Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Cpre537x Final Project

Digital Communication
Via Multipath Fading Channel

Zhiwei Zeng
Instructor: Dr. Russell
November 2000

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Abstract
Multipath fading is a common phenomenon in wireless signal transmission. When a
signal is transmitted over a radio channel, it is subject to reflection, refraction and
diffraction. The communication environment changes quickly and thus introduces more
complexities and uncertainties to the channel response. This simulator offers a better
understanding of this phenomenon. In order to observe the effects of multipath fading
channel on the transmitted signal, a whole digital communication system simulator was
developed. Three kinds of digital communication systems: baseband transmission via
additive white Gaussian noise (AWGN) channel, passband transmission via single
AWGN channel, and passband transmission via multipath fading channel, are simulated.

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Contents

1 Introduction

2 Baseband transmission via AWGN channel

2.1 Signal source
2.2 Sampling
2.3 Waveform encoding and decoding
2.3.1 Pulse Code Modulation (PCM)
2.3.2 Simple DPCM
2.3.3 General DPCM
2.3.4 Delta Modulation
2.4 Error control coding
2.4.1 Block codes
2.4.2 Convolutional codes
2.5 Line coding
2.6 Baseband filtering
2.7 Eye pattern
2.8 Optimum receiver in AWGN channel
2.9 Lowpass filtering

1
2
3
3
3
4
6
6
6
7
8
8
9
10
10
11

3 Passband transmission via single AWGN channel

3.1 M-ASK
3.2 QASK
3.3 M-FSK
3.4 M-PSK

12
13
14
15
15

4 Passband transmission via multipath fading channel

4.1 Time variant multipath fading channel
4.2 Characterization of multipath fading channel
4.3 Channel model for multipath fading channel
4.4 Signal design for multipath fading channel
4.5 RAKE demodulator
4.6 Simulation results

16
17
18
20
22
22
23

Reference

24

II

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

1. Introduction
This report describes a simulator for digital communication systems. Three kinds of
digital communication systems: baseband transmission via additive white Gaussian noise
(AWGN) channel, passband transmission via single AWGN channel, and passband
transmission via multipath fading channel, are simulated, as shown in Figure 1. Passband
transmission via multipath fading channel is emphasized.
Multipath fading is a common phenomenon in wireless signal transmission. When a
signal is transmitted over a radio channel, it is subject to reflection, refraction and
diffraction. Especially in the urban and suburban areas where cellular phones are most
often used, the communication environment changes quickly and thus introduces more
complexities and uncertainties to the channel response. This simulator offers a better
understanding of this phenomenon.
In order to observe the effects of multipath fading channel on the transmitted signal, a
whole digital communication system simulator was developed. For simplicity, complex
techniques, such as multiplexing and spread spectrum communication, were not
simulated. Analyses of the performances of the systems are not given. Section 2 briefly
describes baseband transmission via AWGN channel. Section 3 briefly presents passband
transmission via single AWGN channel. Passband transmission via multipath fading
channel is described in detail in Section 4.

2. Baseband transmission via AWGN channel

Baseband transmission is usually used in wired communication system, such as wired
telephone and wired local area network (LAN). Figure 2 shows the structure of a digital

Fig 1 Contents of the simulator

1

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Fig 2 Baseband transmission via AWGN channel

communication system with baseband transmission, which consists of signal source,
sampler, waveform encoder, error control encoder, code transformer, baseband filter,
AWGN channel, optimum receiver, inverse code transformer, error control decoder, and
lowpass filter. The following subsections describe the above blocks. Users can display
the required figures at different steps in Figure 2 by clicking the small rectangles, whose
color stands for a display format specified in the legend at the bottom of Figure 2.
2.1 Signal source
A signal source is used to transform information such as voice into electric signal. In
this simulator, a sine waveform is generated with the amplitude, frequency, and initial
phase specified by the user. Figure 3 shows the pre-filtered signal to be transmitted with
amplitude 3V, frequency 1000Hz, and initial phase 45. After some processing, this
signal will be transmitted through an AWGN channel, and be recovered by the receiver.
O rig in al s ign al

