41st IEEE CDC

Las Vegas, Nevada

December 9th 2002
Workshop M-5: Wireless Communication Channels:
Modeling, Analysis, Simulations and
Applications
Organizers:

Charalambos D. Charalambous
Nickie Menemenlis

Wireless Communication Channels

Schedule
08:00-08:45 Introduction to Wireless Communication
Channels (C.D. Charalambous)
8:45-9:15 Statistical Analysis of Wireless Fading
Channels (C.D. Charalambous)
9:15-9:25 Break

9:25 -10:10 Stochastic Differential Equations in Modeling

Log-Normal Shadowing (N. Menemenlis)
10:10-10:55 Stochastic Differential Equations in
Modeling Short-Term Fading (N. Menemenlis)
10:55-11:00 Break

11:00-12:00 Applications (C.D. Charalambous)

Additional information can be found at:
[Link]

Introduction to Wireless Communication Channels

Shannons communication channel
Impulse response of wireless fading channels
Large-scale and small scale propagation models
Log-Normal shadowing channel
Short-term fading channel
Autocorrelation functions and power spectral densities
Assumption: WSSUS
Time spreading
Time variations
Channel classification
Channel simulations

Chapter 1: Shannons Wireless Communication System

Channel
code
word

Message
Signal

Source

Source
Encoder

Channel
Encoder

Modulator

Modulated
Transmitted
Signal

Wireless
Channel

User

Source
Decoder
Estimate of
Message

Channel
Decoder
Estimate of

Demodulator

Received
Signal

Chapter 1: Large and Small Scale Propagation Models

Area 1

Area 2

Short-term
fading

Log-normal
shadowing
Transmitter

Chapter 1: Impulse Response Characterization

Tim

ty
r
e
p
ro
p
s
n
io
t
ia
r
va

t2

(t2)

t1

t0

Time spreading property

(t1)

(t0)

Impulse response: Time-spreading : multipath

and time-variations: time-varying environment

Chapter 1: Multipath Fading Components

Complex low-pass representation of impulse response

Response at of the channel at time t due an impulse

applied at time t - .
N (t )

Cl (t ; ) ri (t ; )e ji (t ; ) (t i (t ))
i 1

ri (t ; ): Signal attenuation (R.P.)

i (t ; ) i ( c di ) di t: Phase angle (R.P.)

i (t ) :
N (t ):

Propagation delay (R.P.)

Number of waves impinging on the receiver
antenna at time t (a counting R.P.)

Chapter 1: Band-pass Representation of Impulse Response

Band-pass representation of impulse response:

C (t ; ) Re

N (t )

r (t; )e
i 1

j i ( t ; )

jc t

(t i (t ))

N (t )

I i (t ; ) cos ct Qi (t ; ) sin ct (t i (t ))
i 1

I i (t ; ) ri (t ; ) cos( i (t ; ))
Qi (t ; ) ri (t ; ) sin( i (t ; ))
ri (t ; ) I i2 (t ; ) Qi2 (t ; )
Qi (t ; )

i (t ; ) tan

I
(
t
;

)
i

In-phase component
Quadrature component
Attenuation
Phase

Chapter 1: Representation of Additive Noise Channel

Low-pass and band-pass representation of received signal:
N (t )

yl (t ) ri (t ; i (t ))e
i 1

y (t ) Re{ yl (t )e

j c t

j i ( t ; i ( t ))

sl (t i (t )) nl (t )

N (t )

I i (t ; i (t )) cos ct Qi (t ; i (t )) sin ct sl (t i (t )) n(t )

i 1

sl (t ) :

arbitrary low-pass transmitted signal

{nl (t )}t 0 : complex valued noise process

Chapter 1: Large and Small Scale Propagation Models

Large scale propagation models:
T-R separation distances are large
Main propagation mechanism: reflections
Attenuation of signal strength due to power loss along distance
traveled: shadowing
Distribution of power loss in dBs: Log-Normal
Log-Normal shadowing model
Fluctuations around a slowly
varying mean

Chapter 1: Large and Small Scale Propagation Models

Small scale propagation:
T-R separation distances are small
Heavily populated, urban areas
Main propagation mechanism: scattering
Multiple copies of transmitted signal arriving at the transmitted
via different paths and at different time-delays, add vectotrially at
the receiver: fading
Distribution of signal attenuation
coefficient: Rayleigh, Ricean.
Short-term fading model
Rapid and severe signal
fluctuations around a slowly
varying mean

