTITUT FÜR
NACHRICHTENTECHNIKUND
HOCHFREQUENZTECHNIK
Abstract
This thesis describes the simulation of the proposed IEEE 802.11p Physical layer
(PHY). A MATLAB simulation is carried out in order to analyze baseband pro
cessing of the transceiver. Orthogonal Frequency Division Multiplexing (OFDM)
is applied in this project according to the IEEE 802.11p standard, which allows
transmission data rates from 3 up to 27 Mbps. Distinct modulation schemes,
Binary Phase Shift Keying (BPSK), Quadrate Phase Shift Keying (QPSK) and
Quadrature Amplitude modulation (QAM), are used according to diﬀering data
rates. These schemes are combined with time interleaving and a convolutional
error correcting code. A guard interval is inserted at the beginning of the trans
mitted symbol in order to reduce the eﬀect of Intersymbol Interference (ISI). The
Viterbi decoder is used for decoding the received signal. Simulation results illus
trate the Bit Error Rate (BER), Packet Error Rate (PER) for diﬀerent channels.
Diﬀerent channel implementations are used for the simulations. In addition a
raytracing based software tool for modelling time variant vehicular channels is
integrated into SIMULINK. BER versus Signal to Noise Ratio (SNR) statistics
are as the basic reference for the physical layer of the IEEE 802.11p standard for
all vehicular wireless network simulations.
ii
Acknowledgment
I would like to oﬀer my heartfelt thanks to Professor Christoph F. Mecklenbr¨auker
the principal supervisor providing me with support during the thesis. I also
hereby oﬀer my special gratitude to Alexander Paier, my supervisor at Technical
University of Vienna who provided me with the facilities and data and scientiﬁc
assistants. My thanks are also due to Professor Claes Beckman my examiner and
teacher. My appreciation and thanks go to Rene Wahl, from AWE Communica
tions company who always welcomed my questions with the best possible hints
and helps.
iii
Dedicated to :
My Parents; Ali and Azam
And to my husband Mahdi
iv
Abbreviation
AWGN Additive White Gaussian Noise
ACK Acknowledgment
BER Bit Error Rate
BPSK Binary Phase Shift Keying
CTS Clear to Send
CSMA Carrier Sense Multiple Access
CDF Cumulative Distribution Function
FFT Inverse Fast Fourier Transformation
GI Guard Interval
IFFT Fast Fourier Transformation
ITS Intelligent Transport Systems
IFS Inter Frame Space
ISI Intersymbol Interference
LOS Line of Sight
MAC Medium Access Control
NAV Network Allocation Vector
OFDM Orthogonal Frequency Division Multiplexing
OSI Open System Interconnection
PER Packet Error Rate
PHY Physical layer
PDF Probability Density Functions
PN Pseudorandom Noise
QPSK Quadrate Phase Shift Keying
QAM Quadrature Amplitude modulation
RMS Root Mean square
RTS Request to send
SNR Signal to Noise Ratio
V2V VehicletoVehicle
V2I VehicletoInfrastructure
v
Contents
Abstract ii
Acknowledgment iii
Abbreviation v
1 Introduction 1
1.1 Motivation for Vehicular WLAN Connections . . . . . . . . . . . 1
1.2 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Fading Statistics in Vehicular Mobile Channel . . . . . . . . . . . 2
1.3.1 Rayleigh Fading Distribution . . . . . . . . . . . . . . . . 3
1.3.2 Rician Fading Distribution . . . . . . . . . . . . . . . . . 4
1.3.3 LogNormal Distribution . . . . . . . . . . . . . . . . . . 4
1.4 Principle of OFDM Transmission . . . . . . . . . . . . . . . . . . 5
1.4.1 Using Inverse FFT to Create the OFDM Symbol . . . . . 5
1.4.2 Cyclic Preﬁx Insertion . . . . . . . . . . . . . . . . . . . . 5
2 Wireless LAN According to IEEE 802.11p 7
2.1 Deﬁnition of Key OFDM Parameters . . . . . . . . . . . . . . . . 7
2.2 Deﬁnition of Physical Layer Coding . . . . . . . . . . . . . . . . 8
2.3 Medium Access Control . . . . . . . . . . . . . . . . . . . . . . . 9
2.3.1 The Basic Access Method . . . . . . . . . . . . . . . . . . 10
2.3.2 Frame types . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.3 Most Common Frame Format . . . . . . . . . . . . . . . . 11
3 MATLAB/SIMULINK IEEE 802.11p Physical Layer Model 13
3.1 Transmitter Side . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.1.1 VariableRate Data Source . . . . . . . . . . . . . . . . . 14
3.1.2 Modulator . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.1.3 OFDM symbols . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.4 Pilot insertion . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.5 Preamble . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1.6 Assemble OFDM Frames . . . . . . . . . . . . . . . . . . 18
3.1.7 Padding . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.8 IFFT and FFT . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.9 Cyclic Preﬁx . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1.10 Multiplex OFDM Frame . . . . . . . . . . . . . . . . . . . 19
vi
Contents
3.2 Radio Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.1 AWGN Channel . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.2 Rayleigh Fading Channel with AWGN . . . . . . . . . . . 20
3.2.3 Advanced Model Using WinProp Simulator . . . . . . . . 21
3.3 Receiver Side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3.1 Demultiplex OFDM Frame . . . . . . . . . . . . . . . . . . 23
3.3.2 Remove Cyclic Preﬁx . . . . . . . . . . . . . . . . . . . . . 23
3.3.3 FFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.3.4 Frequency Domain Equalizer . . . . . . . . . . . . . . . . . 24
3.3.5 Disassemble OFDM Frame . . . . . . . . . . . . . . . . . . 25
3.3.6 Demodulator Bank . . . . . . . . . . . . . . . . . . . . . . 25
3.3.7 Adaptive Modulation Control . . . . . . . . . . . . . . . . 26
3.3.8 Evaluation of Reliability Modules . . . . . . . . . . . . . . 27
4 Deﬁnition of Vehicular Scenarios 28
4.1 Scenario I: Cars Are Going in Opposite Directions . . . . . . . . 28
4.2 Scenario II: Cars Are Going in the Same Direction . . . . . . . . 29
5 Discussion of Simulation Results 32
5.1 Simulation Result of AWGN Channel . . . . . . . . . . . . . . . 32
5.2 Simulation Result of a Multipath Rayleigh Fading and AWGN
Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.3 Comparison between Coded and Uncoded AWGN Channel . . . . 35
5.4 Packet Error Rate Calculation . . . . . . . . . . . . . . . . . . . 35
5.5 Simulation Result of Multipath Rayleigh Fading and AWGN Chan
nel with Diﬀerent Tap Delay . . . . . . . . . . . . . . . . . . . . 37
5.6 Advanced Channel Model using WinProp . . . . . . . . . . . . . . 38
6 Conclusions 44
vii
Chapter 1
Introduction
1.1 Motivation for Vehicular WLAN
Connections
Vehicular communication system is a communication network which vehicles and
roadside units are communicating with each other. The transferred information
in this type of communication is warning messages and traﬃc information. Vehic
ular communication systems are eﬀective in decreasing the accidents and traﬃc
congestions. Due to the importance of road safety in recent years, the research
on VehicletoVehicle (V2V) and VehicletoInfrastructure (V2I) communication
is increasing. IEEE 802.11p deﬁnes an international standard for wireless access
in vehicular environments. Generally wireless access in vehicular environments
contains two distinct types of networking which are V2V and V2I. Vehicular com
munications is categorized as a part of Intelligent Transport Systems (ITS). In
addition, Vehicular communication networks will oﬀer a wide range of applica
tions such as providing traﬃc management with a real time data for responding
to road congestions. In the other hand ﬁnding a better path by access to the
real time data can be as the other advantage of vehicular communication system
which cause saving the time and fuel and has large economic beneﬁts. But road
safety is the main goal of this network.
1.2 Overview
IEEE 802.11a is provided for indoor environment with high data rate commu
nication and low user mobility. IEEE 802.11p is designed for operating at high
user mobility (vehicular communication). In this thesis an existing IEEE 802.11a
PHY Simulink, [1], is updated to obtain a 802.11p PHY model. Decreasing of the
signal bandwidth from 20 MHz to 10 MHz in IEEE 802.11p makes the commu
nication more eﬃcient for high mobility vehicular channel such as reducing ISI
caused by multipath channel with using doubled guard interval. This means that
the parameters in the time domain are doubled, compare with the parameters
from IEEE 802.11a [2].
1
1.3 Fading Statistics in Vehicular Mobile Channel
For studying the Medium Access Control (MAC) and higher layers of commu
nication systems, the lower layer (PHY) and wireless channel have to be con
sidered due to their signiﬁcant eﬀect on the computational eﬃciency and the
correctness of simulation results. The draft IEEE 802.11p standard deﬁnes a
MAC and a PHY based on OFDM technique for future vehicular communication
devices. But, the study on the speciﬁcation of a IEEE 802.11p PHY working
in high mobility environments has the potential to be improved. The main goal
of this thesis is the evaluation of V2V and V2I communication PHY based on
IEEE 802.11p standard by MATLAB Simulations. In this work we refer to the
BER versus SNR statistics that can be used as the basic reference for the physical
layer of the IEEE 802.11 p standard for all vehicular wireless network simulations.
In order to get realistic simulation results, a special radio channel simulator is
used. This channel simulator is called WinProp, [3], and is based on ray trac
ing. With this simulator time variant scenarios can be modelled. A description
of the principles of the IEEE 802.11p PHY and working experience with MAT
LAB SIMULINK and WinProp software is explained in this thesis. This thesis
is organized as follows: Section 1 is motivation for vehicular WLAN connection
and principle of OFDM transmission. Section 2 provides properties of WLAN
according to IEEE 802.11p. Section 3 describes a MATLAB SIMULINK model
in software that includes transmitter, receiver and channel models. Section 4
deﬁnes the vehicular scenario in same direction and opposite direction. Section 5
is discussion of simulation results. Section 6 summarizes the thesis and presents
conclusions.
1.3 Fading Statistics in Vehicular Mobile
Channel
Fading occurs due to the multipath propagation in communications systems.
As a result signals reach the receiver from several diﬀerent paths that may
have diﬀerent lengths corresponding to diﬀerent time delays and gains. Time
delay causes additional phase shifts to the main signal component. Therefore the
signal reaching the receiver is the sum of some copies of the original signal with
diﬀerent delays and gains. With this explanation, the channel impulse response
can be modelled as described in [4] with
h
c
(t) =
K−1
k=0
α
k
δ(t −τ
k
). (1.1)
α
k
=Complex path gain
k =Number of paths
2
1.3 Fading Statistics in Vehicular Mobile Channel
τ
k
= Path delay
Two diﬀerent scales of fading have been deﬁned, large scale fading and small
scale fading. Small scale fading happens in very short time duration and is
caused by reﬂectors and scatters that change the amplitude, phase and angle
of the arriving signal. Rayleigh distribution and Rician distribution are often
used to deﬁne small scale fading. Large scale fading is due to shadowing and
the mobile station should move over a large distance to overcome the eﬀects of
shadowing. To deﬁne large scale fading, lognormal distribution is often used.
Figure 1.1 shows a scenario with multipath fading.
Figure 1.1: Multipath propagation.
1.3.1 Rayleigh Fading Distribution
Rayleigh distributions are deﬁned for fading of a channel when all the received
signals are reﬂected signals and there is no dominant component. The Rayleigh
distributions has a Probability Density Functions (PDF) given by, [5],
p(r) =
_
r
σ
2
exp
_
−
r
2
2σ
2
_
(0 ≤ r)
0 (r < 0)
, (1.2)
3
1.3 Fading Statistics in Vehicular Mobile Channel
where σ is the Root Mean Square (RMS) value of voltage in a received signal
, and σ
2
is the timeaverage power of the received signal. The Cumulative Dis
tribution Function (CDF) is deﬁned to specify the probability that the received
signal does not exceed a speciﬁc threshold R. The CDF is given by, [5],
P(R) = P(r ≤ R) =
_
R
0
p(r)dr = 1 −exp
_
−
R
2
2σ
2
_
. (1.3)
1.3.2 Rician Fading Distribution
Rician fading distribution is applied in the case that a Line of Sight (LOS) com
ponent exists between the transmitter and the receiver. The Rician distribution
is given by, [5],
p(r) =
_
r
σ
2
exp
−(r
2
+A
2
)
2σ
2
I
0
_
A
r
σ
2
_
(A ≥ 0, r ≥ 0)
0 (r < 0)
, (1.4)
The A is the amplitude of the dominant component,σ is the RMS value of
voltage in a received signal,σ
2
is the timeaverage power of the received signal
and I
0
(.) is the modiﬁed Bessel function of the ﬁrst kind and zeroorder. Tke
parameter K in Rician distribution is the ratio between the power of the LOS
component and the disperse component,
K(dB) = 10log
A
2
2σ
2
dB (1.5)
Rayleigh distribution is one kind of Rician distribution for K → 0. For K >> 1
the Rician distribution can be approximated by a Gaussian distributation, [5].
1.3.3 LogNormal Distribution
Large scale fading is due to shadowing. In this case one possibility is to use a
lognormal distribution function to deﬁne large scale fading of the channel. The
lognormal distribution has a probability density function that is given by, [6],
f(r; µ, σ) =
_
exp
−(lnx−µ)
2
/2σ
2
xσ
√
2π
(r ≥ 0)
0 (r < 0)
, (1.6)
where µ is the mean deviation and σ is standard deviation of the variable’s
logarithm [6].
4
1.4 Principle of OFDM Transmission
1.4 Principle of OFDM Transmission
Orthogonal Frequency Division Multiplexing (OFDM) is a multiplexing technique
that divides a channel with a higher relative data rate into several orthogonal
subchannels with a lower data rate.
For high data rate transmissions,the symbol duration T
s
is short. Therefore ISI
due to multipath propagation distorts the received signal, if the symbol duration
T
s
is smaller as the maximum delay of the channel. To mitigate this eﬀect a
narrowband channel is needed, but for high data rates a broadband channel is
needed. To overcome this problem the total bandwidth can be split into several
parallel narrowband subcarriers. Thus a block of N serial data symbols with
duration T
s
is converted into a block of N parallel data symbols, each with
duration T = N×T
s
. The aim is that the new symbol duration of each subcarrier
is larger than the maximum delay of the channel, T > T
max
. With many low
data rate subcarriers at the same time, a higher data rate is achieved.
In order to create the OFDM symbol a serial to parallel block is used to convert
N serial data symbols into N parallel data symbols. Then each parallel data
symbol is modulated with a diﬀerent orthogonal frequency subcarriers, and added
to an OFDM symbol, [4].
1.4.1 Using Inverse FFT to Create the OFDM Symbol
All modulated subcarriers are added together to create the OFDM symbol. This
is done by an Inverse Fast Fourier Transformation (IFFT). The advantage of using
IFFT is that the system does not need N oscillators to transmit N subcarriers.
1.4.2 Cyclic Preﬁx Insertion
The cyclic preﬁx is used in OFDM signals as a guard interval and can be deﬁned
as a copy of the end symbol that is inserted at the beginning of each OFDM
symbol. Guard interval is applied to mitigate the eﬀect of ISI due to the multipath
propagation.
Figure 1.2 shows the symbol and its delay. These delay make noise and distort
the beginning of the next symbol as shown.
To overcome this problem, one possibility is to shift the second symbol furthers
away from the ﬁrst symbol. But existence of a blank space for a continuous
communication system is not desired. In order to solve this problem a copy of
the last part of the symbol is inserted at the beginning of each symbol. This
procedure is called adding a cyclic preﬁx. Figure 1.3 shows the insertion of a
cyclic preﬁx. The Cyclic preﬁx is added after the IFFT at the transmitter, and
at the receiver the cyclic preﬁx is removed in order to get the original signal. A
detailed mathematical explanation can be found in [4].
5
1.4 Principle of OFDM Transmission
Symbol 1 Symbol 2
t
Figure 1.2: Delay from front symbol.
Symbol 1 Symbol 2
t
Extention Extention
Copy this part to the front
Copy this part to the front
Figure 1.3: Cyclic preﬁx insertion.
6
Chapter 2
Wireless LAN According to
IEEE 802.11p
2.1 Deﬁnition of Key OFDM Parameters
The IEEE 802.11p PHY has similar speciﬁcations as IEEE 802.11a with some
changes. In IEEE 802.11p, a 10 MHz frequency bandwidth is used, instead of
20 MHz bandwidth in IEEE 802.11a, thus all parameters in the time domain for
IEEE 802.11p are doubled compared with the IEEE 802.11a. The doubled guard
interval reduces ISI more than the guard interval in IEEE 802.11a.
The IEEE 802.11p PHY uses 64 subcarriers OFDM that includes 48 data sub
carriers and 4 pilot subcarriers. The 4 pilot signals are used for tracing the
frequency oﬀset and phase noise, and are located on subcarrier −21, −7, 7 and
21. The short training symbols placed at the ﬁrst part of every data packet (t
1
through t
10
shown in Figure 2.1), relates to the signal detection, time synchro
nization, and coarse frequency oﬀset estimation. The long training symbols (T
1
and T
2
), which are located after the short training symbols, are used for channel
estimation. GI2 is used as guard interval for long training sequence and GI is
used as guard interval for OFDM symbols .The cyclic preﬁx is employed to reduce
the ISI.