S a m p le d s ig na l

-1

-1

-2

-2

-3

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

Fig 3 Original signal

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

-3

1. 8

2
x 10

-3

Fig 4 Sampled signal

2

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

2.2 Sampling
Figure 3 is an analog signal. To be transmitted via a digital communication system,
the signal must be discretized both in time and amplitude. A sampler is to discretize the
analog signal in the time domain. According to the Nyquist sampling theorem, in order to
restore the signal at the receiving site, the sampling frequency should be at least twice of
that of the maximum signal frequency. Figure 4 shows a sampled signal (discrete signal)
with the sampling frequency 50 times of the signal frequency.
2.3 Waveform encoding and decoding
Waveform encoding converts a source signal into a digital code using a quantization
method. The waveform coded signal is represented by a set of integers {1, 2, , N},
where N is finite. Waveform decoding recovers the original information signal sequence
using the waveform-coded signal. This simulator supports four waveform-coding
schemes.
2.3.1 Pulse Code Modulation (PCM)
PCM is the simplest and oldest waveform-coding scheme. PCM is a process that
assigns a signal value to inputs that are within a specified range. Inputs that fall in a
different range of values are assigned a different signal value. The input signal is in effect
digitized by scalar quantization, as show in Figure 5. If the quantizer performs uniform
quantization, the PCM is called uniform PCM, as Figure 5.
As long as the statistics of the input signal are close to the uniform distribution,
uniform PCM works fine. However, in coding of certain signals such as speech, the input
distribution is far from being uniformly distributed. For a speech waveform, in particular,
there exists a higher probability for smaller amplitudes and lower probability for large
amplitudes. Therefore, it makes sense to design a quantizer with more quantization
regions at lower amplitudes and less quantization regions at larger amplitudes. The
resulting quantizer will be a non-uniform quantizer having quantization regions of
various sizes.
S ig nal a ft er wa veform dec o ding
3

-1

-2

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

-3

Fig 5 PCM with uniform quantization

3

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

The usual method for performing non-uniform quantization is to first pass the
samples through a nonlinear element that compresses the large amplitudes (reduces
dynamic range of the signal) and then perform a uniform quantization on the output. At
the receiving end, the inverse (expansion) of this nonlinear operation is applied to obtain
the sampled value. This technique is called companding (compressing-expanding). There
are two types of companders that are widely used for speech coding: A-law compander
and -law compander. Figures 6 and 7 are the characteristics of A-law compander and law compander, respectively. Figures 8 and 9 are the recovered signal using A-law
compander and -law compander, respectively. Except for the discussion of the
waveform schemes in this subsection, the following results are based on PCM with
uniform quantization.
2.3.2 Simple DPCM
When a band-limited random process is sampled at the Nyquist rate or faster, the
sampled values are usually correlated random variables. This means that the previous
samples give some information about the next sample, and this information can be
employed to improve the performance of the PCM system. In the simplest form of

Fig 6 A-law compander characteristics

Fig 7 -law compander characteristics

S ig nal a ft er wa veform dec o ding

S ig nal a ft er wa veform dec o ding

-1

-1

-2

-2

-3

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

-3

Fig 8 PCM with A-law companding

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

-3

Fig 9 PCM with A-law companding

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

differential pulse-code modulation (DPCM), the difference between two adjacent

samples is quantized. Because two adjacent samples are highly correlated, their
difference has small variations and, therefore, to achieve a certain level of performance
fewer bits are required to quantize it.
Figure 10 shows a block diagram of this simple DPCM scheme. As seen in this
figure, the input to the quantizer is not simply X n X n 1 but rather X n Yn 1 . Yn 1 is
closely related to Xn-1, and the above choice has the advantage that accumulation of
quantization noise is prevented. Figure 11 shows the recovered signal using simple 4-bit
DPCM system.

(b) Decoder

(a) Encoder

Fig 10 A simple DPCM scheme

S ig nal a ft er wa veform dec o ding

4

-1

-2

-3

-4
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

-3

Fig 11 Simple DPCM, 4 bits

(a) Encoder

(b) Decoder

Fig 12 A general DPCM scheme

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

2.3.3 General DPCM

In a more complex version of DPCM, instead of using only the previous sample, the
last p samples are used to predict the value of the next sample. Then the difference
between the sample Xn and its predicted value is quantized. Usually a linear predictor of
the form

p
i =1

ai X n i is employed, and the coefficients of the predictor ai are chosen to

minimize the mean-squared error between the sample Xn and its predicted value.
Figure 12 shows a block diagram of a general DPCM system. Because we are using a
p-step predictor, we are using more information in predicting Xn and, therefore, the range
of variations of Yn will be less. This in turn means that even lower bit rates are possible
here. Differential PCM systems find wide applications in speech and image compression.
Figure 13 shows the recovered signal using second-order and 4-bit DPCM.
2.3.4 Delta Modulation
Delta modulation (M) is a simplified version of the simple DPCM scheme. In M
the quantizer is a 1-bit (2-level) quantizer with magnitudes . Figure 14 shows the
recovered signal using M. From Figure 14, we see that when the input signal changes
dramatically, the quantized signal cant follow the input signal. This phenomenon is
called overloading. Many schemes have been developed to solve this problem.
2.4 Error control coding
Error-control coding techniques are used to detect and/or correct errors that occur in the
message transmission in a digital communication system. The transmitting side of the
error-control coding adds redundant bits or symbols to the original information signal
sequence. The receiving side of the error-control coding uses these redundant bits of
symbols to detect and/or correct the errors that occurred during transmission. The
S ig nal a ft er wa veform dec o ding

