4.

Radio Channels

Overview The key features of radio channels are

reviewed.
Source The material is mainly based on Chapters 2
and 3 of the course book [1] and also [2, 3] have
been used.

Wireless Communications I @ University of Oulu, DCE & CWC 1

4. Radio Channels
 Outline of the Lecture
• Introduction
• Path loss
• Shadowing
• Small scale multipath fading
– Multipath channel impulse response
– Channel characterization
– Summary of small scale fading characterization
• Summary of radio channels

Wireless Communications I @ University of Oulu, DCE & CWC 2

4. Radio Channels
 Introduction
• Wireless channel distorts the signal in several
manners:
– free-space path loss (attenuation)
– shadowing (large scale fading)
– (small scale) fading due to multipath propagation.
• Wireless channels are always unpredictable to some
extent.
Wireless channels pose a main
technical / theoretical design / research challenge
for wireless communications.
Understanding the channels is crucial.

Wireless Communications I @ University of Oulu, DCE & CWC 3

4. Radio Channels
 Propagation Characteristics

• Path loss (includes average shadowing)

• Shadowing (due to obstructions)
• Small scale multipath fading

Slow
Pr/Pt Fast
Pt Pr Very slow
v

d =vt
d=vt
Wireless Communications I @ University of Oulu, DCE & CWC 4
4. Radio Channels
 Path Loss
• Signal attenuation due to radio wave propagation.
• More attenuation, the larger the distance.
• Behaviour explained by Maxwell’s equations.
Deterministic phenomenon in principle.
Ray tracing techniques approximate the propagation
of electromagnetic waves by representing the
wavefronts as simple particles and modeling
reflections and refractions.
• Due to complexity, stochastic channel models are
often needed.

Wireless Communications I @ University of Oulu, DCE & CWC 5

4. Radio Channels
 Path Loss Models

• Maxwell’s equations
– complex and impractical.
• Free space path loss model
– often too simple.
• Ray tracing models
– require site-specific information.
• Empirical Models
– do not always generalize to other environments.
• Simplified power fall off models
– main characteristics: good for high-level analysis
– used when path loss dominated by reflections
– most important parameter is the path loss exponent ,
determined empirically:
d0
Pr Pt K ,2 8.
d
Wireless Communications I @ University of Oulu, DCE & CWC 6
4. Radio Channels
 Free Space Line-of-Sight (LOS) Model

d = vt

• Path loss for unobstructed LOS path

• Power falls off:
– Inversely proportional to d 2
– Proportional to 2 (inversely proportional to f 2)

Wireless Communications I @ University of Oulu, DCE & CWC 7

4. Radio Channels
 Ray Tracing Approximation

• Represent wavefronts as simple particles.

• Geometry determines received signal from each
signal component.
• Typically includes reflected rays, can also include
scattered and diffracted rays.
• Requires site parameters:
– geometry
– dielectric properties.
Computer packages often used.
• Complex for a large number of rays.

Wireless Communications I @ University of Oulu, DCE & CWC 8

4. Radio Channels
 Ray Tracing (2)

• Two-ray model
– Parameters: antenna heights and critical distance.
– Power falls off
• proportional to d 2 (small d)
– Constructive add
• proportional to d 4 (d > dc)
– Constructive and destructive add
• independent of (f )
– Destructive add
• General ray tracing
– Reflections
– Scattering
– Diffraction

Wireless Communications I @ University of Oulu, DCE & CWC 9

4. Radio Channels
 Empirical Path Loss Models

• Okumura model (1-100 km, 150-1500 MHz)

– empirically based (site/freq specific)
– awkward (clumsy) (uses graphs).
• Hata model (often called Okumura-Hata model)
– analytical approximation to Okumura model
– commonly applied in cellular system simulations.
• COST 136 model
– extends the Hata model to higher frequency (2 GHz).
• Walfish/Bertoni model
– COST 136 extension to include diffraction from rooftops.
• Indoor models
– each indoor environment is different.

Wireless Communications I @ University of Oulu, DCE & CWC 10

4. Radio Channels
 Summary of Path Loss Models

• Path loss models simplify the Maxwell’s equations.