The total training length is 16 µs. A short OFDM training symbol consists of
12 subcarriers, which are given by
S
−26,26
=
_
(13/6) {0, 0, 1 +j, 0, 0, 0, −1 − j, 0, 0, 0, 1 +j, 0, 0, 0, −1 −j, 0, 0, 0,
−1−j, 0, 0, 0, 1+j, 0, 0, 0, 0, 0, 0, 0, −1−j, 0, 0, 0, −1−j, 0, 0, 0, 1+j, 0, 0, 0, 1+j,
0, 0, 0, 1 +j, 0, 0, 0, 1 +j, 0, 0},
where the modulation is given by the element’s value. The factor
_
(13/6) is
for normalizing the average power in the OFDM symbol. To improve the channel
estimation accuracy, long OFDM training symbols are used. The long training
symbols consist of 53 subcarriers that have a zero value at DC which are given
by
L
−26,26
= {1, 1, −1, −1, 1, 1, −1, 1, −1, 1, 1, 1, 1, 1, 1, −1, −1, 1, 1, −1, 1, −1, 1, 1,
1, 1, 0, 1, −1, −1, 1, 1, −1, 1, −1, 1, −1, −1, −1, −1, −1, 1, 1, −1, −1, 1, −1, 1, −1, 1,
1, 1, 1},
7
2.2 Deﬁnition of Physical Layer Coding
GI2 T
1
T
2
GI Data1 GI
Signal
GI t
1
t
2
t
3
t
4
t
5
t
6
t
7
t
8
t
9
t
10
10*0.8=8 µs 2*0.8+2*3.2=8.0µs 0.8+3.2=4.0µs 0.8+3.2=4.0µs 0.8+3.2=4.0µs
Signal Detect
AGC,Diversity
Selection
Coarse
Frequency
Offset
Estimation,
Timing
Synchronize
Channel and Fine
Frequency offset
Estimation
RATE
LENGTH
Service+
DATA
DATA
Data 2
Figure 2.1: OFDM training structure.
where the modulation is given by the element’s value. Depending on the data
rates, diﬀerent modulation schemes and coding rates must be applied. Table 2.1
illustrates the diﬀerence between IEEE 802.11p and IEEE 802.11a standard.
Table 2.1: Comparisons view on the key parameters of IEEE 802.11p PHY and
IEEE 802.11a PHY (Source: [2])
Parameters IEEE 802.11a IEEE 802.11p
Bitrate Mb/s 6, 9, 12, 18, 24, 3, 4.5, 6, 9,
36, 48, 54 12, 18, 24, 27
Modulation Type BPSK, QPSK, BPSK, QPSK,
16 QAM, 64 QAM 16 QAM, 64 QAM
Code Rate 1/2, 1/3, 1/4 1/2, 1/3, 1/4
Number of Subcarriers 52 52
Symbol Duration 4 µs 8 µs
Guard Time 0.8 µs 1.6 µs
FFT Period 3.2 µs 6.4 µs
Preamble Duration 16 µs 32 µs
Subcarrier 0.3125 MHz 0.15625 MHz
Frequency Spacing
Error Correction Coding K = 7 (64 states) K = 7 (64 states)
Convolutional Code Convolutional Code
2.2 Deﬁnition of Physical Layer Coding
The messages are inﬂuenced by interference and in order to detect and correct the
errors in received signals, the redundancy technique is introduced. For a binary
8
2.3 Medium Access Control
block code, an encoder is used in the transmission system to prepare data for
transmission. A binary convolutional encoder is one kind of block code, which
is used in the IEEE 802.11p standard. The coding rates of R = 1/2, 2/3, or
3/4, that correspond to the desired data rate had been used in 802.11 p. The
convolutional encoder uses the generator polynomials g
0
= 133 and g
1
= 171 in
octal mode. The constraint length of the encoder is 7 with bit rate 1/2. The
bits denoted as ”A” and ”B” are output of the encoder. Puncturing is used to
create higher data rates. Puncturing is a procedure through which the number
of transmitted bits is reduced and the coding rate is increased. The puncturing
patterns are described in section 3.1.2 and in detail in [7]. Figure 2.2 illustrates
the convolutional encoder.
Output Data A
Outpot Data B
Inpot Data Tb Tb Tb Tb
Tb Tb
Figure 2.2: Convolutional encoder (k=7).
The minimum free distance of the code determines the performance of the con
volutional code. d
free
is the minimum Hamming distance between two diﬀerent
code words which is also called free distance.
2.3 Medium Access Control
The Medium Access Control (MAC) in IEEE 802.11p is the second layer of the
lowest protocol layer of the network architecture based on the Open System
Interconnection (OSI) model. Layer 2 is divided into two diﬀerent sublayers, the
logical link control (layer 2b), and the medium access control (layer 2a). The
MAC is carried out to address some wireless communication events and control
the medium access of the node, in order to reduce collisions. Because of medium
access control, several stations can use the same physical medium. The basic
access mechanism is called distributed coordination function.
9
2.3 Medium Access Control
2.3.1 The Basic Access Method
Distributed coordination function is basically a Carrier Sense Multiple Access
(CSMA) in order to avoid collisions. To deﬁne the protocol, some acronyms
must be introduced as follows:
• RTS: A station that wants to transmit a packet ﬁrstly transmits a short
control packet that is called Request to Send (RTS). This packet is ﬁlled
with information considering the source, destination, and duration of the
transmission, [8].
• CTS: If the medium is free the destination station will respond with a
control packet that is called Clear to Send (CTS), [8].
• IFS: Two Inter Frame Spaces IFSs (Distributed IFS and short IFS) are
deﬁned as time interval between frame, to present the delay between sending
and receiving RTS, CTS, DATA and ACK packets, [9].
• ACK: An Acknowledgment (ACK) packet is sent from the receiver back to
the transmitter mentioning the successful completion of the data exchange.
Figure 2.3 deﬁneds a Carrier Sense multiple access protocol.
DIFS
RTS
CTS
SIFS
DATA
ACK
SIFS
SIFS
Transmitter Receiver
Figure 2.3: A CSMA protocol.
A CSMA protocol is implemented when a station willing to transmit a data
packet. Since the medium is busy due to transmission of other stations, the
station has to transmit the data packet to the receiver with a delay. This protocol
10
2.3 Medium Access Control
can be useful if the medium is not loaded heavily. Collision occurs when two or
more stations transmit data at the same time . In order to overcome these
problems, virtual carrier sense mechanism is deﬁned as follows:
A station transmits a data packet, if the medium is free for a speciﬁed interval
time (DIFS), then station decides to transmit a small packet (RTS). The receiver
answers with a delay (SIFS time units ), and sends a response to the source
station as a small packet (CTS). The received CTS at source station shows that
the receiver is ready to receive data. The source station will wait another SIFS
unit and then sends the data packet. Once the receiver has received the data
successfully, the receiver will wait SIFS units of time, and then send a response
to source as a small packet called ACK. When the source receives ACK, it will
be sure that the data exchange was successful. If the sender does not receive
the ACK then it will continue transmitting the data until it gets a response. A
station that receives one of these frames starts a timer, the Network Allocation
Vector (NAV) which marks the medium as busy until the end of the protocol,
[8].
2.3.2 Frame types
There are three main types of frames:
1. Data frames: Used for data transmission.
2. Control Frames: To address some wireless communication phenomena
and control medium access of nodes in order to reduce collisions.
3. Management Frames: Are used to exchange management information,
but they do not belong to the upper layers.
2.3.3 Most Common Frame Format
The Figure 2.4, 2.5 and 2.6 show some of the most common frames formats.
Duration RA TA CRC
Frame
Control
MAC Header
2
6 6
4
Octes : 2
Figure 2.4: RTS frame format.
The frame control deﬁnes the type of frame (e.g. data frames, control frames
and management frames) and depending on the frame type the duration is dif
ferent. It can be the duration value that is used for NAV calculation as deﬁned
11
2.3 Medium Access Control
Duration RA CRC
Frame
Control
MAC Header
2
6 4
Octes : 2
Figure 2.5: ACK frame format.
Duration RA CRC
Frame
Control
MAC Header
2
6 4
Octes : 2
Figure 2.6: CTS frame format.
in the virtual carrier sense mechanism. RA and TA are the receiver address and
transmitter address and CRC is used for cyclic redundancy check, [8].
12
Chapter 3
MATLAB/SIMULINK IEEE
802.11p Physical Layer Model
In order to develop the PHY of IEEE 802.11p, MATLAB SIMULINK is used.
MATLAB has a large number of libraries and tool boxes, especially in the
telecommunication ﬁeld. I started from an available MATLAB/SIMULINK model,
[1], according to IEEE 802.11a to obtain IEEE 802.11p PHY. The IEEE 802.11p
model represents a baseband model for the physical layer in a Wireless Local
Area Network (WLAN). Figure 3.1 illustrates the MATLAB/SIMULINK simu
lator architecture.
Figure 3.1: MATLAB/SIMULINK simulator architecture.
13
3.1 Transmitter Side
3.1 Transmitter Side
3.1.1 VariableRate Data Source
Binary data is created according to a predeﬁned mode. This mode is created
in adaptive modulation control according to the SNR estimation at the receiver.
This mode has to be entered to the data source to create the binary data. In
the subsystem of data source a buﬀer exists whose output is according to the
maximum bits per block which are chosen in the simulation parameter list.
3.1.2 Modulator
IEEE 802.11p OFDM PHY includes diﬀerent data rates which are selected ac
cording to the output of adaptive modulation. The system uses diﬀerent modu
lation schemes due to diﬀerent data rates. Figure 3.2 illustrates the subsystem
of modulator. The modulator is subdivided as following:
• Padding
• Convolutional encoder
• Puncturing convolutional codes
• Matrix interleaver
• General block interleaver
• Rectangular QAM
Figure 3.2: Subsystem of modulator.
Padding
The Padding block changes the dimension of input matrix along its columns, rows
or both of them according to the speciﬁed values. In this system each row is equal
to one subcarrier, it means, rows are in the frequency domain and columns are in
the time domain. In IEEE 802.11p baseband model the padding is employed for
truncating the input signal along column size. The speciﬁed output dimension is
the number of bits per block that is diﬀerent according to the diﬀerent data rate
and corresponding diﬀerent code rate and modulation scheme.
14
3.1 Transmitter Side
Convolutional Encoder
A convolutional encoder is carried out for coding of the transmitted bits. Convo
lution codes have three main parameters, the number of input bits, k, number of
output bits, n, and the number of memory register, m. k.(m + 1) is introduced
as constraint length to deﬁne a convolutional encoder in Matlab Simulink. A
poly2trellis function is used to convert generator polynomials to Trellis struc
ture, see [10].
Trellis = poly2trellis(Constraint Length , Code Generator)
In this system the Trellis structure is poly2trellis (7, [171 133]).
Puncturing Convolutional Codes
Generation of diﬀerent code rates from 1/2 code rate, which is the code rate
of the convolution encoder, is implemented through puncturing. The output
elements will be according to the puncture vector. The k
th
element of the input
vector will be removed if the k
th
element of puncture vector is zero. On the other
hand if the k
th
element of puncture vector is equal to one, then the k
th
element
of the input vector is represented in the output vector. The following example
illustrates creating new code rates from 1/2 code rate:
To create a 3/4 code rate from a 1/2 code rate, one convolutional code with
one puncture vector [ 110110 ] and constraint length 7 can be used. The third
and sixth element from the input vector are removed according to the puncture
vector. Therefore the bit rate resulting from puncture vector is 3/2, ﬁnally bit
rate will be 3/4 = 3/2.1/2, [11].
Matrix Interleaver and General Block Interleaver
Interleaving can be employed in digital data transmission technologies to mitigate
the eﬀect of burst errors. When too many errors exist in one code word, due
to a burst error, the decoding of a code word can not be done correctly. To
reduce the eﬀect of burst error, the bits in one code word are interleaved before
being transmitted. When interleaving occurs the place of bits will change, which
means that a burst error can not disturb a huge part of one code word. Figure
3.3 illustrates the eﬀect of interleaving at the transmitter.
Raw data stream
Interleaved data stream
Data Block
Figure 3.3: Transmission with interleaving.
This example explains that only a small part of each code word is distorted with
interleaving, so the decoding of code word can be done correctly. The interleaving
15
3.1 Transmitter Side
Sudden burst of noise causing errors
Interleaved data stream
Reassembled stream
Figure 3.4: Transmission with a burst error and interleaving.
in this SIMULINK model is deﬁned by two steps. The ﬁrst step is mapping of
the adjacent coded bits into the nonadjacent subcarriers that is implemented
with the matrix interleaver. The second step is mapping the adjacent coded bits
alternately onto signiﬁcant bits of the constellation that is implemented with the
general block interleaver.
Matrix interleaver interleaves the input vector according to the speciﬁed row
and column. In this system the number of rows and columns is given by:
Interleaver Rows = 16
Interleaver Columns = Number of transmitted bits per block / interleaver Rows
Figure 3.5 illustrates an example of the current process of the matrix interleaver
block. The input vector is a column vector. The dimensions of input vector in
matrix interleaver convert to a 2 by 3 matrix. The ﬁrst three elements of input
vector are the ﬁrst row of matrix and the second three elements of input vector
are the second row of matrix. Finally the matrix interleaver block rewrites the
elements column by column. Therefore the ﬁrst two elements of output are the
ﬁrst column of matrix, the second two elements of output are the second column
and the third two elements of output are the third column.
Matrix
interleaver
1
2
3
4
5
6
2by3
1 2 3
4 5 6
1
4
2
5
3
6
Figure 3.5: Matrix interleaver.
The second step of interleaving is implemented with the general block inter
leaver that changes the place of input elements according to the elements vector.
According to the elements vector, the ﬁrst element of output is the forth element
of input, the second element of output is the ﬁrst element of input, the third
element of output is the third element of input and the fourth element of output
is the second element of input. Figure 3.6 explains this process.
The parameters of the matrix and general block interleaver of the IEEE 802.11p
baseband model are following the standard, deﬁned in [7].
16
3.1 Transmitter Side
General
Block
Interleaver
[4 1 3 2]
[40 32 59 1]
[1 40 59 32]
Figure 3.6: General Block interleaver.
Rectangular QAM
The rectangular QAM block is applied to indicate how the binary words are
assigned to points of the signal constellation. In the IEEE 802.11p baseband
model a Graycode is used. Four diﬀerent modulation types are implemented:
• BPSK
• QPSK
• 16 QAM
• 64 QAM
3.1.3 OFDM symbols
To convert a block of N serial data symbols (each has a duration of T
s
) into a
block of N parallel data symbols (each has a duration of T = NT
s
), the modulator
is using a reshape block. The output vector is a number of data subcarriers by
OFDM symbol per frame.
3.1.4 Pilot insertion
Each OFDM symbol in IEEE 802.11p has four pilot subcarriers. The pilot sig
nals are used for tracing frequency oﬀset and phase noise. The location of pilot
subcarriers is −21,−7, 7 and 21.
The Pseudorandom Noise (PN) sequence generator block is carried out creating
the pilot subcarriers. The sample time and the number of samples per frame for
PN sequence Generator is deﬁned as follow:
Sample time = the period of the Block/OFDM symbol per frame
Samples per frame = OFDM symbol per frame
3.1.5 Preamble
Preamble insertion is used for channel estimation in our model in order to improve
the channel estimation accuracy. Four long OFDM training symbols are used
instead of two long training symbol in this system. The long training symbols
consist of 53 subcarriers that have a zero value at DC subcarrier. The long
training symbol is deﬁned in 2.1.
17
3.1 Transmitter Side
3.1.6 Assemble OFDM Frames
The assemble OFDM frames is applied in order to insert the pilot and training
symbols into the OFDM symbols. Figure 3.7 shows the subsystem of this block.
Four pilots are inserted between the subcarriers and then training sequence is
added to the subcarriers.
Figure 3.7: Assemble OFDM frame subsystem.
3.1.7 Padding
The Pad block extends the input vector along its columns. The padding values are
equal to zero, inserted at the end of the columns, where the speciﬁed dimension
of the output is the number of points of the IFFT block.
3.1.8 IFFT and FFT
An inverse Fourier transform converts the frequency domain data stream into
the corresponding time domain. Then a parallel to serial convertor is used to
transmit time domain samples of one symbol. The Fast Fourier Transformation
(FFT) is used to convert data in time domain to the frequency domain at the
receiver. The serial to parallel block convertor is placed to convert this parallel
data into a serial stream to obtain the original input data. Figure 3.8 illustration
the process.
IFFT block allocates the diﬀerent orthogonal subcarrier for transmitted bits
and thus no interference exists between subcarriers. In this situation subcarriers
can be closer together, which means that bandwidth can be saved signiﬁcantly,
[11].
18
3.2 Radio Channel
b0
b1
b2
.
.
.
.
.
.
.
.
.
.
.
bN1
d0
d1
d2
.
.
.
.
.
.
.
.
.
dN1
d0,d1,d2,...,dN1
d0’,d1',....,dN1'
d0'
d1'
d2'
.
.
.
.
.
.
.
.
.
.
.
.

dN1
b0
b1
b2
.
.
.
.
.
.
.
.
.
.
.
bN1
Data coded in frequency
domain:One symbol at a time
IFFT
Inverse fast
Fourier transform
P/S
Parallel to serial
converter
S/P
Serial to parallel
converter
FFT
Fast Fourier
transform
Data in time domain:
One symbol at a time
Transmit time
domain
samples of
one symbol
Receive time
domain
sample of one
symbol
Decode each
frequency bit
independently
Figure 3.8: IFFT/FFT description.
3.1.9 Cyclic Preﬁx
Cyclic preﬁx is used as a guard interval to mitigate the eﬀect of ISI due to the
multipath propagation. A selector block is applied as a cyclic preﬁx inserter to
insert the last 16 subcarriers into the beginning of the OFDM symbols.
3.1.10 Multiplex OFDM Frame
The multiplex block is the last block in the transmitter part to convert the signal
from parallel to serial and to transmit timedomain samples of one symbol.
3.2 Radio Channel
For the ﬁrst simulation a simple Additive White Gaussian Noise (AWGN) channel
model is used, following by simulations with multipath Rayleigh fading together
with AWGN model. In order to get a more realistic channel model for vehicular
scenarios, we use a speciﬁc channel simulator called WinProp.