S ig nal a ft er wa veform dec o ding

3
2
2
1
1

-1
-1
-2
-2
-3

-4

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

Fig 13 General DPCM, p=2, 4 bits

0. 2

0. 4

0. 6

0. 8

1. 2

-3

1. 4

1. 6

1. 8

2
x 10

-3

Fig 14 M

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

transmission coding process is known as encoding, and the receiving coding process is
known as decoding.
2.4.1 Block codes
There are two major classes in error-control code: block and convolutional. In block
coding, successive blocks of K information (message) symbols are formed. The coding
algorithm then transforms each block into a codeword consisting of N symbols where
N>K. This structure is called an (N, K) code. The ratio K/N is called the code rate. A key
point is that each codeword is formed independently from other codewords. This
simulator supports three block-coding schemes: Hamming code, cyclic code, and BCH
code.
Figure 15 shows the corrupted signal via an AWGN with variance 0.02 without errorcontrol coding. Figures 16 through 18 are the signals via the same channel in Figure 15
using Hamming code, cyclic code, and BCH code, respectively, with N=7, K=4. In Figure
15, the bit error rate is 0.015. The bit error rates in Figures 16 through 18 are 0s.
S ig nal a ft er wa veform dec o ding

S ig nal a ft er wa veform dec o ding

-1

-1

-2

-2

-3

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

-3

2
x 10

Fig 15 Without error-control coding

-3

Fig 16 Hamming code, N=7, K=4

S ig nal a ft er wa veform dec o ding

S ig nal a ft er wa veform dec o ding

-1

-1

-2

-2

-3

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

Fig 17 Cyclic code, N=7, K=4

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

-3

2
x 10

-3

Fig 18 BCH code, N=7, K=4

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

2.4.2 Convolutional codes

Convolutional codes differ from block codes in that there are no independent
codewords. The encoding process can be envisioned as a sliding window, M block wide,
which moves over the sequence of information symbols in steps of K symbols. M is
called the constraint length of the code. With each step of the sliding window, the
encoding process generates N symbols based on the MK symbols visible in the window.
A convolutional code so constructed is called an (N, K, M) code. Convolutional codes are
commonly used in applications that require relatively good performance with low
implementation cost.
The Viterbi method is used for decoding the convolutional codes. The Viterbi
algorithm is a maximum likelihood (ML) decoding procedure that takes advantage of the
fact that a convolutional encoder is a finite state machine. The criterion used for decision
-making is the metric for soft decision decoding and the Hamming distance for hard
decision coding. Details on the Viterbi method can be found in most advanced
communication books.
Figure 19 shows the recovered signal via the same channel as Figure 15 with
convolutional coding. The bit error rate is 0.
2.5 Line coding
Binary 1s and 0s, such as in PCM signaling, may be represented in various serial-bit
signaling formats called line codes. The output of the error-control encoder in this
simulator is unipolar (1s and 0s). Since the direct current component of unipolar signal is
not zero, it is not proper to transmit unipolar signals. Instead, bipolar signaling (1s and
1s) schemes are employed. This simulator provides four bipolar signaling schemes:
bipolar non-return-to-zero signaling (NRZ2), bipolar non-return-to-zero inverse
signaling (NRZI2), bipolar return-to-zero signaling (RZ2), and Miller coding, as shown
in Figure 20. Details of the above signaling schemes can be found in most
communication books. The figures in the following section use Miller code.
S ig nal a ft er wa veform dec o ding
3

-1

-2

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

-3

Fig 19 Convolutional coding, (3,2,2) code

Digital Communication via Multipath Fading Channel

0

Zhiwei Zeng November 2000

0

2
6

Fig 20 Line codes

(a) Original (b) NRZ2 (c) NRZI2 (d) RZ2 (e) Miller
2.6 Baseband filtering
Rectangular waveforms in Figure 20 have high frequency components. If the channel
bandwidth is limited, rectangular waveforms are not appropriate for this kind of channels.
In this simulator, a raise cosine filter is employed to make the waveform smooth. Figures
S ig nal a ft er bas eban d filtering