• Models vary in complexity and accuracy.
• Power falloff with distance is proportional to d 2 in free
space, d 4 in a two path model.
• General ray tracing is computationally complex.
• Empirical models often used in cellular system
simulations.
• Main characteristics of path loss captured in simple model
Pr = Pt K [d0/d]
K is unitless constant (antenna, average attenuation)
d0 is reference distance, and pathloss exponent

Wireless Communications I @ University of Oulu, DCE & CWC 11

4. Radio Channels
 Shadowing
• Large scale fading.
• Models attenuation from obstructions.
• Random due to the random number and type of
obstructions.
• Typically follows a log-normal distribution:
– dB value of power is normally distributed (Gaussian) with =0
(mean captured in path loss), 4 < 2 < 12 (empirical).
– Decorrelated over decorrelation distance Xc.
• Distance at which autocovariance equals 1/e of its maximum =>
empirical dependence is removed

Xc
Wireless Communications I @ University of Oulu, DCE & CWC 12
4. Radio Channels
 Combined Path Loss and Shadowing

• Linear model: lognormal Pr d0

K .
Pt d

10logK Slow

Pr /Pt [dB]
Very slow
-10
• dB model:
log d
Pr d
(dB) 10 log10 K 10 log10 dB ,
Pt d0
2
dB ~ N (0, )
Wireless Communications I @ University of Oulu, DCE & CWC 13
4. Radio Channels
 Outage Probability and Cell Coverage
Area
• Path loss: circular cells
• Path loss + random shadowing: amoeba cells:
– tradeoff between coverage and interference.
• Outage probability:
– probability received power below given minimum.
• Cell coverage area
– % of cell locations at desired power
– increases as shadowing variance decreases
– large % indicates interference to other cells. Pr
Path loss and average shadowing

Wireless Communications I @ University of Oulu, DCE & CWC 14

4. Radio Channels
 Model Parameters from Empirical
Measurements
• Fit a model to the observed measurement data.
• Path loss (K, ), d0 known:
– “Best fit” line through dB data.
– K obtained from measurements at d0.
– Exponent is MMSE estimate based on data.
– Captures mean due to shadowing.
• Shadowing variance
– Variance of data relative to path loss model (straight line)
with MMSE estimate for K (dB)
2
y
Pr (dB)
10
log(d )
log(d0)
Wireless Communications I @ University of Oulu, DCE & CWC 15
4. Radio Channels
 Summary of Shadowing

• Random attenuation due to shadowing modeled as

log-normal (empirical parameters).
• Shadowing decorrelates over decorrelation distance.
• Combined path loss and shadowing leads to outage
and amoeba-like cell shapes.
• Cellular coverage area dictates the percentage of
locations within a cell that are not in outage.
• Path loss and shadowing parameters are obtained
from empirical measurements.

Wireless Communications I @ University of Oulu, DCE & CWC 16

4. Radio Channels
 Small Scale Multipath Fading
• Rapid channel state variations due to multiple
propagation paths.
• Causes significant changes over short distances or
time intervals.
A very small movement can cause drastic change in
the signal power.
• The key channel characteristic to be considered in
this course.

Wireless Communications I @ University of Oulu, DCE & CWC 17

4. Radio Channels
 Multipath Channel Impulse Response

• A random number of multipath components, each

with random and time-varying
– amplitude
– phase
– Doppler shift
– delay.
A time-varying channel impulse response.

Wireless Communications I @ University of Oulu, DCE & CWC 18

4. Radio Channels
 Narrowband Channel Model
Signal bandwidth

• Assume delay spread = maxm,n| n(t) – m(t)| << 1/B.

u(t) u(t – ).
Individual multipath components not separable.
Received signal:
N t
j 2 f ct j n t
r t u t e n t e .
n 0

• No signal distortion (spreading in time).

Frequency-flat (non-selective) channel.
• Multipath affects complex scale factor in brackets.

Wireless Communications I @ University of Oulu, DCE & CWC 19

4. Radio Channels
 Wideband Channel Model

• The time difference between components exceeds

signal inverse bandwidth.
Individual multipath components are separable.
Frequency-selective channel.

Signal bandwidth
1 /B 1 /B

1 2

t t
Narrowband Wideband
Wireless Communications I @ University of Oulu, DCE & CWC 20
4. Radio Channels
 Time Varying Impulse Response

• Characterize the scale factor by setting u(t) = (t)

time-varying channel impulse response c( , t).
• Response of the channel at time instant t to input
impulse at time t – :
N
j t
c ,t n t e n
n t
n 1
– t is time when impulse response is observed
– t – is time when impulse put into the channel
is how long ago impulse was put into the channel for the
current observation
• path delay for MP component currently observed.

Wireless Communications I @ University of Oulu, DCE & CWC 21

4. Radio Channels
 Response to Sinusoid

• In-phase (I) and quadrature (Q) responses to a

sinusoid: N (t )
rI t n t cos n t,
n 1
N (t )
rQ t n t sin n t.
n 1

For large N(t) rI(t) and rQ(t) become approximately

jointly Gaussian by the central limit theorem (CLT).
Received signal characterized by its mean, autocorrelation,
and cross correlation.
If n(t) are uniformly distributed, the rI(t) and rQ(t) have
mean zero, are independent, and stationary.