3.2.1 AWGN Channel
To implement the eﬀect of AWGN on the input signal, an AWGN channel is
added to the input signals. This block produces a complex output signal when
the input signal is complex. In this model the variance is speciﬁed from the port
19
3.2 Radio Channel
that inserts SNR, in order to calculate the variance of the noise, as shown in
Figure 3.9, [12].
Figure 3.9: AWGN channel.
3.2.2 Rayleigh Fading Channel with AWGN
The multipath Rayleigh fading is added to an AWGN channel. Since a transmit
ted signal propagates along several paths in multipath channel to reach to the
receiver, it may lead to diﬀerent time delays. In the block, two parameter dialogs
are speciﬁed, the delay vector is used to specify time delay for each path and the
gain vector is used to specify the gain for each path at each delay. The number
of paths is according to the length of the delay vector and the gain vector which
must have the same length.
Figure 3.10: Multipath Rayleigh fading channel.
20
3.2 Radio Channel
3.2.3 Advanced Model Using WinProp Simulator
WinProp [3] is a Software package for the simulation of electromagnetic waves
and radio systems in static and time variant environments. The channel from the
WinProp simulations is used instead of the simpler model (Rayleigh fading and
AWGN) in order to achieve more realistic results. Low cost and fast determi
nation of a channel impulse response, and creating the scenarios easily are some
advantages of WinProp simulator.
WinPropTime Variant scenarios
In this section a short introduction to the time variant usage of WinProp is given.
The introduction explains that the software package can be used for predictions
in time variant scenarios, [13].
To create a realistic channel for V2V communication with WinProp , the fol
lowing steps can be passed:
1. StreetMan is used to create road courses.
2. WallMan is used to create additional objects (buildings, cars,...) and setting
the time variant properties of the objects.
3. ProjectMan is used to create a project with all settings related to the com
putation (transmitter, receiver, result,...).
4. SiMan is used for the computation in time variant scenarios.
5. ProMan is used for visualization of prediction results.
Figure 3.11 illustrates the process for the creation and simulation of a new
scenario.
Addition of static
elements,e.g.
Building
Addition of time variant objects
and their motion data
Database with
street and curves
Definition of he
project settings
Simulation of the
scenarion
Visualization of the
results
StreetMan
WallMan
ProjectMan SiMan ProMan
Figure 3.11: Process for the creation of a time variant scenario.
ProjectMan is applied to create a project, where some properties like transmit
power, center frequency, antenna pattern and antenna placement of transmitter
21
3.2 Radio Channel
are deﬁned. In addition the properties of receiver like antenna pattern and an
tenna placement are deﬁned in ProjectMan. The number of snapshots, type of
polarization, and type of output ﬁles after simulation are other properties that are
deﬁned in ProjectMan. SiMan is used for the computation time variant scenarios.
After ﬁnishing the simulations, the results appear as diﬀerent ﬁle formats:
• cir ﬁle: Channel impulse response
• fpf ﬁle: Field strength result
• fpp ﬁle: Power result ﬁle
• fpl ﬁle: Path loss result ﬁle
• str ﬁle: Propagation paths for visualization in ProMan
• mat ﬁle: A speciﬁc MATLAB output
The mat ﬁle contains the impulse response matrix of the channel and it contains
following variable:
• Matrix
• Date
• Delta delay
• Frequency
• Maximum delay
• Minimum delay
• Power
• Snapshots
• Time
The elements of the matrix are ﬁeld strength in µV/m. Minimum delay, maxi
mum delay and delta delay (resolution) are in ns, center frequency is in MHz and
transmit power is in dBm. Also mat ﬁle contains the date and time of creation
of mat ﬁle.
The following steps describes the implementation of the channel from the mat
ﬁle into the SIMULINK PHY model.
1. Using the digital ﬁlter to convolve the transmit signal with the channel
matrix to obtain the output signal. The convolution for continuous time
can be expressed by, [4],
y(t) =
_
∞
−∞
x(t −τ)h(t, τ)dτ. (3.1)
x(t)= Input signal
y(t)= Output signal
h(t)= Channel impulse response
2. Setting the digital ﬁlter to the FIR ﬁlter
22
3.3 Receiver Side
Finite Impulse Response (FIR) digital ﬁlter operates by convolving the
input signal x(n) with the ﬁlter’s impulse response h(n) to ﬁnd the output
signal as described in Figure 3.12, [14].
b3
...............................
b0
+ + + ....................
b1 b2
bM
+
Z
1
Z
1
Z
1
Z
1
X(n)
y(n)
....................................
Figure 3.12: FIR ﬁlter.
x(n −k) is the input with a delay equal to k, where k = 0, 1, 2, · · · , M and
y(n) = b
0
x(n) +b
1
x(n −1) +b
2
x(n −2) +b
3
x(n −3) +... +b
M
x(n −M)
=
M
m=0
b
m
x(n −m) =
∞
m=−∞
h(m)x(n −m) = (h ∗ x)(n) (3.2)
3. Specifying FIR ﬁlter as time varying ﬁlter
The coeﬃcients of time variant ﬁlter change with time. In this model a time
varying ﬁlter is used that the coeﬃcients of ﬁlter change once per input
frame. The second input port of the FIR ﬁlter is used for the insertion of
the channel impulse responses as coeﬃcients of the ﬁlter. The input of this
port can not be a matrix, therefore the channel matrix is converted with a
buﬀer block to a column vector, to obtain a frame based input for the FIR
ﬁlter.
4. Adding AWGN to the output of the time variant ﬁlter.
.
3.3 Receiver Side
3.3.1 Demultiplex OFDM Frame
To convert a signal from serial to parallel, a demultiplex block is used. A reshap
ing block is a subsystem of this block and is employed to produce a matrix out
of the input vector.
3.3.2 Remove Cyclic Preﬁx
In the receiver the inserted cyclic preﬁx must be removed, to obtain the original
input data. A selector block is used to remove the 16 subcarriers that are inserted
into the beginning of the OFDM symbols.
23
3.3 Receiver Side
3.3.3 FFT
A FFT block computes the fast Fourier transformation (FFT) along each column
for all input matrices to convert a time domain signal to frequency domain.
3.3.4 Frequency Domain Equalizer
To restore the transmitted signal, a zeroforcing equalizer is used that applies
the inverse of the channel frequency response. The combination of channel and
zero forcing equalizer output gives a ﬂat frequency response with linear phase to
obtain the transmitted signal, [5]. Figure 3.13 illustrate a zeroforcing equalizer.
F(f) is channel frequency response and C(f) is inverse of the channel frequency
response.
F(f) C(f)
=
*
f f f
Figure 3.13: ZeroForcing equalizer.
The frequency domain equalizer block has two input signals, the training sym
bols, which are also used at the transmitter, and the received symbols. As de
picted in Figure 3.14, the received signal is splitted up into the training symbols
and data symbols. The channel estimation is done by dividing the received train
ing symbols through the true training symbols. This estimation is used for the
zeroforcing equalizer. The subsystem of the equalizer gaing is shown in Figure
3.15.
Figure 3.14: Subsystem of equalizer.
The estimated channel is the input of the equalizer gain block and the output
is the inverse of the channel estimation. According to the zeroforcing equalizer
24
3.3 Receiver Side
Figure 3.15: Subsystem of equalizer gains.
technique with combination of received data and zero forcing equalizer output
(inverse of channel estimation), the estimated signal before the channel can be
achieved.
3.3.5 Disassemble OFDM Frame
In this part the data subcarriers are separated from pilot subcarriers and the N
parallel data symbols are converted to the N serial data symbols, to achieve the
original signal.
Figure 3.16: Disassemble OFDM frame.
3.3.6 Demodulator Bank
The demodulator subsystem performs the inverse tasks of the modulator subsys
tem. Figure 3.17 illustrates the subsystem of demodulator bank.
Figure 3.17: Subsystem of demodulator bank.
25
3.3 Receiver Side
Zero insertion
The opposite of puncturing is zero insertion. With the puncture vector at the
transmitter, diﬀerent code rates were created and at the receiver a zero insertion
block is used to convert the code rates to the base code rate, 1/2. The following
example determines the inverse process of puncturing.
If the input vector is [ 1 3 4 5 7 9 10 11 ] and the insert zero vector parameter
be a vector like [ 1 0 1 1 1 0 ], the input vector will be divided into two groups,
each with four elements. This happens because of there are four elements one in
the insert zero vector. Based on this, the block inserts zeros after the ﬁrst and
last elements of each group of four elements. This results in a 2/3 code rate. If
the code rate of the encoder is 3/4, the code rate after the insert zero block is
1/2 = 2/3.3/4. The Figure 3.18 explains this process.
Insert Zero
[1 0 3 4 5 0 7 0 9 10 11 0]
[1 3 4 5 7 9 10 11]
Figure 3.18: Zero insertion.
Viterbi decoding
The Viterbi decoder block works according to the maximum likelihood decoding.
It means ﬁnding the most probable transmitted symbol stream from the received
code word.
The Viterbi decoder, [4], deﬁnes a metric for each path and makes a decision
based on this metric. The most common metric is the Hamming distance metric.
When two paths come together on one node, the shortest hamming distance is
kept. The number of trellis branches is deﬁned as trace back depth that is 32 in
this model.
To deﬁne a convolution decoder in MATLAB simulation a poly2trellis function
is used to covert convolution code to trellis description, [10].
trellis =poly2trellis(ConstraintLength , CodeGenerator)
In this system trellis structure is poly2trellis[ 7,[171 133] ].
3.3.7 Adaptive Modulation Control
Adaptive modulation systems improve the rate of transmission. The implemen
tation of adaptive modulation is according to the channel information that is
present at the transmitter. The method of making adaptive modulation in this
model is according to the estimated SNR, a bit rate will be speciﬁed and then
data source generates binary data according to the speciﬁed data rate in adaptive
modulation control.
26
3.3 Receiver Side
3.3.8 Evaluation of Reliability Modules
Bit Error Rate
The error rate calculation block calculates the bit error rate, by comparing the
received data with transmitted data.
It has three inputs, T
x
and R
x
port that are used to accept transmitted and
received signals and the third port is used to indicate the related frame for com
putation.
Packet Error Rate
In this block the numbers of errors in the packet will be divided by the number
of packets for the calculation of packet error rate. Packet error rate is taken over
the last 50 frames.
27
Chapter 4
Deﬁnition of Vehicular Scenarios
The radio propagation characteristics are determined by many factors, such as
operating frequency band, signal bandwidth, time variant properties of objects
and antenna characters. Therefore we deﬁne two V2V scenarios, which were
simulated with the simulation tool WinProp, in this section is detail. In both
scenarios the antennas are placed on the top of the vehicles and an isotropic
radiator is assumed for the transmit and received antennas. In scenario I the
cars are going in opposite directions and in scenario II the cars are going in the
same direction around a corner.
4.1 Scenario I: Cars Are Going in Opposite
Directions
In this scenario the transmitter and receivercar are going in opposite directions.
Further there is a third car, a truck, which is going ahead the receiver car. The
transmittercar is a orange sedan and the receivercar is a blue transporter. In
order to simplify the scenario there is no building next to the road, only a guard
rail and the street are modelled. There is one lane in each direction. Figure 4.1
depicts this scenario. The complete scenario is deﬁned with following parameters:
Transmitter:
• T
x
power: 18 dBm
• Center frequency: 5.9 GHz
• Antenna pattern: Isotropic
• Antenna placement: Roof of the car
• Polarization: Vertical
• Height of the antenna: 1.5 m
• Car type: Sedan (blue)
• Car speed: 15 m/s
Receiver:
• Antenna pattern: Isotropic
28
4.2 Scenario II: Cars Are Going in the Same Direction
• Antenna placement: Roof of the car
• Polarization: Vertical
• Height of the antenna: 1.5 m
• Car type: Transporter (orange)
• Car speed: 15 m/s
Environment and other cars:
• One lane in each direction
• Straight street
• Guard rails along the street
• Third car (truck) ahead the receivercar with speed 10 m/s
In order to deﬁne the number of snapshots in time, the following settings have to
be speciﬁed. The start time, end time and interval time (time resolution). The
interval time shall be equal to the frame length of the SIMULINK PHY model,
which is 24×8µs = 192µs. The start time is equal to 0 s and end time is equal to
5 s. Further the minimum delay, maximum delay and delay resolution has to be
set. The delay resolution is equal to the inverse of the signal bandwidth, which
is 10 MHz in IEEE 802.11p. The maximum delay is equal to the time interval.
• Minimum delay: 0 ns
• Maximum delay: 10000 ns
• Delay resolution: 100 ns
4.2 Scenario II: Cars Are Going in the Same
Direction
In this scenario the transmitter and receivercar are going in the same direction.
One further car is going into the opposite direction. There are several buildings
next to the street, see Figure 4.2. As in Scenario I there is one lane per direction
and a guard rail is along the street. The street is not straight as in Scenario
I, but there is also one curve. The complete scenario is deﬁned with following
parameters:
Transmitter:
• T
x
power: 18 dBm
• Center frequency: 5.9 GHz
• Antenna pattern: Isotropic
• Antenna placement: Roof of the car
• Polarization: Vertical
• Height of the antenna: 1.5 m
29
4.2 Scenario II: Cars Are Going in the Same Direction
Transmittercar
Receivercar
Figure 4.1: Vehicular scenario I .
• Car type: Sedan (blue)
• Car speeds: 10 m/s, 11 m/s, 12 m/s and 13 m/s are related to the distance
0 m, 25 m, 53.50 m and 65 m respectively
Receiver:
• Antenna pattern: Isotropic
• Antenna placement: Roof of the car
• Polarization: Vertical
• Height of the antenna: 1.5 m
• Car type: Transporter (orange)
• Car speeds: 14 m/s, 10 m/s, 10 m/s and 10 m/s are related to the distance
0 m, 40 m, 68.50 m and 81 m respectively
Environment and other cars:
• One lane in each direction
• Street with one curve
• Guard rails along the street
• Seven buildings next to the street
• Third car is going in opposite direction with speeds: 12 m/s, 10 m/s, 10 m/s
and 15 m/s are related to the distance 0 m, 7.20 m, 15.50 m and 49 m re
spectively
30
4.2 Scenario II: Cars Are Going in the Same Direction
The time and delay paramters are the same as in Scenario I.
• Start time: 0 s
• Stop time: 5 s
• Time interval: 192 µs
• Minimum delay: 0 ns
• Maximum delay: 10000 ns
• Delay resolution: 100 ns
Transmittercar
Receivercar
Figure 4.2: Vehicular scenario II .
31
Chapter 5
Discussion of Simulation Results
5.1 Simulation Result of AWGN Channel
In this part the comparison of simulation result with a theoretical curve of error
probability for distinct modulation schemes, BPSK and QPSK is discussed. In
physical layer investigation of the IEEE 802.11p standard, the BER versus SNR
can be the reference for the whole vehicular wireless network.
The theoretical bit error probability ,P
b
, for uncoded BPSK is, [15],
Pb = Q
_
_
2E
b
N
0
_
. (5.1)
In this formula E
b
is the bit energy, N
0
is the noise power spectral density and
Q(x) is deﬁned by
Q(x) =
1
2
erfc
_
x
√
2
_
. (5.2)
The ﬁrst step of comparison the SIMULINK results with theoretical error prob
ability is for the validation. For validation of results a correcting factor is used,
in order to reﬂect the cyclic preﬁx and pilotcarriers, which are carrying no in
formation. Because of that, a part of energy of a transmitted OFDM symbol is
lost, [15].
The correcting factor for BPSK can be expressed as 80/64 that is added to the
SNR in order to obtain E
b
/N
0
. Figure 5.1 illustrates a comparison of simulation
result and theoretical curve of error probability for BPSK modulation with an
AWGN channel. The vertical axis shows the bit error probability and horizontal
axis shows the E
b
/N
0
. In general, the higher the E
b
/N
0
, the fewer the errors
probability in the simulation results. The BER versus E
b
/N
0
for the 3 Mbps
data rate (BPSK modulation scheme with 1/2 coding rate) and theoretical curve
are close together, which validates our simulations.
Figure 5.2 shows the PER versus E
b
/N
0
plot for QPSK modulation with 1/2
coding rate. The PER for simulation results is more than 0.5 at 5 dB E
b
/N
0
, and
followed by a moderate drop to obtain a low value of PER in larger than 7.5 dB
32
5.1 Simulation Result of AWGN Channel
0 2 4 6 8 10 12
10
−6
10
−5
10
−4
10
−3
10
−2
10
−1
E
b
/N
0
(dB)
U
n
c
o
d
e
d
B
i
t
E
r
r
o
r
R
a
t
e
BPSK over AWGN simulink model
BPSK over AWGN theoretical
Figure 5.1: Error probability for BPSK modulation scheme.
E
b
/N
0
. The PER for simulation results in [2] illustrate that PER is more than
0.5 at 5 dB E
b
/N
0
and is followed by a moderate drop to obtain a low value of
PER in 10 dB E
b
/N
0
. Comparison between simulation results for PER over 2304
last frames and simulation result in [2] demonstrates that in both two cases the
values of PER is decreasing in the duration between 5 dB E
b
/N
0
until more than
7.5 dB E
b
/N
0
,which validates our simulations.
3.5 4 4.5 5 5.5 6 6.5 7 7.5
10
−4
10
−3
10
−2
10
−1
10
0
E
b
/N
0
(dB)
C
o
d
e
d
P
a
c
k
e
t
E
r
r
o
r
R
a
t
e
QPSK over AWGN with 1/2 code rate
Figure 5.2: Packet error probability for QPSK modulation scheme.