S ig nal a ft er bas eban d filtering

1. 5

1. 5

0. 5

0. 5

-0 .5

-0 .5

-1

-1

-1 .5

-1 .5

-2

-2
3. 1

3. 2

3. 3

3. 4

3. 5

3. 6

3. 7
x 10

Fig 21 Raise cosine filtered signal

for RZ2 signaling

3. 1
-4

3. 2

3. 3

3. 4

3. 5

3. 6

3. 7
x 10

-4

Fig 22 Raise cosine filtered signal

for Miller coding
9

Digital Communication via Multipath Fading Channel

E y e p att ern of the s ign al aft er bas e band filterin g

Zhiwei Zeng November 2000

E y e p att ern of the s ign al c o rru pte d b y t he A W G N

1
1

0. 8
0. 6

0. 5

0. 4

am plitu de

am plitu de

0. 2
0
-0 .2

-0 .5

-0 .4
-0 .6

-1
-0 .8
-1
1

1. 5

2. 5

tim e (s e c ond )

3. 5
x 10

-6

Fig 23 Eye pattern of the

baseband filtered signal

1. 5

2. 5

tim e (s e c ond )

3. 5
x 10

-6

Fig 24 Eye pattern of signal

corrupted by the AWGN channel

21 and 22 show the raise cosine filtered signals form RZ2 signaling and Miller codes,
respectively.
2.7 Eye pattern
An eye-pattern plot is a simple and convenient tool to study the effects of intersymbol interference (ISI) and other channel impairments for digital transmission. The
received signal is plotted against time. When the x-axis time limit is reached, the signal
goes back to the beginning of the time point. Thus the plots overlay each other. In the
optimal condition, the decision point is at the widest eye opening point. Figures 23 and
24 are the eye patterns of the output of the baseband filtering and the signal corrupted by
the AWGN channel, whose variance is 0.01.

2.8 Optimum receiver in AWGN channel

It is convenient to subdivide the receiver into two parts, the signal demodulator and
the detector. The function of the signal demodulator is to convert the received waveform
r(t) into an N-dimensional vector r=(r1, r2, , rN), where N is the dimension of the
transmitted signal waveforms. The function of the detector is to decide which of the M
possible signal waveforms was transmitted based on observation of the rector r.
The signal and the noise are expanded into a series of linearly weighted orthonormal
basis functions {n(t)}. Two realizations of the signal demodulator are simulated. One is
based on the use of signal correlators, named correlation-type demodulator, as shown in
Figure 25. The second is based on the use of matched filters (MF), named matched-filtertype demodulator, as shown in Figure 26. In this simulator, the signal is one-dimensional,
and the realizations of above two demodulators are the same.
The maximum a posteriori probability (MAP) criterion is used to maximize the
probability of a correct decision and, hence, minimizes the probability of error. This
10

Digital Communication via Multipath Fading Channel

Fig 25 Correlation-type demodulator

Zhiwei Zeng November 2000

Fig 26 Matched-filter-type demodulator

S ig nal c orrupt ed by t he A W G N

O ut put of the opt im u m rec eiver

1. 5

1. 5

0. 5

0. 5

-0 .5

-0 .5

-1

-1

-1 .5

-1 .5

-2

-2
3. 1

3. 2

3. 3

3. 4

3. 5

3. 6

3. 7
x 10

62

64

66

68

70

72

74

-4

Fig 27 Corrupted signal of Figure 22

Fig 28 Output of the optimum

receiver for Figure 25

detector is optimum when the signal has no memory. When the transmitted signal has
memory, the optimum detector is a maximum-likelihood (ML) sequence detector, which
bases its decisions on observation of a sequence of received signals over successive
signal intervals. Although this simulator does not provide ML sequence detector, it has
been shown that a detector based on the MAP criterion and one that is based on the ML
criterion make the same decisions as long as the a priori probabilities are all equal. Figure
27 is the signal, corresponding to that in Figure 22, corrupted by the AWGN channel.
Figure 28 is the decisions made by the optimum receiver.
2.9 Lowpass filtering
In order to recover the analog signal from the output of the waveform decoder, a
lowpass filter is needed. This simulator provides a butterworth filter with the order and
digital cutoff frequency specified by the user. Figure 29 shows the lowpass filtered signal,

11

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Re s to red s ign al
4
O rig in al s ign al
Re s to red s ign al
3

-1

-2

-3
0

0. 2

0. 4

0. 6

0. 8

1. 2

1. 4

1. 6

1. 8

2
x 10

-3

Fig 29 Lowpass filtered signal compared with the original signal

with order of 4 and cutoff frequency of 0.1, compared with the original analog signal. It
is obvious that there is a delay due to the infinite impulse response (IIR) filter.
3. Passband transmission via single AWGN channel
Wireless communication systems must use passband transmission. Figure 30 shows
the structure of a digital communication system with passband transmission via single
AWGN channel, which consists of signal source, sampler, waveform encoder, error
control encoder, digital modulation (digital mapping and analog modulation), AWGN
channel, digital demodulation (analog demodulation and digital demapping), error control
decoder, and lowpass filter. The following subsections describe the above blocks that did
not show up in Section 2, that is, digital modulation and demodulation. Users can display
the required figures at different steps in Figure 30 by clicking the small rectangles, whose
color stands for a display format specified in the legend at the bottom of Figure 30.