Chapter 1: Log-Normal Shadowing Model

Average power of received signal at d

d 0
P(d ) P(d0 ) , d d 0
d
d 0 : Reference distance, i.e. d 0 1 m. for indoors

: Characterizes the environment

Chapter 1: Log-Normal Shadowing Model

Average power path-loss in dBs at distance d

d 0
PL d (d )[dB ] PL d (d 0 )[dB ] 10 log , d d 0
d
Power path-loss in dB's
PL(d )[dB ] PL d (d )[dB ] X%, X%: N (0, 2 )
Signal Attenuation Coefficient
r (d )

P%
(d )
kPL ( d )[ dB ]
e
, log-normal, k ln10
20
Pt

P%
(d ) : received power at d
Pt :

power of transmitted signal

Chapter 1: Log-Normal Shadowing Model

Power path-loss in dBs, x, and
attenuation coefficient, r, vs d

Distributions: x : normal and

r=ekx : log-normal

1
( x x )2 / 2 x2
f ( x)
e
, x E[ x], x2 Var ( x)
2 x

1
x2
2
2
f (r )
exp (ln r x ) / 2 x , E[ r ] exp x , r 0
2
2 x r

Chapter 1: Short-Term Fading Model

3-Dimensional Model [Clarke 68, Aulin 79]
z
z0

y0
y

ve , N
a
w 1,
g
n n=
i
m n};
o
inc , n,
O(x0 ,y0 ,z0)
h
t
n
n r n,
:{
=
En
n
n
x0

v
direction of motion of mobile
on x-y plane

Chapter 1: Short-Term Fading Model

3-D Model [Clarke 68, Aulin 79]
Transmitted signal: Re{ejct}
Total field at mobile, or receiving location, O(x0, y0, z0)
N

E (t ) En (t )
n 1

En (t ) Re rn e jn e jc (t n ) rn cos(c (t n ) n ), n 1, K N
2
x0 cos n cos n y0 sin n cos n z0 sin n

c f c , f c : carrier frequency, c : speed of light

c n

rn , n , n , n n1 :
N

random variables, statistically independent,

E0
rn : signal attenuation coefficient; n : U 0, 2 , E[r ]
N
2
n

Chapter 1: Short-Term Fading Model

3-D Model [Clarke 68, Aulin 79]
Total field at receiving location when mobile moves
O(x0, y0, z0) => (x0+vtcos, y0 +vtsin, z0), v: velocity of mobile

En (t ) Re rn e jn e

j c n ( t n )

rn cos( nt n ) cos c t rn sin( nt n ) sin ct

I n (t ) cos c t Qn (t )sin c t
2 v
n
cos n cos n , n n c n

E (t ) I (t ) cos ct Q(t ) sin ct

N

n 1

n 1

I (t ) rn cos( nt n ), Q(t ) rn sin( nt n )

I (t ) : In-phase component, Q(t ) : Quadrature component

n :

Doppler shift

Chapter 1: Short-Term Fading Model

3-D Model [Clarke 68, Aulin 79]
Statistical characterization of {I(t), Q(t)}
for N large ( 6)
x @{I (t ), Q(t )},

I (t ),

Q(t ) : N ( x , x2 ), i.e. Gaussian

x E[ I (t )] E[Q(t )],

x2 Var I (t ) Var Q(t )

Usually assume I (t ), Q(t ) uncorrelated and therefore independent

No specular component: x 0 :

I (t ),

Q(t ) : N (0, x2 )

With specular component

or a direct path between transmitter and receiver: x 0
E (t ) I (t ) I d (t ) cos ct I (t ) I d (t ) sin ct

I (t ) I

d (t ) ,

I (t ) I d (t ) :

N ( x , x2 )

Chapter 1: Short-Term Fading Model

Statistical characterization of rn
From above r (t ) I 2 (t ) Q 2 (t ),
No specular component: r (t ) : Rayleigh
r
f (r ) 2 exp r 2 / 2 x2 ,
x

E[r ]
x,
2
With specular component: r : Ricean
r
f (r ) 2 exp rs2 r 2 / 2 x2 I 0 rrs / x2 ,
x

rs2 E I (t ) E Q(t ) , E r 2 2 x2 rs2

I 0 : modified Bessel function of the first kind of order 0
2

Chapter 1: Short-Term Fading Model

Autocorrelation functions
Autocorrelation of E (t ) : RE ( )

RE ( ) @E E (t ) E (t ) ( ) cos c c( ) sin c

E0
( ) E I (t ) I (t )
E cos n ,
2
E
c( ) E Q(t )Q(t ) 0 E sin n
2
For n : uniform in (0, 2 ]
E0
2 v

( )
J
cos

0
p( ) d , c( ) 0

J 0 : Bessel function of the first kind of order 0.