33
5.2 Simulation Result of a Multipath Rayleigh Fading and AWGN Channel
5.2 Simulation Result of a Multipath Rayleigh
Fading and AWGN Channel
In this part the comparison between simulation results and theoretical error prob
ability is shown. A Rayleigh fading channel with AWGN for uncoded transmis
sion is considered. Furthermore QPSK modulation scheme with 1/2 coding rate
is used as modulation technique.
The correcting factor in this case is the same as for BPSK , see section 5.1,
and the number of tap delay is 1 and maximum delay is 100 ns.
Figure 5.3 demonstrates the simulation result of a QPSK modulation with 1/2
coding rate over a AWGN and fading channel. The start point of simulation result
curve is BER of 0.5 at −30 dB SNR for the QPSK modulation. This is followed
by a slight drop in 0 dB SNR. BER continues to fall moderately until 30 dB SNR,
when a low values of BER is obtained. Furthermore Figure 5.3 illustrates the
simulation result of a QPSK modulation over a AWGN and fading channel is
matched with theoretical curve, which validates our simulations.
−30 −20 −10 0 10 20 30 40
10
−4
10
−3
10
−2
10
−1
10
0
SNR (dB)
U
n
c
o
d
e
d
B
i
t
E
r
r
o
r
R
a
t
e
QPSK over AWGN and fading simulink model
QPSK over AWGN and fading theoretical
Figure 5.3: Error probability for BPSK modulation scheme.
Bit error probability ,P
b
, for QPSK modulation scheme is given by equation
,[16],
Pb =
1
2
_
1 −
µ
_
2 −µ
2
_
K
k=0
_
2k
k
_
_
1−µ
2
4−2µ
2
_
k
, (5.3)
34
5.3 Comparison between Coded and Uncoded AWGN Channel
where K is the number of bit symbol and µ is
µ =
_
SNR
1 +SNR
. (5.4)
5.3 Comparison between Coded and Uncoded
AWGN Channel
Figure 5.4 shows the BER performance versus E
b
/N
0
of the coded and uncoded
transmission over an AWGN channel for BPSK modulation scheme. The start
point of BER versus E
b
/N
0
for uncoded transmission is BER of 0.024 at 1 dB
E
b
/N
0
. The BER is decreasing moderately until 5 dB E
b
/N
0
, when BER reaches
to 7 × 10
−3
. The start point of the BER versus E
b
/N
0
for coded transmission
is BER of 0.079, which is 5% less than uncoded transmission. BER for uncoded
transmission continues to fall sharply in comparison with coded transmition.
BER for uncoded transmission at 2.5 dB E
b
/N
0
is 4.5 ×10
−2
and BER for coded
transmission at 2.5 dB E
b
/N
0
is 5.5 × 10
−4
which demonstrate the provment of
BER over coded transmission. Figure 5.4 describes that the simulation results
with coding have lower error compared with the simulation results without cod
ing, because of error detection and correction with convolutional encoder.
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
10
−4
10
−3
10
−2
10
−1
E
b
/N
0
(dB)
B
i
t
E
r
r
o
r
R
a
t
e
BPSK over AWGN uncoded
BPSK over AWGN coded
Figure 5.4: Bit error probability for BPSK modulation scheme.
5.4 Packet Error Rate Calculation
Figure 5.5 and 5.6 show the PER performance versus SNR of the coded and
uncoded transmission over 50 last frame. The higher the SNR, the fewer the PER
35
5.4 Packet Error Rate Calculation
in the simulation results. Furthermore comparison between Figure 5.5 and 5.6
illustrates that coded transmission has lower PER compared with the simulation
results for uncoded transmission.
−2 0 2 4 6 8 10
0
20
40
60
80
100
SNR (dB)
P
E
R
(
%
)
PER for uncoded
transmission
Figure 5.5: Packet Error Rate of BPSK over AWGN channel.
−1 0 1 2 3 4
0
20
40
60
80
100
SNR (dB)
P
E
R
(
%
)
PER for coded transmission
Figure 5.6: Packet Error Rate of BPSK over AWGN channel.
36
5.5 Simulation Result of Multipath Rayleigh Fading and AWGN Channel with
Diﬀerent Tap Delay
5.5 Simulation Result of Multipath Rayleigh
Fading and AWGN Channel with Diﬀerent
Tap Delay
Since in the multipath channel signals are reﬂected at multiple places, they reach
the receiver from several diﬀerent paths that may have diﬀerent lengths and
diﬀerent corresponding time delays. To specify the channel with diﬀerent tap
delays in fading channel a tapmodel is used which is called ITU vehicular channel
A [17]. This model is not allocated to V2V communication, this is a Channel
impulse response model based on a tappeddelay. The model is speciﬁed by the
number of taps, the time delay relative to the ﬁrst tap and the average power
relative to the strongest tap. The model is set up according to the Table 5.1.
Table 5.1: ITU Vehicular Channel Model (Channel A)(Source: [17])
Tap Relative delay(ns) Average power(dB)
1 0 0.0
2 310 1.0
3 710 9.0
4 1090 10.0
5 1730 15.0
6 2510 20.0
The comparison of BER for simulation results for vehicular channel A and
QPSK over AWGN and fading SIMULINK model is illustrated in the Figure 5.7.
The correcting factor is the same as before. Simulation results in this part are
somewhat diﬀerent which is due to more than one tap delay in vehicular channel
A.
Maximum excess delay in vehicular channel is 2.510µs > 1.6µs (cyclic preﬁx)
the RMS delay spread and mean excess delay are 604.1 ns and 466.10 ns. The
RMS delay spread in this case is less than cyclic preﬁx.
Mean Excess delay and RMS delay are,[5],
Mean Excess Delay =
K
k=0
p(τ
k
).τ
k
K0
p(τ
k
)
(5.5)
RMS delay =
_
(τ
2
) − (τ)
2
(5.6)
τ
2
=
K
k=0
p(τ
k
)τ
2
K
K
k=0
p(τ
k
)
(5.7)
where K is number of paths.
37
5.6 Advanced Channel Model using WinProp
−30 −20 −10 0 10 20 30 40
10
−3
10
−2
10
−1
10
0
SNR (dB)
U
n
c
o
d
e
d
B
i
t
E
r
r
o
r
R
a
t
e
QPSK over AWGN and fading
simulink model
QPSK over vehicular A model
Figure 5.7: Error probability for QPSK modulation scheme.
5.6 Advanced Channel Model using WinProp
The channel resulting from WinProp is used in the SIMULINK model instead of
the simpler model,mention in the last sections, to achieve more realistic results.
Distinct vehicular scenarios described in section 4 are used in this part.
The resolution of the delay is deﬁned according to the Tx sample rate in Matlab
simulation that is 100 ns. In this part the normalization of impulse response
matrix is taken into account to obtain more realistic result. Equation 5.8 deﬁne
the relation between transmit signal and received signal,
Y = (H ∗ X) +N, (5.8)
equation 5.9 is used to deﬁne a certain value of SNR. To simplify obtaining
the SNR, the values of mean power (E {H(x)
2
}) and Average transmit power
(E {x
2
}) is set to one,
SNR =
E {H(x)
2
}
E {N
2
}
.E
_
x
2
_
, (5.9)
Therfore
SNR =
1
E {N
2
}
. (5.10)
Equation 5.11 and Equation 5.12 are the deﬁnition of mean power,[4],
38
5.6 Advanced Channel Model using WinProp
Continuous:
MeanPower = lim
T→∞
1
T
_
T/2
−T/2
_
∞
0
h(t, τ)
2
dτdt, (5.11)
Discrete:
MeanPower = lim
T→∞
1
T
k
h=−k
∆t
L
l=0
h(k∆t, l∆t)
2
∆τ, (5.12)
in MATLAB the impulse response channel matrix is divided to square root of
mean power to normalize the (E {H(x)
2
}) to one.
Simulation Result of Advance Channel Model in Opposite Direction
The MATLAB output of WinProp contains the ﬁeld strength in µV/m in three
components, E(x), E(y), and E(z). Figure 5.8 illustrates the magnitude of E(z
) as a function of time. Due to the vertical polarization the magnitude of E(z)
is more than the magnitude of E(x) and E(y). Figure 5.8 illustrates that the
magnitude of E(z) is increasing until two cars reaching to each other at t = 2.9 s.
The magnitude is decreasing due to the increasing the distance between two cars
after passing from each other.
0 1 2 3 4 5
20
40
60
80
100
120
Time (s)
M
a
g
n
i
t
u
d
e
o
f
E
(
z
)
v
e
c
t
o
r
(
d
B
)
Figure 5.8: Magnitude of E(z) vector (dB).
Figure 5.9 shows the BER for vehicular scenario in opposite direction. The
comparison between Figure 5.8 and Figure 5.9 illustrates that when a fading
accrues, the magnitude of E(z) vector is decreasing and the bit error rate is
increasing. When two cars reach to each other, the highest values of magnitude
39
5.6 Advanced Channel Model using WinProp
for E(z) occurs, and therefore a decreasing of values for BER are obtained. Then
the values of BER increase due to the increasing the distance between two cars
after passing from each other. Figure 5.10 shows the diﬀerent values of data rate
that generated according the adaptive modulation.
0 1 2 3 4 5
0
0.2
0.4
0.6
0.8
1
1.2
1.4
x 10
−3
Time (s)
B
E
R
Figure 5.9: BER of vehicular scenario in opposite direction.
0 1 2 3 4 5
3
3.5
4
4.5
Time (s)
D
a
t
a
R
a
t
e
Figure 5.10: Data rate of vehicular scenario in opposite direction .
Simulation Result of Advance Channel Model in Same Direction
Figure 5.11 illustrates the magnitude of E(z) as a function of time. Due to
the vertical polarization the magnitude of E(z) is more than the magnitude of
40
5.6 Advanced Channel Model using WinProp
E(x) and E(y). The magnitude of E(z) is changing between 95 dB to 105 dB in
the same direction of transmitter and receiver scenario. The curve of E(z) in
the same direction scenario is varying less than variation in opposite direction
scenario, because of the higher variation of the Doppler shift in the opposite
direction scenario. The magnitude of E(z) decreases at t = 3.7 s, when the third
car is passing from the transmitter and receiver in the opposite direction. Figure
5.12 shows the BER for vehicular scenario in the same direction. The comparison
between Figure 5.11 and Figure 5.12 illustrates that when the third car is passing
from the transmitter and receiver, the lowest values of magnitude for E(z) occurs,
and therefore an increasing of values for BER are obtained. Figure 5.13 shows the
diﬀerent values of data rate that generated according to the adaptive modulation.
0 1 2 3 4 5
80
90
100
110
120
Time (s)
M
a
g
n
i
t
u
d
e
o
f
E
(
z
)
v
e
c
t
o
r
(
d
B
)
Figure 5.11: Magnitude of E(z) vector (dB) .
Analyzing of received power for Advance Channel Model
Figures 5.14 and 5.15 illustrate the received power and pathloss related to the
transmitter that is on the roof of the orange car was showed in ﬁgure 4.1. Received
power is decreasing by distance and will not be understandable in blue area due
to sensitivity level of ITS transceivers in practice.
The Path loss is increased by distance. In the near area to the transmitter, we
can see that pathloss is less than 60 dB and it is increasing to more than 100 dB
in higher distances. This phenomenon is illustrated in Figure 5.15.
41
5.6 Advanced Channel Model using WinProp
1 2 3 4 5
0
2
4
6
8
10
12
x 10
−5
Time (s)
B
E
R
Figure 5.12: Bit error rate of vehicular scenario in same direction .
0 1 2 3 4 5
3
3.5
4
4.5
Time (s)
D
a
t
a
R
a
t
e
Figure 5.13: Data rate of vehicular scenario in same direction .
42
5.6 Advanced Channel Model using WinProp
Figure 5.14: Received power.
Figure 5.15: Path loss.
43
Chapter 6
Conclusions
At this work a MATLAB simulation was carried out in order to analyze base
band processing of the proposed IEEE 802.11p physical layer. Diﬀerent channel
implementations are used for the simulations. For the ﬁrst simulation a simple
AWGN channel model is used, following by simulations with multipath Rayleigh
fading together with AWGN model. In order to get a more realistic channel
model for vehicular scenarios, I use a speciﬁc channel simulator called WinProp.
The WinProp is a Software package for the simulation of electromagnetic waves
and radio systems in static and time variant environments.
I compared my simulation results with theoretical curves of error probability
to validation of my simulation result. The comparison of BER versus E
b
/N
0
for
AWGN channel and theoretical curve illustrate that simulation result over BPSK
modulation scheme with 1/2 coding rate are close together. The comparison of
simulation results over Rayleigh fading and AWGN channel with theoretical curve
for QPSK by 1/2 coding rate illustrates that the simulation results and theoret
ical curve are matched together, which validates our simulation results.
I have integrated a raytracing based software tool, WinProp, for modelling
time variant vehicular channel into SIMULINK. Normalization of impulse re
sponse matrix is taken into account to obtain more realistic results. The compar
ison between the simulation results of advanced channel model using WinProp in
same direction scenario and opposite direction scenario illustrates that the curve
of E(z) in the same direction scenario is varying less than variation in opposite
direction scenario. Furthermore the value of BER for vehicular scenario in same
direction is less than the values of BER for vehicular scenario in opposite direc
tion because of the higher variation of the Doppler shift in the opposite direction
scenario.
A description of the principles of the IEEE 802.11p PHY and working expe
rience with MATLAB SIMULINK and WinProp software was obtained through
this thesis.
44
Bibliography
[1] http://www.mathworks.com/matlabcentral/ﬁleexchange/3540, M. Clark,
“IEEE 802.11a WLAN model.”
[2] Y. Zang, L. Stibor, G. Orfanos, S. Guo, and H. Reumerman, “An error
model for intervehicle communications in highway scenarios at 5.9 GHz,”
in International Workshop on Modeling Analysis and Simulation of Wireless
and Mobile Systems, Montreal, Quebec, Canada, 2005.
[3] http://www.awe com.de/Automotive/, AWE Communications GmbH.
[4] A. F. Molisch, Wireless Communications. IEEEPress  Wiley and Sons,
2005.
[5] T. S. Rappaport, Wireless Communications: Principle and Practice. NJ:
PrenticeHall, 1996.
[6] E. Limpert, W. A. Stahel, and M. Abbt, “Lognormal distributions across
the sciences: Keys and clues,” Bio Science, vol. 51, pp. 341–352, 2001.
[7] IEEE 802.11, “Wireless LAN medium access control (MAC) and physical
layer (PHY) speciﬁcations,” June 2003.
[8] P. Brenner, “A technical tutorial on the IEEE 802.11 protocol,” Breezecom
Wireless Communications, July 1996.
[9] P. Pant and T. Castelli, “Simulation of a wireless network using the 802.11
MAC protocol,” Final Report.
[10] www.mathworks.com/access/helpdesk/help/toolbox/commblks, communi
cations blockset, Convolutional encoder.
[11] L. Z. Fuertes, “OFDM PHY layer implementation based on the 802.11a
standard and system performance analysis,” B.Sc. thesis, 2005, university of
Link¨oping, Link¨oping, Sweden.
[12] www.mathworks.com/access/helpdesk/help/toolbox/commblks, communi
cations blockset, AWGN channel.
[13] R. Wahl, “An introduction to WinProp time variant,” AWE Communica
tions GmbH.
[14] H. Monson, Statistical Digital Signal Processing and Modeling. Wiley, 1996.
[15] E. Mark, Wireless OFDM Systems: How to Make Them Work? Kluwer
Academics Publishers, 2002.
[16] J. G. Proakis, Digital Communications. McGrawHill, 1995.
[17] TUR Rec. M 1225, “Guidelines for evaluation of radio transmission tech
nologies (RTTs) for IMT2000,” 1997.