Fig 30 Passband transmission via single AWGN channel

12

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

This section discusses passband digital modulation and demodulation methods.

Digital modulation includes two parts: digital to analog mapping and analog modulation,
as shown in Figure 31. Digital to analog mapping techniques convert the received digital
data into analog signals. Analog modulation then modulates the mapped signals.
Typically, digital signals form a finite set of symbols. For example, in binary
transmission it is possible to consider pairs of binary numbers as symbols. In this case,
the symbol set is {00, 01, 10, 11}, making four distinct symbols. A digital
mapping algorithm must have at least four mapping points to map the symbols uniquely.
The number of points in the signal set is called the M-ary number, also known as
alphabet size. The arrangement of the signal set in the signal space, the space that digital
mapping algorithm uses to map the mapping points, is called the constellation.
In general, digital demodulation is the inverse of the modulation process. A
modulated signal is demodulated and then demapped from an analog signal to a digital
signal. Figure 32 shows how digital demodulation divides into an analog demodulation
and an analog to digital mapping. This demodulation structure applies to all
demodulation methods except for M-FSK and P-FSK, which use correlation techniques.
3.1 M-ASK
M-ary amplitude shift keying (ASK) modulation includes two parts: M-ASK mapping
followed by analog amplitude modulation (AM). M-ASK mapping is a one-dimensional
coding process that maps the input digital symbols into real numbers in the interval [-x,
x], where x is the specified maximum number. The input symbols are integers in the
range [0, M-1], where M is the M-ary number.
Three AM methods are supported in this simulator: double-sideband with
transmission carrier amplitude modulation (AM-DSB-TC), double-sideband carrier

Fig 31 Digital modulation

Fig 32 Digital demodulation

13

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

suppressed amplitude modulation (AM-DSB-SC), and single-sideband amplitude

modulation (AM-SSB). Details on the AM methods can be found in references.
Figure 33 shows the constellation of 4-ASK. Figures 34 through 36 are the spectrums
of the modulated signals using AM-DSB-TC, AM-DSB-SC, and AM-SSB, respectively.
3.2 QASK
The M-ary quadrature amplitude shift keying M-QASK method is the most
commonly used digital modulation method in communication systems. In general, an MQASK process takes the input digital symbol and maps it into two independent
components: in-phase and quadrature. An analog quadrature amplitude modulation
(QAM) is used to modulate the in-phase and quadrature signals. On the receiving side,
the signal is demodulated into in-phase and quadrature signals. A demapping process
recovers the message signal by using the in-phase and quadrature signals.
There are many mapping schemes available to map the input symbol to the in-phase
and quadrature components. This simulator supports square constellation, as shown in
Figure 37.
A S K c on s te llat ion

S p ec t rum o f th e a nalog m o dulat ed s ignal

10 000

1
90 00
0. 8
80 00

0. 6

70 00

0. 4
0. 2

60 00

50 00

-0 .2

40 00

-0 .4

30 00

-0 .6

20 00

-0 .8
10 00
-1
-1

-0 .8

-0 .6

-0 .4

-0 .2
0
0. 2
In-phas e c o m p onen t

0. 4

0. 6

0. 8

0
1. 3095

S p ec t rum o f th e a nalog m o dulat ed s ignal

90 00

90 00

80 00

80 00

70 00

70 00

60 00

60 00

50 00

50 00

40 00

40 00

30 00

30 00

20 00

20 00

10 00

10 00

1. 311

1. 311 5

1. 312

1. 311 5

S p ec t rum o f th e a nalog m o dulat ed s ignal

10 000

1. 310 5

1. 311

Fig 34 Spectrum using AM-DSB-TC

10 000

1. 31

1. 310 5

x 10

Fig 33 4-ASK constellation

0
1. 3095

1. 31

1. 312
x 10

Fig 35 Spectrum using AM-DSB-SC

0
1. 3095

1. 31

1. 310 5

1. 311

1. 311 5

1. 312
x 10

Fig 36 Spectrum using AM-SSB

14

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Q A S K Co ns t alla tion

F S K c o ns t ellation

1. 5
2
1

1. 5

0. 5

0
1

-0 .5
0. 5
-1

-1 .5
-1 .5

-1

-0 .5

0. 5

1. 5

0
-1

B a s eba nd s pec trum for F S K (H z )