Chapter 1: Short-Term Fading Model

Power Spectral Density of E (t ) : S D ( f ) @F RE ( )
For n : uniform in (0, 2 ] and p( )

cos
,
2sin m

0,
E0
,
4 f m sin m

SD ( f )
E0
4 f m sin m

,
2

f fm
f m cos m f f m

1 2 cos m 1 ( f / f m )
sin
,
1 ( f / fm )

f f m cos m

f m : Maximum
Doppler frequency

Chapter 1: Time Delays of Paths

Complex low-pass representation of impulse response:
N (t )

Cl (t ; ) ri (t ; )e ji (t ; ) (t i (t ))
i 1

i (t ) :
N (t ):

Propagation delay (R.P)

Number of waves impinging on the receiver
antenna at time t (a counting R.P.)

Typically the time delays are modeled using exponential

distribution, implying that the number of paths is a Poisson
counting process
In reality this representation is not very accurate.

Chapter 1: Channel Autocorrelation Functions

Linear time-varying filter

N (t )

Cl (t ; ) ri (t ; )e ji (t ; ) (t i (t ))
i 1

General autocorrelation function is defined by:

c t1 , t2 ; 1 , 2 =E Cl (t1 ; 1 )C l (t2 ; 2 )
General expressions for the Autocorrelation function are
introduced by Bello 63 for a widely accepted Wide-Sense
Stationary Uncorrelated Scattering (WSSUS) channel
WSS in the time-domain
US attenuation and phase shift of paths i and j are
uncorrelated

Chapter 1: Channel Autocorrelation Functions

Time-spreading: Multipath characteristics of channel

1. Autocorrelation in the Time-Domain

E Cl (t ; 1 )C l (t t ; 2 ) c t ; 1 2 1

c t ; : Average power output of the channel as a

function of time-delay, , and the difference in the
observation time t.
2. Power-delay profile
c @ c 0; c t ;

t 0

c : Average power output of the channel as a

function of the time-delay, or excess delay, .

Chapter 1: Channel Autocorrelation Functions

Time-spreading: Multipath characteristics of channel

3. Space-frequency, space-time
c t ; f @F c t ; c t ; e j 2f d

Note: US condition WSS in the frequency domain

observation time t.
4. Power-delay spectrum
c f @ c t ; f

t 0

F c

5. Frequency variations of the channel

Cl (t ; f ) @F Cl (t ; ) Cl (t ; )e j 2 f d

Chapter 1: Channel Autocorrelation Functions

Time-spreading: Multipath characteristics of channel
Multi-path delay spread, Tm
Characterizes time dispersiveness of the channel,
Obtained from power delay-profile, c()
Indicates delay during which the power of the received
signal is above a certain value.
Coherence bandwidth, Bc approx. 1/ Tm
Indicates frequencies over which the channel can be
considered flat
Two sinusoids separated by more than Bc: affected
differently by the channel
Indicates frequency selectivity during transmission.

Chapter 1: Channel Autocorrelation Functions

Time variations of channel: Frequency-spreading

1a. Double Fourier transform of Cl (t ; )

Sc ; f @Ft F c t ; Ft c t ; f

c t ; f e j 2t d t

1b. Doppler Power Spectum of channel

Sc @Sc ; f

f 0

c t ; 0 e j 2t d t

Power output of the channel as a function of the

Doppler frequency shift , , (w.r.t. carrier frequency f c ).
No time variations: Sc a delta function

Chapter 1: Channel Autocorrelation Functions

Time variations of channel: Frequency-spreading

2. Scattering function
Sc ; @Ft c t ;
1

@Ff

S ; f
c

c t ; e j 2t d t

Sc ; f e j 2f d f

Represents average power of the channel as a function

of different time-delays, , and the Doppler frequency, .

Power delay profile : c Sc ; d

Doppler power spectrum: Sc Sc ; d

Chapter 1: Channel Autocorrelation Functions

Time variations of channel: Frequency-spreading
Doppler Spread, Bd
Characterizes frequency dispersiveness of the channel,
or the spreading of transmitted frequency due to different
Doppler shifts
Obtained from Doppler spectrum, Sc()
Indicates range of frequencies over which the received
Doppler spectrum is above a certain value
Coherence time, Tc approx. 1/ Bd
Time over which the channel is time-invariant
A large coherence time: Channel changes slowly

Chapter 1: Channel Autocorrelation Functions

c()

Power Delay
Profile

t=0

Power Delay
Spectrum
|c(f)|

|c(t;f)|

t=0
Bc

Tm

c( t;)

F
f
t

Ft

WSSUS
Channel

Sc(; f)

S ( ; ) d

Ft
Sc( ;) Scattering
Function

|c(t)|

S ( ; )d

Sc( )
Tc

Ft

Doppler Power
Spectrum
Bd

f=0

f=0

Sc()

Chapter 1: Channel Classification

Based on Time-Spreading

Flat Fading

Frequency Selective

1. BS < BC Tm < Ts
2. Rayleigh, Ricean distrib.
3. Spectral chara. of transmitted
signal preserved

1. BS > BC Tm > Ts
2. Intersymbol Interference
3. Spectral chara. of transmitted
signal not preserved
4. Multipath components resolved
Channel

Channel

Signal

Signal
BC

BS

freq.