45
List of Figures
1.1 Multipath propagation . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Delay from front symbol . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Cyclic preﬁx insertion . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1 OFDM training structure . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Convolutional encoder (k=7) . . . . . . . . . . . . . . . . . . . . . 9
2.3 A CSMA protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4 RTS frame format . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5 ACK frame format . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.6 CTS frame format . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1 MATLAB/SIMULINK simulator architecture . . . . . . . . . . . 13
3.2 Subsystem of modulator . . . . . . . . . . . . . . . . . . . . . . . 14
3.3 Transmission with interleaving . . . . . . . . . . . . . . . . . . . . 15
3.4 Transmission with a burst error and interleaving . . . . . . . . . . 16
3.5 Matrix interleaver . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.6 General Block interleaver . . . . . . . . . . . . . . . . . . . . . . . 17
3.7 Assemble OFDM frame subsystem . . . . . . . . . . . . . . . . . 18
3.8 IFFT/FFT description . . . . . . . . . . . . . . . . . . . . . . . . 19
3.9 AWGN channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.10 Multipath Rayleigh fading channel . . . . . . . . . . . . . . . . . 20
3.11 Process for the creation of a time variant scenario . . . . . . . . . 21
3.12 FIR ﬁlter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.13 ZeroForcing equalizer . . . . . . . . . . . . . . . . . . . . . . . . 24
3.14 Subsystem of equalizer . . . . . . . . . . . . . . . . . . . . . . . . 24
3.15 Subsystem of equalizer gains . . . . . . . . . . . . . . . . . . . . . 25
3.16 Disassemble OFDM frame . . . . . . . . . . . . . . . . . . . . . . 25
3.17 Subsystem of demodulator bank . . . . . . . . . . . . . . . . . . . 25
3.18 Zero insertion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.1 Vehicular scenario I . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.2 Vehicular scenario II . . . . . . . . . . . . . . . . . . . . . . . . 31
5.1 Error probability for BPSK modulation scheme . . . . . . . . . . 33
5.2 Packet error probability for QPSK modulation scheme . . . . . . 33
5.3 Error probability for BPSK modulation scheme . . . . . . . . . . 34
5.4 Bit error probability for BPSK modulation scheme . . . . . . . . 35
46
List of Figures
5.5 Packet Error Rate of BPSK over AWGN channel . . . . . . . . . 36
5.6 Packet Error Rate of BPSK over AWGN channel . . . . . . . . . 36
5.7 Error probability for QPSK modulation scheme . . . . . . . . . . 38
5.8 Magnitude of E(z) vector (dB) . . . . . . . . . . . . . . . . . . . . 39
5.9 BER of vehicular scenario in opposite direction . . . . . . . . . . 40
5.10 Data rate of vehicular scenario in opposite direction . . . . . . . 40
5.11 Magnitude of E(z) vector (dB) . . . . . . . . . . . . . . . . . . . 41
5.12 Bit error rate of vehicular scenario in same direction . . . . . . . 42
5.13 Data rate of vehicular scenario in same direction . . . . . . . . . 42
5.14 Received power . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.15 Path loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
47
List of Tables
2.1 Comparisons view on the key parameters of IEEE 802.11p PHY
and IEEE 802.11a PHY (Source: [2]) . . . . . . . . . . . . . . . . 8
5.1 ITU Vehicular Channel Model (Channel A)(Source: [17]) . . . . . 37
48
Abstract
This thesis describes the simulation of the proposed IEEE 802.11p Physical layer (PHY). A MATLAB simulation is carried out in order to analyze baseband processing of the transceiver. Orthogonal Frequency Division Multiplexing (OFDM) is applied in this project according to the IEEE 802.11p standard, which allows transmission data rates from 3 up to 27 Mbps. Distinct modulation schemes, Binary Phase Shift Keying (BPSK), Quadrate Phase Shift Keying (QPSK) and Quadrature Amplitude modulation (QAM), are used according to diﬀering data rates. These schemes are combined with time interleaving and a convolutional error correcting code. A guard interval is inserted at the beginning of the transmitted symbol in order to reduce the eﬀect of Intersymbol Interference (ISI). The Viterbi decoder is used for decoding the received signal. Simulation results illustrate the Bit Error Rate (BER), Packet Error Rate (PER) for diﬀerent channels. Diﬀerent channel implementations are used for the simulations. In addition a raytracing based software tool for modelling time variant vehicular channels is integrated into SIMULINK. BER versus Signal to Noise Ratio (SNR) statistics are as the basic reference for the physical layer of the IEEE 802.11p standard for all vehicular wireless network simulations.
ii
Acknowledgment
I would like to oﬀer my heartfelt thanks to Professor Christoph F. Mecklenbr¨uker a the principal supervisor providing me with support during the thesis. I also hereby oﬀer my special gratitude to Alexander Paier, my supervisor at Technical University of Vienna who provided me with the facilities and data and scientiﬁc assistants. My thanks are also due to Professor Claes Beckman my examiner and teacher. My appreciation and thanks go to Rene Wahl, from AWE Communications company who always welcomed my questions with the best possible hints and helps.
iii
Dedicated to :
M y P arents; Ali and Azam And to my husband M ahdi
iv
Abbreviation
AWGN ACK BER BPSK CTS CSMA CDF FFT GI IFFT ITS IFS ISI LOS MAC NAV OFDM OSI PER PHY PDF PN QPSK QAM RMS RTS SNR V2V V2I Additive White Gaussian Noise Acknowledgment Bit Error Rate Binary Phase Shift Keying Clear to Send Carrier Sense Multiple Access Cumulative Distribution Function Inverse Fast Fourier Transformation Guard Interval Fast Fourier Transformation Intelligent Transport Systems Inter Frame Space Intersymbol Interference Line of Sight Medium Access Control Network Allocation Vector Orthogonal Frequency Division Multiplexing Open System Interconnection Packet Error Rate Physical layer Probability Density Functions Pseudorandom Noise Quadrate Phase Shift Keying Quadrature Amplitude modulation Root Mean square Request to send Signal to Noise Ratio VehicletoVehicle VehicletoInfrastructure
v
. . . . . . .4 Pilot insertion . .3 Medium Access Control . . . . 2. . . . . . . .1. . . . 3. . . . . . . 2. . . . . . . 3. . . . . . . . . 1.11p 2. . . . . . . . . . . . . . . .6 Assemble OFDM Frames .2 Deﬁnition of Physical Layer Coding . . . . . . . . . . . . . . . . . . . 2. .3. . . . . . . . . . . . . . . . . . . . . . . . . . 3. 1. . .10 Multiplex OFDM Frame . . .4. . Model . . . . . . .3 LogNormal Distribution . . . . . . . . . . . . . . . . . . .1. . . . . . . . .8 IFFT and FFT . . 3. . . . . . . .1 Deﬁnition of Key OFDM Parameters . . . . . 1. .1. .5 Preamble . . . . . . . . . . . . . . . . . . . . . . . . vi .3 OFDM symbols . . . . . . . . . . .2 Cyclic Preﬁx Insertion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. . . .2 Rician Fading Distribution . . . . . . 3. . . . 2 Wireless LAN According to IEEE 802.3. . . . . . . . . . 1. . . .1. . . .1. . . . . . . . . . . .1 Motivation for Vehicular WLAN Connections . . . . .1. . . . .1. . . . . . . . . . . 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. ii iii v 1 1 1 2 3 4 4 5 5 5 7 7 8 9 10 11 11 13 14 14 14 17 17 17 18 18 18 19 19 Physical Layer . . . . . . . . . . .1 Transmitter Side . . . . . . . . . 1. . . . .2 Frame types . . . . 3. . . . . . . . . . . .3. . . . . . . . . . . 2. .11p 3. . . . .2 Modulator . . .1 Rayleigh Fading Distribution . .9 Cyclic Preﬁx . .1 Using Inverse FFT to Create the OFDM Symbol 1. .3. . . . . . . . .Contents Abstract Acknowledgment Abbreviation 1 Introduction 1. . . . . . . . . . . . .2 Overview . . 3. . . . 3. . . . . . . . . . . . . .1. . . .1. . . . . . . . . .3 Most Common Frame Format .1. . . . . 3. . . . . . . . . 1. . . . . . . . . . . . . .3 Fading Statistics in Vehicular Mobile Channel . . . . . . . . . .4 Principle of OFDM Transmission . . . . 3 MATLAB/SIMULINK IEEE 802. . . . .3. .7 Padding . . . . . .1 VariableRate Data Source . . . . . . .1 The Basic Access Method . . . . . . .3. .4. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . .4 Frequency Domain Equalizer . . . . . . . . . . . . . . .1 Demultiplex OFDM Frame . . . . . . . . . . . . .6 Demodulator Bank . . . . . . . . . . . .5 Disassemble OFDM Frame . . . . . 19 19 20 21 23 23 23 24 24 25 25 26 27 28 28 29 32 32 34 35 35 37 38 44 3. 5. . . . . .2 Rayleigh Fading Channel with AWGN . . . . . . . . . . . . . . . . . . .2 Remove Cyclic Preﬁx . . . . . . . . . . . . . . . . . .2. . . . 3.3 4 Deﬁnition of Vehicular Scenarios 4. . . . . . 3. . . . . . . . . . . . .3. . . . . . . . . . . . . . . 3. . . . . . . . . . . . . .7 Adaptive Modulation Control . . . . . . .1 Scenario I: Cars Are Going in Opposite Directions 4. . . .2 Scenario II: Cars Are Going in the Same Direction . . . . . . .3. . . . 3. . . . . . . .1 Simulation Result of AWGN Channel . . . . . . . . . .3. . . . . . . . . . 5 Discussion of Simulation Results 5.3. . . . . . .4 Packet Error Rate Calculation . . . . . . . . . . . . .2 Radio Channel . . . . . . . . . . . . . . . . . . . . . . 3.3. . 3. . . 5.2 Simulation Result of a Multipath Rayleigh Fading and AWGN Channel . .2.3. . . 3. . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . . .3. . . .Contents 3. . 3. . . . . . . 3. . . . . . . . 3. . . . . . . . 5. . . . . . . . .3 Advanced Model Using WinProp Simulator Receiver Side . . . . . . . . 3. . . .1 AWGN Channel . .3 FFT . 5. .2. . . . . .3. . . . . . . .3 Comparison between Coded and Uncoded AWGN Channel . . . 6 Conclusions vii . . . . . . . .5 Simulation Result of Multipath Rayleigh Fading and AWGN Channel with Diﬀerent Tap Delay . . . . . . . . . .8 Evaluation of Reliability Modules . . . . . . . . . . . . . . . . . . .6 Advanced Channel Model using WinProp . . .
IEEE 802.Chapter 1 Introduction 1.2 Overview IEEE 802.11p deﬁnes an international standard for wireless access in vehicular environments. But road safety is the main goal of this network. Vehicular communication networks will oﬀer a wide range of applications such as providing traﬃc management with a real time data for responding to road congestions. Due to the importance of road safety in recent years. is updated to obtain a 802. The transferred information in this type of communication is warning messages and traﬃc information. the research on VehicletoVehicle (V2V) and VehicletoInfrastructure (V2I) communication is increasing. 1 . Vehicular communications is categorized as a part of Intelligent Transport Systems (ITS). Generally wireless access in vehicular environments contains two distinct types of networking which are V2V and V2I. compare with the parameters from IEEE 802. Decreasing of the signal bandwidth from 20 MHz to 10 MHz in IEEE 802. This means that the parameters in the time domain are doubled.11a is provided for indoor environment with high data rate communication and low user mobility.11a PHY Simulink. 1. In the other hand ﬁnding a better path by access to the real time data can be as the other advantage of vehicular communication system which cause saving the time and fuel and has large economic beneﬁts. Vehicular communication systems are eﬀective in decreasing the accidents and traﬃc congestions. In addition.11p is designed for operating at high user mobility (vehicular communication). [1].11a [2].11p makes the communication more eﬃcient for high mobility vehicular channel such as reducing ISI caused by multipath channel with using doubled guard interval. IEEE 802.11p PHY model.1 Motivation for Vehicular WLAN Connections Vehicular communication system is a communication network which vehicles and roadside units are communicating with each other. In this thesis an existing IEEE 802.
In order to get realistic simulation results. A description of the principles of the IEEE 802. This channel simulator is called WinProp. The draft IEEE 802. Section 2 provides properties of WLAN according to IEEE 802. With this simulator time variant scenarios can be modelled. In this work we refer to the BER versus SNR statistics that can be used as the basic reference for the physical layer of the IEEE 802. (1.3 Fading Statistics in Vehicular Mobile Channel For studying the Medium Access Control (MAC) and higher layers of communication systems.1) αk =Complex path gain k =Number of paths 2 . Section 6 summarizes the thesis and presents conclusions. Section 3 describes a MATLAB SIMULINK model in software that includes transmitter. receiver and channel models. The main goal of this thesis is the evaluation of V2V and V2I communication PHY based on IEEE 802.11p PHY and working experience with MATLAB SIMULINK and WinProp software is explained in this thesis. and is based on ray tracing. the study on the speciﬁcation of a IEEE 802. the channel impulse response can be modelled as described in [4] with K−1 hc (t) = k=0 αk δ(t − τk ). But. Time delay causes additional phase shifts to the main signal component. a special radio channel simulator is used.11p PHY working in high mobility environments has the potential to be improved. Therefore the signal reaching the receiver is the sum of some copies of the original signal with diﬀerent delays and gains. Section 4 deﬁnes the vehicular scenario in same direction and opposite direction.11 p standard for all vehicular wireless network simulations. This thesis is organized as follows: Section 1 is motivation for vehicular WLAN connection and principle of OFDM transmission.11p standard by MATLAB Simulations. As a result signals reach the receiver from several diﬀerent paths that may have diﬀerent lengths corresponding to diﬀerent time delays and gains. With this explanation. the lower layer (PHY) and wireless channel have to be considered due to their signiﬁcant eﬀect on the computational eﬃciency and the correctness of simulation results.11p. 1. [3]. Section 5 is discussion of simulation results.11p standard deﬁnes a MAC and a PHY based on OFDM technique for future vehicular communication devices.1.3 Fading Statistics in Vehicular Mobile Channel Fading occurs due to the multipath propagation in communications systems.
1. Rayleigh distribution and Rician distribution are often used to deﬁne small scale fading. lognormal distribution is often used.1 Rayleigh Fading Distribution Rayleigh distributions are deﬁned for fading of a channel when all the received signals are reﬂected signals and there is no dominant component.1: Multipath propagation. The Rayleigh distributions has a Probability Density Functions (PDF) given by. Large scale fading is due to shadowing and the mobile station should move over a large distance to overcome the eﬀects of shadowing.2) 3 . Figure 1. Small scale fading happens in very short time duration and is caused by reﬂectors and scatters that change the amplitude. 1. [5].1 shows a scenario with multipath fading. To deﬁne large scale fading. Figure 1.3. p(r) = r exp σ2 r − 2σ2 2 0 (0 ≤ r) (r < 0) . (1.3 Fading Statistics in Vehicular Mobile Channel τk = Path delay Two diﬀerent scales of fading have been deﬁned. large scale fading and small scale fading. phase and angle of the arriving signal.
4 . [5]. The lognormal distribution has a probability density function that is given by. σ) = exp−(lnx−µ) √ xσ 2π 2 /2σ 2 0 (r ≥ 0) (r < 0) . µ. (1.6) where µ is the mean deviation and σ is standard deviation of the variable’s logarithm [6]. K(dB) = 10log 1.3. R P (R) = P (r ≤ R) = 0 p(r)dr = 1 − exp − R2 2σ 2 .3 LogNormal Distribution Large scale fading is due to shadowing. (1.3) 1. The Cumulative Distribution Function (CDF) is deﬁned to specify the probability that the received signal does not exceed a speciﬁc threshold R.) is the modiﬁed Bessel function of the ﬁrst kind and zeroorder.σ is the RMS value of voltage in a received signal. [5].3. and σ 2 is the timeaverage power of the received signal. For K >> 1 the Rician distribution can be approximated by a Gaussian distributation.3 Fading Statistics in Vehicular Mobile Channel where σ is the Root Mean Square (RMS) value of voltage in a received signal . r ≥ 0) (r < 0) .1. Tke parameter K in Rician distribution is the ratio between the power of the LOS component and the disperse component.σ 2 is the timeaverage power of the received signal and I0 (. [5]. The CDF is given by. A2 dB (1.2 Rician Fading Distribution Rician fading distribution is applied in the case that a Line of Sight (LOS) component exists between the transmitter and the receiver. In this case one possibility is to use a lognormal distribution function to deﬁne large scale fading of the channel. p(r) = r exp σ2 −(r 2 +A2 ) 2σ 2 I0 Ar σ2 0 (A ≥ 0.4) The A is the amplitude of the dominant component.5) 2σ 2 Rayleigh distribution is one kind of Rician distribution for K → 0. The Rician distribution is given by. (1. f (r. [6].
a higher data rate is achieved. Then each parallel data symbol is modulated with a diﬀerent orthogonal frequency subcarriers. A detailed mathematical explanation can be found in [4]. if the symbol duration Ts is smaller as the maximum delay of the channel.4 Principle of OFDM Transmission Orthogonal Frequency Division Multiplexing (OFDM) is a multiplexing technique that divides a channel with a higher relative data rate into several orthogonal subchannels with a lower data rate. To overcome this problem. Thus a block of N serial data symbols with duration Ts is converted into a block of N parallel data symbols. T > Tmax . but for high data rates a broadband channel is needed. each with duration T = N ×Ts .2 Cyclic Preﬁx Insertion The cyclic preﬁx is used in OFDM signals as a guard interval and can be deﬁned as a copy of the end symbol that is inserted at the beginning of each OFDM symbol. To mitigate this eﬀect a narrowband channel is needed. This procedure is called adding a cyclic preﬁx.2 shows the symbol and its delay.4 Principle of OFDM Transmission 1. and at the receiver the cyclic preﬁx is removed in order to get the original signal.4.1 Using Inverse FFT to Create the OFDM Symbol All modulated subcarriers are added together to create the OFDM symbol. The aim is that the new symbol duration of each subcarrier is larger than the maximum delay of the channel. 5 . In order to create the OFDM symbol a serial to parallel block is used to convert N serial data symbols into N parallel data symbols. Therefore ISI due to multipath propagation distorts the received signal.3 shows the insertion of a cyclic preﬁx. 1. [4]. Guard interval is applied to mitigate the eﬀect of ISI due to the multipath propagation. The Cyclic preﬁx is added after the IFFT at the transmitter. 1. In order to solve this problem a copy of the last part of the symbol is inserted at the beginning of each symbol. With many low data rate subcarriers at the same time. To overcome this problem the total bandwidth can be split into several parallel narrowband subcarriers. But existence of a blank space for a continuous communication system is not desired.4. These delay make noise and distort the beginning of the next symbol as shown. For high data rate transmissions. This is done by an Inverse Fast Fourier Transformation (IFFT). The advantage of using IFFT is that the system does not need N oscillators to transmit N subcarriers. Figure 1.1. and added to an OFDM symbol. one possibility is to shift the second symbol furthers away from the ﬁrst symbol. Figure 1.the symbol duration Ts is short.
1.2: Delay from front symbol.4 Principle of OFDM Transmission t Symbol 1 Symbol 2 Figure 1.3: Cyclic preﬁx insertion. 6 . Extention Extention t Copy this part to the front Symbol 1 Copy this part to the front Symbol 2 Figure 1.