Fig 37 4-ary QASK square constellation

Fig 38 4-ary FSK constellation

3.3 M-FSK
M-ary frequency shift keying (FSK) modulation modulates a digital signal by
changing the frequency of the output signal depending on the value of the input signal.
The M-FSK modulation divides into two parts: mapping and analog modulation. The
mapping process maps the input symbol into the value of the frequency shift from the
carrier frequency, and the analog modulation is analog frequency modulation (FM). If the
carrier frequency is Fc, and the tone space f (the frequency separation between two
consecutive frequencies in the modulated signal), then the frequency range of a
modulated signal is in the range [Fc, Fc+(M-1) f]. Figure 38 shows the constellation of
4-FSK.
The demodulation process of the M-FSK uses a length-M vector signal where the
frequency of the ith element in the vector signal matches the modulated signal when the
input symbol is i. The demodulation process computes the correlation values between the
signal array and the received signal and, after calculating the maximum correlation value,
decides what symbol was most likely transmitted. There are two different methods to
compute the correlation value: the coherent and noncoherent method, as shown in Figures
39 and 40, respectively. Using the coherent method, you must know the phase
information of the modulated signal from the receiving side where phase-locked loops are
used. The noncoherent method does not require phase information; it recovers the phase
of the modulated signal during the demodulation. However, the noncoherent
demodulation method is more computationally complex than the coherent demodulation
method.
3.4 M-PSK
The M-ary phase shift keying (PSK) modulation modulates a signal by changing the
phase values in the modulated output signal. The M-PSK modulation divides into two
parts: mapping and analog phase modulation (PM). M-PSK distinguishes between the

15

Digital Communication via Multipath Fading Channel

Fig 39 Coherent demodulation of M-FSK

Zhiwei Zeng November 2000

Fig 40 Noncoherent demodulation of M-FSK

digital messages by setting different initial phase shifts in the modulation. The digital
input signals to a M-PSK modulator are in the range [0, M-1]. The phase shift for input
digit i is 2i/M. The structure of the demodulator for M-PSK is similar to that of M-FSK.
The demodulation process calculates the correlation value between the input signal and a
vector of carrier frequency sinusoidal signal. Each sinusoidal signal in the vector has its
phase set to a possible result from the signal set. Figure 41 shows the constellation of 4PSK. Figure 42 is part of the modulated signal using 4-PSK. We can see phase change in
Figure 42.

4. Passband transmission via multipath fading channel

This section discusses the most important contexts of this report. Figure 43 shows the
A n alog m o dulat ed s ignal
1. 5

0. 5

-0 .5

-1

-1 .5
0

0. 2

0. 4

0. 6

0. 8

1. 2

A S K /P S K C ons tellation

x 10

Fig 41 Constellation of 4-PSK

-5

Fig 42 Modulated signal using 4-PSK

16

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Fig 43 Digital transmission via multipath fading channel

structure of a digital communication system with passband transmission via multipath
fading channel, which consists of signal source, sampler, waveform encoder, error
control encoder, digital modulation (digital mapping and analog modulation), AWGN
channel, filter bank demodulator, error control decoder, and lowpass filter. The following
subsections discuss the multipath fading channels. Users can display the required figures
at different steps in Figure 43 by clicking the small rectangles, whose color stands for a
display format specified in the legend at the bottom of Figure 43.
4.1 Time variant multipath fading channel
Fundamentally, mobile radio communication channels are time varying, multipath
fading channels. In a radio communication system, there are many paths for a signal to
travel from a transmitter to a receiver. Sometimes there is a direct path where the signal
travels without being obstructed. In most cases, components of the signal are reflected by
the ground and objects between the transmitter and the receiver such as buildings,

Fig 44 Multipath fading channel

17

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

vehicles, and hills or refracted by different atmospheric layers. Theses components travel
in different paths and merge at the receiver. Figure 44 illustrates this phenomenon. Each
path has a different physical length. Thus, signals on each path suffer different
transmission delays due to the finite propagation velocity. The superposition of these
signals at the receiver results in destructive of constructive interference, depending on the
relative delays involved. The fact that the environment changes as time passes leads to
signal variation. This is called time variant. Signals are also influenced by the motion of a
terminal. A short distance movement can cause an apparent change in the propagation
paths and in turn the strength of the received signals.