Wireless Communications I @ University of Oulu, DCE & CWC 22

4. Radio Channels
 Received Signal Characteristics

• Received signal consists of many multipath components.

• Amplitudes change slowly.
• Phases change rapidly.
Constructive and destructive addition of signal components.
Amplitude fading of received signal (both wideband and
narrowband signals).
Rapid signal level fluctuation, called fading.
• Rapid either in
– space
• short distance changes the phase remarkably
– time
• due to mobility.

Wireless Communications I @ University of Oulu, DCE & CWC 23

4. Radio Channels
 Channel Characterization

• Received signal characterized by its mean,

autocorrelation, and cross correlation, since it is
Gaussian.
• Assume n ~ U[0,2 ]
• Recall that n is the multipath arrival angle
• Autocorrelation of I and Q signals ( p = E( n )):
2

ArI ArQ 1
2 p
E n [cos 2 f Dn ], f Dn v cos n / .
• Cross correlation:
1
ArI ,rQ 2 p E n [sin 2 f Dn ] ArQ ,rI .
• Autocorrelation of the received signal:
Ar ArI cos 2 f c ArI ,rQ sin 2 f c .
Wireless Communications I @ University of Oulu, DCE & CWC 24
4. Radio Channels
 Uniform Scattering

• A common assumption: the scatterers are uniformly

distributed and their number approaches infinity.
Clarke’s autocorrelation function:
ArI ArQ 1
2 p J 0 2 fD .
The corresponding power spectral density (PSD)
know as Jakes’s Doppler spectrum:
Sr f 1 S rI f fc S rI f fc , Sr(f)
2

S rI f F 1
2 pJ 0 2 fD .
– Widely applied in simulations.
fc-fD fc fc+fD

Wireless Communications I @ University of Oulu, DCE & CWC 25

4. Radio Channels
 Signal Envelope Distributions

• Signal envelope: z t r t rI2 t r Q2 t .

• If rI(t) and rQ(t) are Gaussian with zero mean (no
LOS component) Rayleigh distribution:
z z2
p z 2
exp z 0.
z 2 2z
2
– Chi-square distribution for r t r I2 t r Q2 t .
• In case of a dominant LOS component, rI(t) and
rQ(t) are Gaussian with non-zero mean Rician.
• Measurements: Nakagami-m distribution.
– Rayleigh (m=1) and Rician are special cases.

Wireless Communications I @ University of Oulu, DCE & CWC 26

4. Radio Channels
 Average Fade Duration

• How long time the signal envelope stays below a

target level Z with SNR of ?
t1 t2 t3
Z
T

• Derived from level crossing rate of fading process.

• For Rayleigh fading:
2
TLZ
Pr z t Z e 1
tR 1
TLZ
ti .
i 1 LZ 2 fD
– Depends on ratio of target to average level Z .
Pr
– Inversely proportional to Doppler frequency fD.
2
– Level crossing rate: LZ 2 f De .

Wireless Communications I @ University of Oulu, DCE & CWC 27

4. Radio Channels
 Scattering Function

• Fourier transform of the time-varying impulse

response c ( t) relative to time t.
• Since the underlying process is Gaussian, the mean
(often zero) and correlation are needed only.
• Autocorrelation is Ac( 1, 2, t)=Ac( , t).
• The statistical scattering function:
S ( )=F t[Ac( , t)].

Wireless Communications I @ University of Oulu, DCE & CWC 28

4. Radio Channels
 Summary of Small Scale Fading
Characterization
• Statistical multipath model leads to a time-varying
channel impulse response.
Received signal has random amplitude fluctuations.
• Narrowband model and CLT lead to I and Q
components that are stationary Gaussian processes.
Completely characterized by their mean, autocorrelation,
and cross correlation.
• Assuming uniform phase offsets:
– The process has zero mean.
– Auto and cross correlation depend on AoAs of multipath
components.

Wireless Communications I @ University of Oulu, DCE & CWC 31

4. Radio Channels
 Summary of Small Scale Fading
Characterization (2)
• Uniform scattering makes autocorrelation of I and Q
to follow Bessel function
Clarke’s autocorrelation or Jakes’s Doppler spectrum.
– Signal components are decorrelated over half wavelength.
– Cross correlation is zero, i.e., I and Q components are
independent.
• Fading distribution depends on environment.
– Rayleigh, Ricean, and Nakagami all commonly applied.
• Average fade duration important for system issues.
– Determines how long a user is in continuous outage.
• Wideband models characterized by scattering
function: measures power vs. delay and Doppler
frequency.
Wireless Communications I @ University of Oulu, DCE & CWC 32
4. Radio Channels
 Summary of Channel Selectivity Classes

• Frequency selectivity
– Frequency-flat (non-selective) if Tm << T or Bc >> B.
No frequency dispersion or intersymbol interference (ISI).
– Otherwise frequency-selective.
Frequency dispersion or intersymbol interference (ISI).
• Time selectivity
– Time-flat or slowly fading if Tc >> T or Bd << B.
Channel approximately constant from symbol to symbol.
– Otherwise time-selective or fast fading.
Coherent communications may be impossible.