1). 0. which are given by S−26. 1. 0. 0. 1. In IEEE 802. 1. 0. 1.The cyclic preﬁx is employed to reduce the ISI. 1. 0. −1 − j. 0. −1. 1 + j. −1. where the modulation is given by the element’s value. 0. a 10 MHz frequency bandwidth is used. 1 + j. time synchronization. 1. 0.11a. The 4 pilot signals are used for tracing the frequency oﬀset and phase noise. 0. 0. 0.11a. 0. 1.1 Deﬁnition of Key OFDM Parameters The IEEE 802. relates to the signal detection. −1. 1. 0. 1. −1. 1. The long training symbols (T1 and T2 ). −1. −1. which are located after the short training symbols. 1. 0. 1. −1. 1 + j. 0. 0. 0. −1. 0.26 = (13/6) {0. 0. − 1 − j. long OFDM training symbols are used. 1.11p 2. 0. −1. 1. 1. 0. are used for channel estimation. A short OFDM training symbol consists of 12 subcarriers. 0}. 1. 7 and 21. 0. 0. 0. 0. 1. 1. −1. 1. −7.11p are doubled compared with the IEEE 802.11p PHY has similar speciﬁcations as IEEE 802. −1. 0. 0. −1 − j. 1. 1. 0. −1. and coarse frequency oﬀset estimation. The factor (13/6) is for normalizing the average power in the OFDM symbol. 0. The total training length is 16 µs. −1. 1. −1.26 = {1. 0. 1. 1. 0. The long training symbols consist of 53 subcarriers that have a zero value at DC which are given by L−26.11p PHY uses 64 subcarriers OFDM that includes 48 data subcarriers and 4 pilot subcarriers. 1 + j. 0. 1. −1. GI2 is used as guard interval for long training sequence and GI is used as guard interval for OFDM symbols . 1 + j.11a. −1. 0. 0. −1 − j. −1. 0. −1.11p.Chapter 2 Wireless LAN According to IEEE 802. The short training symbols placed at the ﬁrst part of every data packet (t1 through t10 shown in Figure 2. 1}. 1.11a with some changes. −1. 7 . instead of 20 MHz bandwidth in IEEE 802. −1. 0. The IEEE 802. The doubled guard interval reduces ISI more than the guard interval in IEEE 802. 0. 1 + j. 0. and are located on subcarrier −21. 0. 1. To improve the channel estimation accuracy. 1 + j. 0. 1. −1 − j. thus all parameters in the time domain for IEEE 802. −1. 1.
8+3.1: OFDM training structure.6 µs 6. 36. the redundancy technique is introduced.11p PHY and IEEE 802. 1/3.2=4. 9.11p 3. 16 QAM.11a 6. 12.2 µs 16 µs 0.8+3. Timing Synchronize Channel and Fine Frequency offset Estimation RATE LENGTH Service+ DATA DATA Figure 2.1 illustrates the diﬀerence between IEEE 802. 4. 54 BPSK.2.8+2*3.8+3. 1/3.11a PHY (Source: [2]) Parameters Bitrate Mb/s Modulation Type Code Rate Number of Subcarriers Symbol Duration Guard Time FFT Period Preamble Duration Subcarrier Frequency Spacing Error Correction Coding IEEE 802.2=4. QPSK.15625 MHz K = 7 (64 states) Convolutional Code 2. 18.11p and IEEE 802.8 µs 3. where the modulation is given by the element’s value. 64 QAM 1/2.3125 MHz K = 7 (64 states) Convolutional Code IEEE 802. 64 QAM 1/2. For a binary 8 . 27 BPSK. QPSK. 12.8=8 µs 2*0. 24.Diversity Selection Coarse Frequency Offset Estimation. 16 QAM. diﬀerent modulation schemes and coding rates must be applied. 48. Table 2.0µs 0. Table 2.11a standard. Depending on the data rates. 18.2=8. 24.1: Comparisons view on the key parameters of IEEE 802. 1/4 52 8 µs 1.5.2 Deﬁnition of Physical Layer Coding 10*0.0µs t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 GI2 T1 T2 GI Signal GI Data1 GI Data 2 Signal Detect AGC. 1/4 52 4 µs 0.4 µs 32 µs 0. 9.0µs 0.2=4. 6.2 Deﬁnition of Physical Layer Coding The messages are inﬂuenced by interference and in order to detect and correct the errors in received signals.0µs 0.
Puncturing is used to create higher data rates.2. The coding rates of R = 1/2. The MAC is carried out to address some wireless communication events and control the medium access of the node. Because of medium access control. The constraint length of the encoder is 7 with bit rate 1/2. The basic access mechanism is called distributed coordination function.1. Puncturing is a procedure through which the number of transmitted bits is reduced and the coding rate is increased. The puncturing patterns are described in section 3. in order to reduce collisions.2 and in detail in [7].3 Medium Access Control The Medium Access Control (MAC) in IEEE 802. df ree is the minimum Hamming distance between two diﬀerent code words which is also called free distance.2 illustrates the convolutional encoder. an encoder is used in the transmission system to prepare data for transmission. and the medium access control (layer 2a).11p standard. or 3/4. several stations can use the same physical medium. Layer 2 is divided into two diﬀerent sublayers. The bits denoted as ”A” and ”B” are output of the encoder. Output Data A Inpot Data Tb Tb Tb Tb Tb Tb Outpot Data B Figure 2. 9 . the logical link control (layer 2b). 2. The minimum free distance of the code determines the performance of the convolutional code. Figure 2. 2/3.2: Convolutional encoder (k=7). which is used in the IEEE 802.11 p. A binary convolutional encoder is one kind of block code. that correspond to the desired data rate had been used in 802.11p is the second layer of the lowest protocol layer of the network architecture based on the Open System Interconnection (OSI) model. The convolutional encoder uses the generator polynomials g0 = 133 and g1 = 171 in octal mode.3 Medium Access Control block code.
3. • IFS: Two Inter Frame Spaces IFSs (Distributed IFS and short IFS) are deﬁned as time interval between frame. DATA and ACK packets. This packet is ﬁlled with information considering the source. [8]. Figure 2. • CTS: If the medium is free the destination station will respond with a control packet that is called Clear to Send (CTS). destination. A CSMA protocol is implemented when a station willing to transmit a data packet. • ACK: An Acknowledgment (ACK) packet is sent from the receiver back to the transmitter mentioning the successful completion of the data exchange. to present the delay between sending and receiving RTS.3 deﬁneds a Carrier Sense multiple access protocol. This protocol 10 . [9]. the station has to transmit the data packet to the receiver with a delay. CTS.3 Medium Access Control 2. [8].1 The Basic Access Method Distributed coordination function is basically a Carrier Sense Multiple Access (CSMA) in order to avoid collisions. To deﬁne the protocol.2. Since the medium is busy due to transmission of other stations. and duration of the transmission.3: A CSMA protocol. some acronyms must be introduced as follows: • RTS: A station that wants to transmit a packet ﬁrstly transmits a short control packet that is called Request to Send (RTS). Transmitter Receiver DIFS RTS SIFS CTS SIFS DATA SIFS ACK Figure 2.
5 and 2. 2.2 Frame types There are three main types of frames: 1. Data frames: Used for data transmission.4: RTS frame format. control frames and management frames) and depending on the frame type the duration is different. 2.3. virtual carrier sense mechanism is deﬁned as follows: A station transmits a data packet. Octes : 2 2 6 6 4 Frame Control Duration RA TA CRC MAC Header Figure 2. the Network Allocation Vector (NAV) which marks the medium as busy until the end of the protocol.3. Collision occurs when two or more stations transmit data at the same time . then station decides to transmit a small packet (RTS). data frames.6 show some of the most common frames formats. but they do not belong to the upper layers.4. Management Frames: Are used to exchange management information. it will be sure that the data exchange was successful.g.3 Most Common Frame Format The Figure 2. Once the receiver has received the data successfully. A station that receives one of these frames starts a timer. It can be the duration value that is used for NAV calculation as deﬁned 11 . if the medium is free for a speciﬁed interval time (DIFS). The received CTS at source station shows that the receiver is ready to receive data. In order to overcome these problems. 3. the receiver will wait SIFS units of time. and sends a response to the source station as a small packet (CTS). 2. [8]. The frame control deﬁnes the type of frame (e.3 Medium Access Control can be useful if the medium is not loaded heavily. and then send a response to source as a small packet called ACK. The source station will wait another SIFS unit and then sends the data packet. 2. The receiver answers with a delay (SIFS time units ). When the source receives ACK. Control Frames: To address some wireless communication phenomena and control medium access of nodes in order to reduce collisions. If the sender does not receive the ACK then it will continue transmitting the data until it gets a response.2.
[8]. 12 .6: CTS frame format. RA and TA are the receiver address and transmitter address and CRC is used for cyclic redundancy check.3 Medium Access Control Octes : 2 2 6 4 Frame Control Duration RA CRC MAC Header Figure 2.5: ACK frame format. in the virtual carrier sense mechanism. Octes : 2 2 6 4 Frame Control Duration RA CRC MAC Header Figure 2.2.
Figure 3. MATLAB SIMULINK is used.1 illustrates the MATLAB/SIMULINK simulator architecture.11a to obtain IEEE 802. I started from an available MATLAB/SIMULINK model.Chapter 3 MATLAB/SIMULINK IEEE 802. [1].11p Physical Layer Model In order to develop the PHY of IEEE 802. according to IEEE 802.11p. especially in the telecommunication ﬁeld.1: MATLAB/SIMULINK simulator architecture. The IEEE 802. Figure 3.11p model represents a baseband model for the physical layer in a Wireless Local Area Network (WLAN).11p PHY. MATLAB has a large number of libraries and tool boxes. 13 .
3.1.11p baseband model the padding is employed for truncating the input signal along column size. This mode has to be entered to the data source to create the binary data.2 illustrates the subsystem of modulator. it means. This mode is created in adaptive modulation control according to the SNR estimation at the receiver. rows or both of them according to the speciﬁed values.1 VariableRate Data Source Binary data is created according to a predeﬁned mode. 14 .1 Transmitter Side 3. The modulator is subdivided as following: • • • • • • Padding Convolutional encoder Puncturing convolutional codes Matrix interleaver General block interleaver Rectangular QAM Figure 3. rows are in the frequency domain and columns are in the time domain.1 Transmitter Side 3.2 Modulator IEEE 802.3. In this system each row is equal to one subcarrier. The speciﬁed output dimension is the number of bits per block that is diﬀerent according to the diﬀerent data rate and corresponding diﬀerent code rate and modulation scheme.11p OFDM PHY includes diﬀerent data rates which are selected according to the output of adaptive modulation. In the subsystem of data source a buﬀer exists whose output is according to the maximum bits per block which are chosen in the simulation parameter list. Padding The Padding block changes the dimension of input matrix along its columns. The system uses diﬀerent modulation schemes due to diﬀerent data rates. Figure 3. In IEEE 802.2: Subsystem of modulator.1.
Matrix Interleaver and General Block Interleaver Interleaving can be employed in digital data transmission technologies to mitigate the eﬀect of burst errors. k.1/2. The output elements will be according to the puncture vector. Raw data stream Data Block Interleaved data stream Figure 3.(m + 1) is introduced as constraint length to deﬁne a convolutional encoder in Matlab Simulink. due to a burst error. one convolutional code with one puncture vector [ 110110 ] and constraint length 7 can be used. k. This example explains that only a small part of each code word is distorted with interleaving. number of output bits. Code Generator) In this system the Trellis structure is poly2trellis (7.3. see [10].3 illustrates the eﬀect of interleaving at the transmitter. then the k th element of the input vector is represented in the output vector. the bits in one code word are interleaved before being transmitted. The following example illustrates creating new code rates from 1/2 code rate: To create a 3/4 code rate from a 1/2 code rate. When too many errors exist in one code word. [11]. A poly2trellis function is used to convert generator polynomials to Trellis structure. To reduce the eﬀect of burst error. which is the code rate of the convolution encoder. The interleaving 15 . the decoding of a code word can not be done correctly. [171 133]). Puncturing Convolutional Codes Generation of diﬀerent code rates from 1/2 code rate. which means that a burst error can not disturb a huge part of one code word. The third and sixth element from the input vector are removed according to the puncture vector. m. is implemented through puncturing. The k th element of the input vector will be removed if the k th element of puncture vector is zero. When interleaving occurs the place of bits will change.3: Transmission with interleaving.1 Transmitter Side Convolutional Encoder A convolutional encoder is carried out for coding of the transmitted bits. T rellis = poly2trellis(Constraint Length . Figure 3. Therefore the bit rate resulting from puncture vector is 3/2. On the other hand if the k th element of puncture vector is equal to one. ﬁnally bit rate will be 3/4 = 3/2. the number of input bits. so the decoding of code word can be done correctly. Convolution codes have three main parameters. n. and the number of memory register.
3. deﬁned in [7]. the second two elements of output are the second column and the third two elements of output are the third column. 16 . The second step is mapping the adjacent coded bits alternately onto signiﬁcant bits of the constellation that is implemented with the general block interleaver. Finally the matrix interleaver block rewrites the elements column by column. The second step of interleaving is implemented with the general block interleaver that changes the place of input elements according to the elements vector. In this system the number of rows and columns is given by: Interleaver Rows = 16 Interleaver Columns = Number of transmitted bits per block / interleaver Rows Figure 3.5 illustrates an example of the current process of the matrix interleaver block. According to the elements vector. The ﬁrst three elements of input vector are the ﬁrst row of matrix and the second three elements of input vector are the second row of matrix.5: Matrix interleaver. The parameters of the matrix and general block interleaver of the IEEE 802. the ﬁrst element of output is the forth element of input. Figure 3.11p baseband model are following the standard.6 explains this process. The ﬁrst step is mapping of the adjacent coded bits into the nonadjacent subcarriers that is implemented with the matrix interleaver. Therefore the ﬁrst two elements of output are the ﬁrst column of matrix. the third element of output is the third element of input and the fourth element of output is the second element of input. the second element of output is the ﬁrst element of input. 1 2 3 4 5 6 Matrix interleaver 2by3 123 456 1 4 2 5 3 6 Figure 3. The dimensions of input vector in matrix interleaver convert to a 2 by 3 matrix. The input vector is a column vector.1 Transmitter Side Sudden burst of noise causing errors Interleaved data stream Reassembled stream Figure 3. Matrix interleaver interleaves the input vector according to the speciﬁed row and column.4: Transmission with a burst error and interleaving. in this SIMULINK model is deﬁned by two steps.
6: General Block interleaver.1. The sample time and the number of samples per frame for PN sequence Generator is deﬁned as follow: Sample time = the period of the Block/OFDM symbol per frame Samples per frame = OFDM symbol per frame 3.1 Transmitter Side [40 32 59 1] General Block Interleaver [4 1 3 2] [1 40 59 32] Figure 3.11p baseband model a Graycode is used. Four diﬀerent modulation types are implemented: • • • • BPSK QPSK 16 QAM 64 QAM 3.3. The Pseudorandom Noise (PN) sequence generator block is carried out creating the pilot subcarriers. the modulator is using a reshape block. In the IEEE 802. Rectangular QAM The rectangular QAM block is applied to indicate how the binary words are assigned to points of the signal constellation. 3. The output vector is a number of data subcarriers by OFDM symbol per frame. The long training symbols consist of 53 subcarriers that have a zero value at DC subcarrier.1.11p has four pilot subcarriers.−7.3 OFDM symbols To convert a block of N serial data symbols (each has a duration of Ts ) into a block of N parallel data symbols (each has a duration of T = N Ts ). The long training symbol is deﬁned in 2. The location of pilot subcarriers is −21. The pilot signals are used for tracing frequency oﬀset and phase noise. 7 and 21.4 Pilot insertion Each OFDM symbol in IEEE 802.1. 17 . Four long OFDM training symbols are used instead of two long training symbol in this system.1.5 Preamble Preamble insertion is used for channel estimation in our model in order to improve the channel estimation accuracy.
The padding values are equal to zero. Then a parallel to serial convertor is used to transmit time domain samples of one symbol.7 shows the subsystem of this block. Figure 3.1. Figure 3. which means that bandwidth can be saved signiﬁcantly.8 IFFT and FFT An inverse Fourier transform converts the frequency domain data stream into the corresponding time domain. [11].6 Assemble OFDM Frames The assemble OFDM frames is applied in order to insert the pilot and training symbols into the OFDM symbols. 3. The Fast Fourier Transformation (FFT) is used to convert data in time domain to the frequency domain at the receiver.1. Four pilots are inserted between the subcarriers and then training sequence is added to the subcarriers. where the speciﬁed dimension of the output is the number of points of the IFFT block.7: Assemble OFDM frame subsystem.8 illustration the process. 18 . 3.7 Padding The Pad block extends the input vector along its columns.1.3.1 Transmitter Side 3. inserted at the end of the columns. In this situation subcarriers can be closer together. Figure 3. The serial to parallel block convertor is placed to convert this parallel data into a serial stream to obtain the original input data. IFFT block allocates the diﬀerent orthogonal subcarrier for transmitted bits and thus no interference exists between subcarriers.
.3.d2. 3. ..dN1 dN1 bN1 Transmit timedomain samples of one symbol Data coded in frequency domain:One symbol at a time Data in time domain: One symbol at a time d0' d1' d2' . 3. 3. ..d1'..10 Multiplex OFDM Frame The multiplex block is the last block in the transmitter part to convert the signal from parallel to serial and to transmit timedomain samples of one symbol. .2 Radio Channel b0 b1 b2 . .2 Radio Channel For the ﬁrst simulation a simple Additive White Gaussian Noise (AWGN) channel model is used.dN1' S/P Receive timedomain sample of one symbol Serial to parallel converter FFT Fast Fourier transform bN1 Figure 3.2.1... frequency bit . . . dN1 b0 b1 b2 . . This block produces a complex output signal when the input signal is complex. . In order to get a more realistic channel model for vehicular scenarios.8: IFFT/FFT description. . . . . . . following by simulations with multipath Rayleigh fading together with AWGN model. .d1. A selector block is applied as a cyclic preﬁx inserter to insert the last 16 subcarriers into the beginning of the OFDM symbols. . 3. .. . d0 d1 d2 . . . .1. In this model the variance is speciﬁed from the port 19 . . independently . . d0’.. Decode each .. . . . . . . an AWGN channel is added to the input signals. . . . IFFT Inverse fast Fourier transform P/S Parallel to serial converter d0. . we use a speciﬁc channel simulator called WinProp.9 Cyclic Preﬁx Cyclic preﬁx is used as a guard interval to mitigate the eﬀect of ISI due to the multipath propagation.. .1 AWGN Channel To implement the eﬀect of AWGN on the input signal. .