4.2 Characterization of multipath fading channel

Both the propagation delays and the attenuation factors are time-variant as a result of
changes in the structure of the medium. The received bandpass signal may be expressed
in the form
x(t ) = n (t )s (t n (t ))

(1)

where s(t) is the transmitted signal, n (t ) is the attenuation factor for the signal received

on the nth path and n (t ) is the propagation delay for the nth path. s(t) can be expressed as

s (t ) = Re s l (t )e j 2f c t

(2)

where sl(t) is the equivalent lowpass transmitted signal. Substitute (2) into (1) yields the
result

x(t ) = Re n (t )e j 2f c n (t )s l (t n (t )) e j 2f ct

(3)

It is apparent from (3) that the equivalent lowpass received signal is

rl (t ) = n (t )e j 2f c n (t )s l (t n (t ))

(4)

It follows that the equivalent lowpass channel is described by the time-variant impulse
response
c( ; t ) = n (t )e j 2f c n (t ) ( n (t ))
(5)
n

c( ; t ) represents the response of the channel at time t due to an impulse applied at time t.
When there are a large number of signal propagation paths, the central limit theorem
can be applied. Thus c( ; t ) can be modeled as a complex-valued Gaussian random
process. The envelope c( ; t ) at any instant t is Rayleigh-distributed, as shown in Figure
45. In this case the channel is said to be a Rayleigh fading channel. In the event that there
are fixed scatterers or signal reflectors in the medium, in addition to randomly moving

18

Digital Communication via Multipath Fading Channel

6000

Zhiwei Zeng November 2000

5000
4500

5000
4000
3500
4000
3000
3000

2500
2000

2000
1500
1000
1000
500
0

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Fig 45 Rayleigh distribution

0.2

0.4

0.6

0.8

1.2

1.4

Fig 46 Ricean distribution

scatterers, c( ; t ) can no longer be modeled as having zero mean. In this case, the
envelope c( ; t ) has a Rice distribution, as shown in Figure 46, and the channel is said to
be a Ricean fading channel.
Assuming that c(;t) is wide-sense-stationary (WSS), we can define the
autocorrelation function of c(;t) as
1
c ( 1 , 2 ; t ) = E c * ( 1 ; t )c ( 2 ; t + t )
(6)
2
where * means conjugation. In most radio transmission media, the attenuation and phase
shift of the channel associated with path delay 1 is uncorrelated with the attenuation and
phase shift associated with path delay 2. This is usually called uncorrelated scattering.
Under this assumption, we have

c ( 1 , 2 ; t ) = c ( 1 ; t ) ( 1 2 )

(7)

c ( ) = c ( ;0 ) is simply the average power output of the channel as a function of the

time delay . For this reason, c() is called the multipath intensity profile or the delay
power spectrum of the channel. The range of values of over which c() is essentially
nonzero is called the multipath spread or delay spread of the channel and is denoted by
Tm.
The time-variant transfer function C(f;t) can be defined as the Fourier transform of
c(;t). That is,
C ( f ; t ) = c( ; t )e j 2f d

(8)

If c(;t) is modeled as a complex-valued zero-mean Gaussian random process in the t

variable, it follows that C(f;t) also has the same statistics. Under the assumption that the
channel is WSS, the autocorrelation function of C(f;t) is the Fourier transform of the
multipath intensity profile, i.e.,

C (f ; t ) = C ( f 1 , f 2 ; t ) =

{ c ( ; t )}

(9)

19

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

where f=f2-f1. C(f;t) is called the spaced-frequency, spaced-time correlation function

of the channel. Let t=0, then

C (f ) =

{ c ( )}

(10)

As a result of the Fourier transform relationship between C(f) and c(), the reciprocal
of the multipath spread is a measure of the coherent bandwidth (f)c of the channel. That
is,

(f )c

1
Tm

(11)

If (f)c is small in comparison to the bandwidth of the transmitted signal, the channel is
said to be frequency-selective. In this case, the signal is severely distorted by the channel.
On the other hand, if (f)c is large in comparison with the bandwidth of the transmitted
signal, the channel is said to be frequency-nonselective.
Define the Fourier transform of C(f;t) with respective to the variable t to be the
function SC(f;). With f set to zero, the relation becomes
S C ( ) = C (0; t )e j 2t dt

(12)

The function SC() is a power spectrum that gives the signal intensity as a function of the
Doppler frequency . Hence, we call SC() the Doppler spectrum of the channel. The
range of values of over which SC() is essentially nonzero is called the Doppler spread
Bd of the channel. Due to the Fourier transform relationship between SC() and C(t), the
reciprocal of Bd is a measure of the coherence time (t)c of the channel. That is,

(t )c

1
Bd

(13)

If the coherence time is larger than the symbol period, the channel is said to be a slowfading channel. On the other hand, if the coherence time is smaller than the symbol
period, the channel is a fast fading channel.
The scattering function of the channel is defined as
S ( ; ) =

C (f ; t )e j 2t e j 2f dtdf

(14)