BS

BC

freq.

Chapter 1: Channel Classification

Based on Time-Variations

Fast Fading

Slow Fading

1. High Doppler Spread

2. 1/Bd TC < Ts

1. Low Doppler Spread

2. 1/Bd TC> Ts

Signal

Signal
Doppler
BD

BS

freq.

Doppler
BS

BD

freq.

Chapter 1: Channel Classification

Underspread channel: TmBd << 1

Channel characteristics vary slowly (Bd small) or paths
obtained within a short interval of time (Tm small).
Easy to extract channel parameters.

Overspread channel: TmBd >> 1

Hard to extract parameters as channel characteristics
vary fast and channel changes before all paths can be
obtained.

Chapter 1: Flat Fading Channel Simulations

Flat Fading

yl (t ) Cl t ; sl (t )d Cl t ; f Sl ( f )e j 2 ft df

Cl t ;0 Sl ( f )e j 2 ft df C t ;0 sl (t )

(t )e j (t ) sl (t );

Sl ( f ) @F sl (t )

Cl t ; f @F Cl t ;
(t): Rayleigh or Ricean

Gaussian
noise
source

Shaping
filter

90o phase
shift

Input
Signal
xl (t)
Gaussian
noise
source

Balanced
modulator

Shaping
filter

Balanced
modulator

Rayleigh
fading
signal

Chapter 1: Frequency Selective Channel Simulations

Frequency Selective
1 N (Ts ) n j 2 fn / W
Cl t ; f
Cl t ; e
;

W i 1
W

yl (t ) Cl t ; sl (t )d Cl t ; f Sl ( f )e j 2 ft df
1 N (Ts ) n
n

Cl t ; sl t

W i 1
W
W
W: bandwidth of real band-pass signal
Input
Signal
xL(t)

Tapped-delay line
Delay-line

r0(t)ej0(t)

direct
path

...

r1(t)ej1(t)

rL(t)ejL(t)

Fading Signal Output

Receiver

Chapter 1: References
G.L. Turin. Communication through noisy, random-multipath channels. IRE
Convention Record, pp. 154-166, 1956.
P. Bello. Characterization of random time-variant linear channels. IEEE
Transactions in Communications, pp 360-393, 1963.
J.F. Ossanna. A model for mobile radio fading due to building reflections:
Theoretical and experimental waveform power spectra. Bell Systems
Technical Journal, 43:2935-2971, 1964.
R.H. Clarke. A statistical theory of mobile radio reception. Bell Systems
Technical Journal, 47:957-1000, 1968.
M.J Gans. A power-spectral theory of propagation in the mobile-radio
environment. IEEE Transactions on Vehicular Technology, VT-21(1):27-38,
1972.
H. Suzuki. A statistical model for urban radio propagation. IEEE
Transactions in Communications, 25:673-680, 1977.
T. Aulin. A modified model for the fading signal at a mobile radio channel.
IEEE Transactions on Vehicular Technology, VT-28(3):182-203, 1979.
[Link], [Link]. A statistical model for indoor multi-path
propagation. IEEE Journal on Selected Areas in Communications, 5(2):128137, 1987.

Chapter 1: References
M. Gudamson. Correlation model for shadow fading in mobile radio
systems. Electronics Letters, 27(23):2145-2146, 1991.
D. Giancristofaro. Correlation model for shadow fading in mobile radio
channels. Electronics Letters, 32(11):956-958, 1996.
A.J. Coulson, G. Williamson, R.G. Vaughan. A statistical basis for lognormal shadowing effects in multipath fading channels. IEEE Transactions
in Communications, 46(4):494-502, 1998.
E. Biglieri, J. Proakis, S. Shamai. Fading channels: Information-theoretic
and communication aspects. IEEE Transactions on Information Theory,
44(6):2619-2692, October 1998.
[Link]. Microwave mobile communications, New York, WileyInterscience, 1974.
K. Pahlavan, A.H. Levesque. Wireless Information Networks, New York,
Wiley-Interscience, 1995.
J.G. Proakis. Digital Communications, Mc-Graw-Hill, New-York, 1995.
T.S. Rappaport. Wireless Communications, Prentice Hall, 1996.