10: Multipath Rayleigh fading channel. two parameter dialogs are speciﬁed.9.3. 20 . Since a transmitted signal propagates along several paths in multipath channel to reach to the receiver. 3.2 Rayleigh Fading Channel with AWGN The multipath Rayleigh fading is added to an AWGN channel. in order to calculate the variance of the noise. the delay vector is used to specify time delay for each path and the gain vector is used to specify the gain for each path at each delay. [12]. it may lead to diﬀerent time delays. Figure 3. as shown in Figure 3.9: AWGN channel.2.2 Radio Channel that inserts SNR. In the block. Figure 3. The number of paths is according to the length of the delay vector and the gain vector which must have the same length.
cars. receiver... result.) and setting the time variant properties of the objects. Low cost and fast determination of a channel impulse response.3.11: Process for the creation of a time variant scenario. SiMan is used for the computation in time variant scenarios. where some properties like transmit power.2 Radio Channel 3.. ProjectMan is used to create a project with all settings related to the computation (transmitter. Building Addition of time variant objects and their motion data WallMan Definition of he project settings ProjectMan Simulation of the scenarion SiMan Visualization of the results ProMan Figure 3. Figure 3. 5. To create a realistic channel for V2V communication with WinProp .). the following steps can be passed: 1. The channel from the WinProp simulations is used instead of the simpler model (Rayleigh fading and AWGN) in order to achieve more realistic results. WallMan is used to create additional objects (buildings. ProMan is used for visualization of prediction results. antenna pattern and antenna placement of transmitter 21 . The introduction explains that the software package can be used for predictions in time variant scenarios. and creating the scenarios easily are some advantages of WinProp simulator..11 illustrates the process for the creation and simulation of a new scenario.. ProjectMan is applied to create a project. center frequency. 4. 2. Database with street and curves StreetMan Addition of static elements.e. StreetMan is used to create road courses.g. [13]..3 Advanced Model Using WinProp Simulator WinProp [3] is a Software package for the simulation of electromagnetic waves and radio systems in static and time variant environments. WinPropTime Variant scenarios In this section a short introduction to the time variant usage of WinProp is given. 3.2.
3. type of polarization. The convolution for continuous time can be expressed by. Minimum delay. Setting the digital ﬁlter to the FIR ﬁlter 22 . y(t) = ∞ −∞ x(t − τ )h(t. SiMan is used for the computation time variant scenarios. Also mat ﬁle contains the date and time of creation of mat ﬁle. Using the digital ﬁlter to convolve the transmit signal with the channel matrix to obtain the output signal.1) x(t)= Input signal y(t)= Output signal h(t)= Channel impulse response 2. and type of output ﬁles after simulation are other properties that are deﬁned in ProjectMan. In addition the properties of receiver like antenna pattern and antenna placement are deﬁned in ProjectMan. After ﬁnishing the simulations. 1. τ )dτ.2 Radio Channel are deﬁned. the results appear as diﬀerent ﬁle formats: • cir ﬁle: Channel impulse response • fpf ﬁle: Field strength result • fpp ﬁle: Power result ﬁle • fpl ﬁle: Path loss result ﬁle • str ﬁle: Propagation paths for visualization in ProMan • mat ﬁle: A speciﬁc MATLAB output The mat ﬁle contains the impulse response matrix of the channel and it contains following variable: • • • • • • • • • Matrix Date Delta delay Frequency Maximum delay Minimum delay Power Snapshots Time The elements of the matrix are ﬁeld strength in µV/m. (3. center frequency is in MHz and transmit power is in dBm. [4]. The number of snapshots. maximum delay and delta delay (resolution) are in ns. The following steps describes the implementation of the channel from the mat ﬁle into the SIMULINK PHY model.
3.. A selector block is used to remove the 16 subcarriers that are inserted into the beginning of the OFDM symbols... 4. ... 3..... 1....3 Receiver Side Finite Impulse Response (FIR) digital ﬁlter operates by convolving the input signal x(n) with the ﬁlter’s impulse response h(n) to ﬁnd the output signal as described in Figure 3.2 Remove Cyclic Preﬁx In the receiver the inserted cyclic preﬁx must be removed.12.. where k = 0........ therefore the channel matrix is converted with a buﬀer block to a column vector... + bM x(n − M ) M = m=0 bm x(n − m) = ∞ m=−∞ h(m)x(n − m) = (h ∗ x)(n) (3.... + + + .. In this model a time varying ﬁlter is used that the coeﬃcients of ﬁlter change once per input frame............ Adding AWGN to the output of the time variant ﬁlter.. x(n − k) is the input with a delay equal to k. X(n) Z 1 Z1 Z1 . The second input port of the FIR ﬁlter is used for the insertion of the channel impulse responses as coeﬃcients of the ﬁlter.... · · · ........ to obtain the original input data. 2.......3........1 Demultiplex OFDM Frame To convert a signal from serial to parallel... The input of this port can not be a matrix..... to obtain a frame based input for the FIR ﬁlter..... 23 . + y(n) Figure 3. [14]..... Z 1 b0 b1 b2 b3 bM .. A reshaping block is a subsystem of this block and is employed to produce a matrix out of the input vector.3 Receiver Side 3.. 3. a demultiplex block is used..2) 3...3..... Specifying FIR ﬁlter as time varying ﬁlter The coeﬃcients of time variant ﬁlter change with time...12: FIR ﬁlter. M and y(n) = b0 x(n) + b1 x(n − 1) + b2 x(n − 2) + b3 x(n − 3) + .
The estimated channel is the input of the equalizer gain block and the output is the inverse of the channel estimation. Figure 3. which are also used at the transmitter. F (f ) is channel frequency response and C(f) is inverse of the channel frequency response.3. a zeroforcing equalizer is used that applies the inverse of the channel frequency response. Figure 3.3 FFT A FFT block computes the fast Fourier transformation (FFT) along each column for all input matrices to convert a time domain signal to frequency domain. This estimation is used for the zeroforcing equalizer.13 illustrate a zeroforcing equalizer.3 Receiver Side 3. The channel estimation is done by dividing the received training symbols through the true training symbols. 3. F(f) C(f) * f f = f Figure 3. the received signal is splitted up into the training symbols and data symbols. The subsystem of the equalizer gaing is shown in Figure 3.3. and the received symbols. the training symbols. According to the zeroforcing equalizer 24 .13: ZeroForcing equalizer.3. As depicted in Figure 3. [5].15. The combination of channel and zero forcing equalizer output gives a ﬂat frequency response with linear phase to obtain the transmitted signal.14. The frequency domain equalizer block has two input signals.14: Subsystem of equalizer.4 Frequency Domain Equalizer To restore the transmitted signal.
25 .3. 3. to achieve the original signal. the estimated signal before the channel can be achieved.5 Disassemble OFDM Frame In this part the data subcarriers are separated from pilot subcarriers and the N parallel data symbols are converted to the N serial data symbols. 3.3. Figure 3.16: Disassemble OFDM frame. technique with combination of received data and zero forcing equalizer output (inverse of channel estimation).3.17: Subsystem of demodulator bank.3 Receiver Side Figure 3.6 Demodulator Bank The demodulator subsystem performs the inverse tasks of the modulator subsystem. Figure 3. Figure 3.17 illustrates the subsystem of demodulator bank.15: Subsystem of equalizer gains.
This happens because of there are four elements one in the insert zero vector. The most common metric is the Hamming distance metric. With the puncture vector at the transmitter. deﬁnes a metric for each path and makes a decision based on this metric. The method of making adaptive modulation in this model is according to the estimated SNR. [10].3/4. the code rate after the insert zero block is 1/2 = 2/3. each with four elements. diﬀerent code rates were created and at the receiver a zero insertion block is used to convert the code rates to the base code rate. 3. 26 . If the input vector is [ 1 3 4 5 7 9 10 11 ] and the insert zero vector parameter be a vector like [ 1 0 1 1 1 0 ]. This results in a 2/3 code rate. [4]. The implementation of adaptive modulation is according to the channel information that is present at the transmitter. [1 3 4 5 7 9 10 11] Insert Zero [1 0 3 4 5 0 7 0 9 10 11 0] Figure 3.18: Zero insertion. When two paths come together on one node.3. the input vector will be divided into two groups.3 Receiver Side Zero insertion The opposite of puncturing is zero insertion. a bit rate will be speciﬁed and then data source generates binary data according to the speciﬁed data rate in adaptive modulation control. CodeGenerator) In this system trellis structure is poly2trellis[ 7. the block inserts zeros after the ﬁrst and last elements of each group of four elements.3. The Figure 3. The Viterbi decoder. The number of trellis branches is deﬁned as trace back depth that is 32 in this model. If the code rate of the encoder is 3/4. trellis =poly2trellis(ConstraintLength . the shortest hamming distance is kept. It means ﬁnding the most probable transmitted symbol stream from the received code word.18 explains this process. 1/2. Based on this. The following example determines the inverse process of puncturing. To deﬁne a convolution decoder in MATLAB simulation a poly2trellis function is used to covert convolution code to trellis description. Viterbi decoding The Viterbi decoder block works according to the maximum likelihood decoding.7 Adaptive Modulation Control Adaptive modulation systems improve the rate of transmission.[171 133] ].
by comparing the received data with transmitted data.8 Evaluation of Reliability Modules Bit Error Rate The error rate calculation block calculates the bit error rate. Packet error rate is taken over the last 50 frames.3 Receiver Side 3. It has three inputs. 27 . Tx and Rx port that are used to accept transmitted and received signals and the third port is used to indicate the related frame for computation.3.3. Packet Error Rate In this block the numbers of errors in the packet will be divided by the number of packets for the calculation of packet error rate.
Therefore we deﬁne two V2V scenarios. Further there is a third car. time variant properties of objects and antenna characters. a truck. only a guard rail and the street are modelled. 4.1 depicts this scenario. In scenario I the cars are going in opposite directions and in scenario II the cars are going in the same direction around a corner. There is one lane in each direction. Figure 4.Chapter 4 Deﬁnition of Vehicular Scenarios The radio propagation characteristics are determined by many factors. The complete scenario is deﬁned with following parameters: Transmitter: • • • • • • • • Tx power: 18 dBm Center frequency: 5. which is going ahead the receiver car. The transmittercar is a orange sedan and the receivercar is a blue transporter. such as operating frequency band. in this section is detail.and receivercar are going in opposite directions. signal bandwidth. In order to simplify the scenario there is no building next to the road. which were simulated with the simulation tool WinProp.5 m Car type: Sedan (blue) Car speed: 15 m/s Receiver: • Antenna pattern: Isotropic 28 .1 Scenario I: Cars Are Going in Opposite Directions In this scenario the transmitter.9 GHz Antenna pattern: Isotropic Antenna placement: Roof of the car Polarization: Vertical Height of the antenna: 1. In both scenarios the antennas are placed on the top of the vehicles and an isotropic radiator is assumed for the transmit and received antennas.
The street is not straight as in Scenario I. • Minimum delay: 0 ns • Maximum delay: 10000 ns • Delay resolution: 100 ns 4.2. One further car is going into the opposite direction. maximum delay and delay resolution has to be set.2 Scenario II: Cars Are Going in the Same Direction In this scenario the transmitter. The complete scenario is deﬁned with following parameters: Transmitter: • • • • • • Tx power: 18 dBm Center frequency: 5. There are several buildings next to the street. see Figure 4. which is 10 MHz in IEEE 802.11p. The delay resolution is equal to the inverse of the signal bandwidth.and receivercar are going in the same direction. Further the minimum delay.9 GHz Antenna pattern: Isotropic Antenna placement: Roof of the car Polarization: Vertical Height of the antenna: 1. The start time is equal to 0 s and end time is equal to 5 s.2 Scenario II: Cars Are Going in the Same Direction • • • • • Antenna placement: Roof of the car Polarization: Vertical Height of the antenna: 1. the following settings have to be speciﬁed. The maximum delay is equal to the time interval. but there is also one curve.5 m Car type: Transporter (orange) Car speed: 15 m/s Environment and other cars: • • • • One lane in each direction Straight street Guard rails along the street Third car (truck) ahead the receivercar with speed 10 m/s In order to deﬁne the number of snapshots in time.4. The interval time shall be equal to the frame length of the SIMULINK PHY model. The start time. end time and interval time (time resolution).5 m 29 . which is 24×8µ s = 192µ s. As in Scenario I there is one lane per direction and a guard rail is along the street.
20 m. 10 m/s and 10 m/s are related to the distance 0 m. 10 m/s. 7. 68.50 m and 81 m respectively Environment and other cars: • One lane in each direction • Street with one curve • Guard rails along the street • Seven buildings next to the street • Third car is going in opposite direction with speeds: 12 m/s. • Car type: Sedan (blue) • Car speeds: 10 m/s.1: Vehicular scenario I .50 m and 65 m respectively Receiver: • Antenna pattern: Isotropic • Antenna placement: Roof of the car • Polarization: Vertical • Height of the antenna: 1. 25 m. 40 m. 11 m/s.4. 10 m/s and 15 m/s are related to the distance 0 m.2 Scenario II: Cars Are Going in the Same Direction Receivercar Transmittercar Figure 4.5 m • Car type: Transporter (orange) • Car speeds: 14 m/s. 12 m/s and 13 m/s are related to the distance 0 m. 15.50 m and 49 m respectively 30 . 10 m/s. 53.
4.2 Scenario II: Cars Are Going in the Same Direction The time and delay paramters are the same as in Scenario I. • • • • • • Start time: 0 s Stop time: 5 s Time interval: 192 µs Minimum delay: 0 ns Maximum delay: 10000 ns Delay resolution: 100 ns Transmittercar Receivercar Figure 4.2: Vehicular scenario II . 31 .
1 illustrates a comparison of simulation result and theoretical curve of error probability for BPSK modulation with an AWGN channel. Pb = Q 2Eb N0 . The correcting factor for BPSK can be expressed as 80/64 that is added to the SNR in order to obtain Eb /N0 .2) The ﬁrst step of comparison the SIMULINK results with theoretical error probability is for the validation. The BER versus Eb /N0 for the 3 Mbps data rate (BPSK modulation scheme with 1/2 coding rate) and theoretical curve are close together. the BER versus SNR can be the reference for the whole vehicular wireless network. Figure 5. Figure 5.5 at 5 dB Eb /N0 . Because of that.2 shows the PER versus Eb /N0 plot for QPSK modulation with 1/2 coding rate.11p standard.1) In this formula Eb is the bit energy. for uncoded BPSK is. which are carrying no information.1 Simulation Result of AWGN Channel In this part the comparison of simulation result with a theoretical curve of error probability for distinct modulation schemes. The theoretical bit error probability . the fewer the errors probability in the simulation results. which validates our simulations. in order to reﬂect the cyclic preﬁx and pilotcarriers. N0 is the noise power spectral density and Q(x) is deﬁned by 1 Q(x) = erf c 2 x √ 2 . (5. and followed by a moderate drop to obtain a low value of PER in larger than 7.5 dB 32 . (5. In physical layer investigation of the IEEE 802. [15]. The vertical axis shows the bit error probability and horizontal axis shows the Eb /N0 . For validation of results a correcting factor is used.Chapter 5 Discussion of Simulation Results 5. In general. [15]. BPSK and QPSK is discussed. the higher the Eb /N0 . The PER for simulation results is more than 0.Pb . a part of energy of a transmitted OFDM symbol is lost.
5 4 4. Eb /N0 .5 at 5 dB Eb /N0 and is followed by a moderate drop to obtain a low value of PER in 10 dB Eb /N0 . 10 0 Coded Packet Error Rate 10 −1 10 −2 10 −3 QPSK over AWGN with 1/2 code rate 10 −4 3.1 Simulation Result of AWGN Channel 10 −1 BPSK over AWGN simulink model BPSK over AWGN theoretical −2 Uncoded Bit Error Rate 10 10 −3 10 −4 10 −5 10 −6 0 2 4 6 Eb/N0 (dB) 8 10 12 Figure 5. The PER for simulation results in [2] illustrate that PER is more than 0.5.5 dB Eb /N0 .1: Error probability for BPSK modulation scheme. 33 .5 Figure 5.5 7 7.which validates our simulations. Comparison between simulation results for PER over 2304 last frames and simulation result in [2] demonstrates that in both two cases the values of PER is decreasing in the duration between 5 dB Eb /N0 until more than 7.5 5 5.2: Packet error probability for QPSK modulation scheme.5 6 Eb/N0 (dB) 6.