The relationships among the functions C(f;t), c(;t), C(f;), and s(;) are
summarized in Figure 47.
4.3 Channel model for multipath fading channel
We may view the channel response in (5) as the sum of a number of vectors
(phasors), each of which has a time-variant amplitude n(t) and phase n(t). In general, it
takes large dynamic changes in the physical medium to cause a large change in {n(t)}.
On the other hand, the phase { n(t)} change by 2 or more radians with relatively small

20

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Fig 47 Relationships among the channel correlation functions and power spectra
changes of the medium characteristics. {n(t)} is expected to be in an unpredictable
(random) manner. Thus c(;t) can be modeled as a complex-valued Gaussian random
process using the central limit theorem.
This simulator simulates a slow-fading frequency-selective channel. A general model
for a time-variant slow-fading frequency-selective channel is illustrated in Figure 48. The
channel model consists of a tapped delay line with uniformly spaced taps. The tap
spacing between adjacent taps is 1/W, where W is the bandwidth of the signal transmitted
through the channel. The tap coefficients, denoted as c n (t ) = n (t )e j n (t ) , are usually
modeled as complex-valued Gaussian random processes that are mutually uncorrelated.
The length of the delay line corresponds to the multipath spread. That is,

Tm =

L
W

(15)

where L represents the maximum number of possible multipath signal components. Using
the simulator, the user can specify the means and variances of the possible channels or
use the default values.
21

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Fig 48 Tapped delay line model of frequency-selective channel

4.4 Signal design for multipath fading channel
There are lots of carrier modulation methods that are available for digital and analog
modulation, as discussed in Section 3. However, signal fading generally causes large
fluctuations in the received signal amplitude, and consequently it is extremely difficult to
discriminate among multiple amplitude levels in the received signal. For this reason,
ASK and QASK are generally avoided for digital communication through fading
channels. In this simulator, users can choose 2-FSK (coherent), 2-FSK/noncoherent, or 2PSK as the digital modulation method. ASK, QASK, and the corresponding analog
modulations are disabled in Figure 43. For simplicity, spread spectrum and ISI are not
discussed in this report.
4.5 RAKE demodulator
Observing Figure 48, it is apparent that the tapped delay line model with statistically
independent tap weights provides us with L replicas of the same transmitted signal at the
receiver. Hence, a receiver that processes the received signal in an optimum manner will
achieve the performance of an equivalent Lth-order diversity communications system.
This simulator employs binary signaling over the multipath-fading channel. We have
two equal-energy signals sl1(t) and sl2(t). The received signal is expressed as
L

rl (t ) = c k (t )s li (t k W ) + z (t )
k =1

= v i (t ) + z (t ),

0 t T,

i = 1, 2

(16)

where z(t) is a complex-valued zero-mean AWGN process. Assume that the channel
tap weights are known. Then the optimum receiver consists of two filters matched to v1(t)

22

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Fig 49 Optimum demodulator for wideband binary signals

and v2(t), followed by samplers and a decision circuit that selects the signal
corresponding to the largest output. An equivalent optimum receiver employs cross
correlation instead of matched filtering. In either case, the decision variables for coherent
detection of the binary signals can be expressed as
T
U m = Re rl (t )v m* (t )dt ,
0

m = 1, 2

(17)

Figure 49 illustrates the operations involved in the computation of the decision

variables. In this realization of the optimum receiver, the two reference signals are
delayed and correlated with the received signal rl(t).
4.6 Simulation results
Figure 50 shows part of the transmitted signal and the corresponding received signal
through the multipath fading channel (without AWGN), whose means and variances of
the tap coefficients are default values shown in Figure 43. Figure 51 shows part of the
original binary data and the output of the RAKE demodulator. The bit error rate of this
example is 0.

23

Digital Communication via Multipath Fading Channel

Zhiwei Zeng November 2000

Input and O utput of the m ultipath fading c hannel

O riginal digital signal

1.5

2
Input of t he m ultipath c hannel
O ut put of the m ult ipath c hannel
1

0
0.5
-1
55

60

65

70

75

80

85

90

95

100

90

95

100

0
O ut put of the R A K E dem odulater
2
-0.5
1

-1

-1.5

-1
0

0.5

1.5

2.5

3.5

4.5

5
x 10

Fig 50 Input and output of the

multipath-fading channel

55

60

65

70

75

80

85

-6

Fig 51 Original binary data and the

output of the RAKE demodulator

Reference:
1. J. G. Proakis, Digital Communications, McGraw Hill, 1995;
2. L. W. Couch, Digital and Analog Communication Systems, Prentice Hall,
1997;
3. Z. Y. Shen, et al, Principals of Communication Systems, Xidian Univ. Press,
1997;
4. J. H. Schiller, Mobile Communications, Addison-wesley, 2000;
5. X. M. Kong, A Simulator for Time-varying Multipath Fading Channels, Master
thesis, 2000.

24