5 at −30 dB SNR for the QPSK modulation. which validates our simulations. Furthermore QPSK modulation scheme with 1/2 coding rate is used as modulation technique.3: Error probability for BPSK modulation scheme. for QPSK modulation scheme is given by equation .2 Simulation Result of a Multipath Rayleigh Fading and AWGN Channel In this part the comparison between simulation results and theoretical error probability is shown. Bit error probability .5. A Rayleigh fading channel with AWGN for uncoded transmission is considered. see section 5. The start point of simulation result curve is BER of 0.[16].3 demonstrates the simulation result of a QPSK modulation with 1/2 coding rate over a AWGN and fading channel.Pb . The correcting factor in this case is the same as for BPSK . 10 0 Uncoded Bit Error Rate 10 −1 10 −2 10 −3 QPSK over AWGN and fading simulink model QPSK over AWGN and fading theoretical 10 −30 −4 −20 −10 0 10 SNR (dB) 20 30 40 Figure 5.3 illustrates the simulation result of a QPSK modulation over a AWGN and fading channel is matched with theoretical curve.1. This is followed by a slight drop in 0 dB SNR. BER continues to fall moderately until 30 dB SNR. when a low values of BER is obtained. and the number of tap delay is 1 and maximum delay is 100 ns.3) 34 .2 Simulation Result of a Multipath Rayleigh Fading and AWGN Channel 5. 1 Pb = 2 1− µ 2 − µ2 K 2k k k=0 1−µ2 4−2µ2 k . Furthermore Figure 5. (5. Figure 5.
079.5 4 4. which is 5% less than uncoded transmission.5 1 1.5 × 10−4 which demonstrate the provment of BER over coded transmission. 1 + SN R (5.4: Bit error probability for BPSK modulation scheme.4 Packet Error Rate Calculation Figure 5.5 5 Figure 5. 5.5 dB Eb /N0 is 4.5 × 10−2 and BER for coded transmission at 2.6 show the PER performance versus SNR of the coded and uncoded transmission over 50 last frame. 10 −1 Bit Error Rate 10 −2 10 −3 BPSK over AWGN uncoded BPSK over AWGN coded 10 −4 0.5 dB Eb /N0 is 5.4 shows the BER performance versus Eb /N0 of the coded and uncoded transmission over an AWGN channel for BPSK modulation scheme. when BER reaches to 7 × 10−3 . The start point of BER versus Eb /N0 for uncoded transmission is BER of 0.4) 5.3 Comparison between Coded and Uncoded AWGN Channel Figure 5.5 and 5. Figure 5. The higher the SNR.024 at 1 dB Eb /N0 . The BER is decreasing moderately until 5 dB Eb /N0 . BER for uncoded transmission continues to fall sharply in comparison with coded transmition.5.5 3 E /N (dB) b 0 3. the fewer the PER 35 .4 describes that the simulation results with coding have lower error compared with the simulation results without coding. The start point of the BER versus Eb /N0 for coded transmission is BER of 0. because of error detection and correction with convolutional encoder. BER for uncoded transmission at 2.3 Comparison between Coded and Uncoded AWGN Channel where K is the number of bit symbol and µ is µ= SN R .5 2 2.
5 and 5.4 Packet Error Rate Calculation in the simulation results. 36 .5. Furthermore comparison between Figure 5. 100 PER for coded transmission 80 60 40 20 0 −1 PER(%) 0 1 SNR (dB) 2 3 4 Figure 5.6: Packet Error Rate of BPSK over AWGN channel. 100 80 60 40 20 0 −2 PER for uncoded transmission PER(%) 0 2 4 SNR (dB) 6 8 10 Figure 5.6 illustrates that coded transmission has lower PER compared with the simulation results for uncoded transmission.5: Packet Error Rate of BPSK over AWGN channel.
The RMS delay spread in this case is less than cyclic preﬁx.0 10. Table 5. The model is set up according to the Table 5.7. Mean Excess delay and RMS delay are. Mean Excess Delay = RMS delay = K k=0 p(τk ).5 Simulation Result of Multipath Rayleigh Fading and AWGN Channel with Diﬀerent Tap Delay Since in the multipath channel signals are reﬂected at multiple places.0 15. the time delay relative to the ﬁrst tap and the average power relative to the strongest tap. The model is speciﬁed by the number of taps.τ k K0 p(τk ) (5.0 9. Simulation results in this part are somewhat diﬀerent which is due to more than one tap delay in vehicular channel A.6) (τ 2 ) − (τ )2 τ2 where K is number of paths.510µ s > 1.6µ s (cyclic preﬁx) the RMS delay spread and mean excess delay are 604.0 1.5 Simulation Result of Multipath Rayleigh Fading and AWGN Channel with Diﬀerent Tap Delay 5. this is a Channel impulse response model based on a tappeddelay.5.1: ITU Vehicular Channel Model (Channel A)(Source: [17]) Tap 1 2 3 4 5 6 Relative delay(ns) 0 310 710 1090 1730 2510 Average power(dB) 0.0 20.[5].0 The comparison of BER for simulation results for vehicular channel A and QPSK over AWGN and fading SIMULINK model is illustrated in the Figure 5. To specify the channel with diﬀerent tap delays in fading channel a tapmodel is used which is called ITU vehicular channel A [17].1 ns and 466. This model is not allocated to V2V communication.5) (5.10 ns. they reach the receiver from several diﬀerent paths that may have diﬀerent lengths and diﬀerent corresponding time delays.1. Maximum excess delay in vehicular channel is 2.7) 37 . = K 2 k=0 p(τk )τK K k=0 p(τk ) (5. The correcting factor is the same as before.
Y = (H ∗ X) + N.E x2 . To simplify obtaining the SNR. E {N 2 } 1 .6 Advanced Channel Model using WinProp 10 0 Uncoded Bit Error Rate 10 −1 10 −2 QPSK over AWGN and fading simulink model QPSK over vehicular A model 10 −30 −3 −20 −10 0 10 SNR (dB) 20 30 40 Figure 5.mention in the last sections. 38 . SN R = Therfore E {H(x)2 } .5.8) equation 5.10) Equation 5. 5. E {N 2 } (5.8 deﬁne the relation between transmit signal and received signal.9) SN R = (5. In this part the normalization of impulse response matrix is taken into account to obtain more realistic result. to achieve more realistic results.12 are the deﬁnition of mean power. Distinct vehicular scenarios described in section 4 are used in this part. (5.9 is used to deﬁne a certain value of SNR.6 Advanced Channel Model using WinProp The channel resulting from WinProp is used in the SIMULINK model instead of the simpler model. the values of mean power (E {H(x)2 }) and Average transmit power (E {x2 }) is set to one. Equation 5.11 and Equation 5.[4].7: Error probability for QPSK modulation scheme. The resolution of the delay is deﬁned according to the Tx sample rate in Matlab simulation that is 100 ns.
E(y).8 and Figure 5.12) in MATLAB the impulse response channel matrix is divided to square root of mean power to normalize the (E {H(x)2 }) to one. τ )2 dτ dt. Figure 5.8 illustrates that the magnitude of E(z) is increasing until two cars reaching to each other at t = 2. Simulation Result of Advance Channel Model in Opposite Direction The MATLAB output of WinProp contains the ﬁeld strength in µ V /m in three components.8 illustrates the magnitude of E(z ) as a function of time. When two cars reach to each other. The magnitude is decreasing due to the increasing the distance between two cars after passing from each other.8: Magnitude of E(z) vector (dB). (5.11) ∆t h=−k l=0 h(k∆t. Figure 5.9 s. E(x). 120 Magnitude of E(z) vector (dB) 100 80 60 40 20 0 1 2 3 4 5 Time (s) Figure 5. the highest values of magnitude 39 .6 Advanced Channel Model using WinProp Continuous: M eanP ower = limT →∞ Discrete: 1 M eanP ower = limT →∞ T k L 1 T T /2 −T /2 0 ∞ h(t. (5. Figure 5. Due to the vertical polarization the magnitude of E(z) is more than the magnitude of E(x) and E(y). the magnitude of E(z) vector is decreasing and the bit error rate is increasing.9 shows the BER for vehicular scenario in opposite direction.5. l∆t)2 ∆τ.9 illustrates that when a fading accrues. The comparison between Figure 5. and E(z).
1.6 Advanced Channel Model using WinProp for E(z) occurs.6 0.5 3 0 1 2 Time (s) 3 4 5 Figure 5.4 0. and therefore a decreasing of values for BER are obtained.5 4 Data Rate 3. 4.8 0.2 0 0 1 2 3 4 5 x 10 −3 Time (s) Figure 5.2 1 BER 0. Then the values of BER increase due to the increasing the distance between two cars after passing from each other.4 1.5.11 illustrates the magnitude of E(z) as a function of time.10: Data rate of vehicular scenario in opposite direction . Simulation Result of Advance Channel Model in Same Direction Figure 5. Due to the vertical polarization the magnitude of E(z) is more than the magnitude of 40 .10 shows the diﬀerent values of data rate that generated according the adaptive modulation.9: BER of vehicular scenario in opposite direction. Figure 5.
14 and 5.5. we can see that pathloss is less than 60 dB and it is increasing to more than 100 dB in higher distances. and therefore an increasing of values for BER are obtained. The magnitude of E(z) decreases at t = 3.15 illustrate the received power and pathloss related to the transmitter that is on the roof of the orange car was showed in ﬁgure 4.7 s. The magnitude of E(z) is changing between 95 dB to 105 dB in the same direction of transmitter and receiver scenario. Received power is decreasing by distance and will not be understandable in blue area due to sensitivity level of ITS transceivers in practice. This phenomenon is illustrated in Figure 5.13 shows the diﬀerent values of data rate that generated according to the adaptive modulation. 41 . Figure 5.15.11: Magnitude of E(z) vector (dB) . Figure 5. Analyzing of received power for Advance Channel Model Figures 5. when the third car is passing from the transmitter and receiver in the opposite direction. The curve of E(z) in the same direction scenario is varying less than variation in opposite direction scenario.12 illustrates that when the third car is passing from the transmitter and receiver.11 and Figure 5. In the near area to the transmitter. The Path loss is increased by distance. the lowest values of magnitude for E(z) occurs.12 shows the BER for vehicular scenario in the same direction. The comparison between Figure 5.1. because of the higher variation of the Doppler shift in the opposite direction scenario. 120 Magnitude of E(z) vector (dB) 110 100 90 80 0 1 2 Time (s) 3 4 5 Figure 5.6 Advanced Channel Model using WinProp E(x) and E(y).
12: Bit error rate of vehicular scenario in same direction . 42 .5 4 Data Rate 3.5.6 Advanced Channel Model using WinProp x 10 12 10 BER 8 6 4 2 0 −5 1 2 Time (s) 3 4 5 Figure 5.5 3 0 1 2 Time (s) 3 4 5 Figure 5.13: Data rate of vehicular scenario in same direction . 4.
14: Received power. 43 .15: Path loss.5. Figure 5.6 Advanced Channel Model using WinProp Figure 5.
I compared my simulation results with theoretical curves of error probability to validation of my simulation result. The WinProp is a Software package for the simulation of electromagnetic waves and radio systems in static and time variant environments. I have integrated a raytracing based software tool.Chapter 6 Conclusions At this work a MATLAB simulation was carried out in order to analyze baseband processing of the proposed IEEE 802. Furthermore the value of BER for vehicular scenario in same direction is less than the values of BER for vehicular scenario in opposite direction because of the higher variation of the Doppler shift in the opposite direction scenario. In order to get a more realistic channel model for vehicular scenarios. which validates our simulation results. 44 . The comparison of simulation results over Rayleigh fading and AWGN channel with theoretical curve for QPSK by 1/2 coding rate illustrates that the simulation results and theoretical curve are matched together. The comparison between the simulation results of advanced channel model using WinProp in same direction scenario and opposite direction scenario illustrates that the curve of E(z) in the same direction scenario is varying less than variation in opposite direction scenario. WinProp.11p physical layer. for modelling time variant vehicular channel into SIMULINK. The comparison of BER versus Eb /N0 for AWGN channel and theoretical curve illustrate that simulation result over BPSK modulation scheme with 1/2 coding rate are close together. A description of the principles of the IEEE 802. Normalization of impulse response matrix is taken into account to obtain more realistic results. Diﬀerent channel implementations are used for the simulations. I use a speciﬁc channel simulator called WinProp.11p PHY and working experience with MATLAB SIMULINK and WinProp software was obtained through this thesis. For the ﬁrst simulation a simple AWGN channel model is used. following by simulations with multipath Rayleigh fading together with AWGN model.
Zang. 2001. 1996. F.11a WLAN model. communications blockset. Convolutional encoder.” June 2003.mathworks. 2005. Statistical Digital Signal Processing and Modeling. “Wireless LAN medium access control (MAC) and physical layer (PHY) speciﬁcations.11 MAC protocol.” B. M. July 1996.9 GHz.11a standard and system performance analysis. and M. Stibor. Pant and T. Abbt. Wiley. 1996. o o [12] www. M 1225. A. communications blockset. Fuertes. [13] R.11.” [2] Y. pp. and H. G. Brenner. AWE Communications GmbH.” 1997.mathworks. [6] E.com/access/helpdesk/help/toolbox/commblks. “A technical tutorial on the IEEE 802. Orfanos.Bibliography [1] http://www. S. Wireless Communications: Principle and Practice. Canada. “An error model for intervehicle communications in highway scenarios at 5. [10] www.Wiley and Sons. Proakis.11 protocol.” Final Report. Z. Wahl. 2005.” in International Workshop on Modeling Analysis and Simulation of Wireless and Mobile Systems. Wireless Communications. 1995. 2002. S. Digital Communications. G. 51. thesis. [9] P. W. Castelli. 2005. [4] A. Rappaport. Link¨ping. “IEEE 802. [17] TUR Rec.” AWE Communications GmbH. IEEEPress . “Guidelines for evaluation of radio transmission technologies (RTTs) for IMT2000. vol. Reumerman. university of Link¨ping. Quebec.awe com.com/access/helpdesk/help/toolbox/commblks. Stahel. L. [7] IEEE 802. Sweden.Sc.de/Automotive/. Clark. “OFDM PHY layer implementation based on the 802. 341–352. Limpert. 45 . [3] http://www. McGrawHill. “Simulation of a wireless network using the 802. AWGN channel.mathworks. [11] L. Mark. [15] E.” Breezecom Wireless Communications. [8] P. “An introduction to WinProp time variant. [16] J. Monson.” Bio Science. Wireless OFDM Systems: How to Make Them Work? Kluwer Academics Publishers. Molisch. Montreal. Guo. [5] T. [14] H. “Lognormal distributions across the sciences: Keys and clues.com/matlabcentral/ﬁleexchange/3540. NJ: PrenticeHall.
. .3 3. Subsystem of equalizer . . . . . . . . . . . . . . . . .7 3. . . . . . . . . . . . .4 2. . . . . .4 3. . . .2 5. . Assemble OFDM frame subsystem . . . . ACK frame format . . . . .2 5. . . . . . . . . . . . . . . . . . . . . . . . Convolutional encoder (k=7) A CSMA protocol . . . . . . . .5 3. . . . . . . RTS frame format . . . .6 3. . . . . . . . .1 5.11 3. . . . . . . . .1 1. . . . . . . . . . . . .18 4. . . . . . . . Vehicular scenario I . . . . . . . . . . . . . .3 2. . . . . . . Multipath Rayleigh fading channel . ZeroForcing equalizer . . . . . . . . . . . . . . . .10 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bit error probability for BPSK modulation scheme . . . . . . . . . . Packet error probability for QPSK modulation scheme Error probability for BPSK modulation scheme . . Transmission with a burst error and interleaving . . . . Vehicular scenario II . . . . . . . . . . . . . . . . . . . . .13 3. . . . . . . . .8 3. . . . . . . . Error probability for BPSK modulation scheme . .12 3. . . . . .2 1. . . . . . . . . . . . . . . . . . . . . . 3 6 6 8 9 10 11 12 12 13 14 15 16 16 17 18 19 20 20 21 23 24 24 25 25 25 26 30 31 33 33 34 35 MATLAB/SIMULINK simulator architecture . . . . . . . . .15 3. . .4 Multipath propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subsystem of demodulator bank . . . . . . . . . . . . . . . . . . . . . . . . . . . General Block interleaver . . . . . Subsystem of equalizer gains . . . . . . . . . . . . . Matrix interleaver . . . . . . . . . . . . . . . IFFT/FFT description . . . . .9 3. . . .2 2. . . . . . . . . . . . . . . . . . . . . . . .1 2. . . . . . . . . . . . Delay from front symbol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zero insertion . . CTS frame format . . . . . . . . . .1 3. Cyclic preﬁx insertion . . . . . . . . . . . . Subsystem of modulator . . . . . . . . . . . . . . . . . . . . . . . . . . OFDM training structure .2 3. . Disassemble OFDM frame . . . . . .5 2. . . . . . Process for the creation of a time variant scenario FIR ﬁlter . . . . . . . . . . . . . . .14 3. . . . . . . . . . . .16 3. . . . . .6 3. . . . . . . . . . . . . . . . . . . Transmission with interleaving . . . . . . . . . . . . . . . . . AWGN channel . . . . . . . . . . .3 5. . .17 3. . . . . 46 . . . .1 4. . . . .List of Figures 1. .3 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . Error probability for QPSK modulation scheme . . . . . .List of Figures 5. . . . .14 5. . . . . . . . . . . Path loss . . . .6 5. . . . . . . . BER of vehicular scenario in opposite direction . . .13 5. . . . . . . . . . . . . . .8 5. . .11 5. Bit error rate of vehicular scenario in same direction Data rate of vehicular scenario in same direction . . .15 Packet Error Rate of BPSK over AWGN channel . . . . 36 36 38 39 40 40 41 42 42 43 43 47 . . . . .10 5. . . . . .12 5. . .7 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 5. . . . . . . . . . . . . . . . . . . . . Magnitude of E(z) vector (dB) . . .9 5. . . . . . . Packet Error Rate of BPSK over AWGN channel . Received power . . . . . . . . . . . . . . Data rate of vehicular scenario in opposite direction Magnitude of E(z) vector (dB) . . . . . . . .
. . . . ITU Vehicular Channel Model (Channel A)(Source: [17]) . . . .List of Tables 2. . .11a PHY (Source: [2]) . . . .1 5. . . 8 37 48 . .1 Comparisons view on the key parameters of IEEE 802. . . . .11p PHY and IEEE 802.