Wireless Communications I @ University of Oulu, DCE & CWC 35

4. Radio Channels
Delay spread
Delay spread
Realistic example
…ctd
Coherence bandwidth
 …ctd
• .
Example
...ctd
Types of fading
…Ctd
All types of fading summary
 Summary of Radio Channels

• Path loss (includes average shadowing)

• Shadowing (due to obstructions)
• Small scale multipath fading

Slow
Pr/Pt Fast
Pt Pr Very slow
v

d =vt
d=vt
Wireless Communications I @ University of Oulu, DCE & CWC 36
4. Radio Channels
 Summary: Path Loss and Shadowing

• Path loss models simplify the Maxwell’s equations.

– Models vary in complexity and accuracy.
– Power falloff with distance is proportional to d 2 in free
space, d 4 in a two path model.
– General ray tracing is computationally complex.
– Empirical models often used in cellular system simulations.
– Main characteristics of path loss captured in simple model
P r = P t K [ d0 / d] .
• Random attenuation due to shadowing modeled as
log-normal.
– Decorrelates over decorrelation distance.
– Combined path loss and shadowing leads to outage and
amoeba-like cells.
– Path loss and shadowing parameters are obtained from
empirical measurements.
Wireless Communications I @ University of Oulu, DCE & CWC 37
4. Radio Channels
 Summary: Small Scale Multipath Fading

• Statistical multipath model leads to a time-varying

channel impulse response.
Received signal has random amplitude fluctuations.
• Narrowband model and CLT lead to I and Q
components that are stationary Gaussian processes.
• Uniform scattering makes autocorrelation of I and Q
to follow Bessel function
• Fading distribution depends on environment.
• Average fade duration important for system issues.
• Wideband models characterized by scattering
function: measures power vs. delay and Doppler
frequency.
Wireless Communications I @ University of Oulu, DCE & CWC 38
4. Radio Channels
 Summary of Channel Selectivity Classes

• Frequency selectivity
– Frequency-flat (non-selective) if Tm << T or Bc >> B.
No frequency dispersion or intersymbol interference (ISI).
– Otherwise frequency-selective.
Frequency dispersion or intersymbol interference (ISI).
• Time selectivity
– Time-flat or slowly fading if Tc >> T or Bd << B.
Channel approximately constant from symbol to symbol.
– Otherwise time-selective or fast fading .
Coherent communications may be impossible.

Wireless Communications I @ University of Oulu, DCE & CWC 39

4. Radio Channels
 Introduction to Shannon Channel Capacity
The fundamental problem of communication is that of reproducing at one
point, either exactly or approximately, a message selected at another point.
~ Claude Shannon

Channel

Channel Capacity :
C = B log2 (1 + SN R)

2017/3/7 Lecture 2: Capacity of Wireless Channels 3

 Introduction to Shannon Channel Capacity
• Consider a discrete-time additive white Gaussian noise (AWGN) channel
with channel input/output relationship y[i] = x[i] + n[i] , where x[i] is the
channel input at time i, y[i] is the corresponding channel output,
and n[i] is a white Gaussian noise random process.
• Assume a channel bandwidth B and transmit power P.
• The channel SNR, the power in x[i] divided by the power in n[i] , is
constant and given by = P/(N0 B) , where N0 is the power spectral
density of the noise.
• The capacity of this channel is given by Shannon’s well-known formula
C = B log2 (1 + )
where the capacity units are bits/second (bps).
• Shannon’s coding theorem proves that there exists a code that achieves
data rates arbitrarily close to capacity with arbitrarily small probability
of bit error.
2017/3/7 Lecture 2: Capacity of Wireless Channels 4
 Introduction to Shannon Channel Capacity
• Shannon proved that channel capacity equals the mutual information of
the channel maximized over all possible input distributions:
X ✓ ◆
p(x, y)
C = max I(X; Y ) = max p(x, y) log . (4.3)
p(x) p(x)
x,y
p(x)p(y)

For the AWGN channel, the maximizing input distribution is Gaussian,

which results in the channel capacity C = B log2 (1 + ) .
• At the time that Shannon developed his theory of information, data rates
over standard telephone lines were on the order of 100 bps. Thus, it was
believed that Shannon capacity, which predicted speeds of roughly 30
Kbps over the same telephone lines, was not a very useful bound for real
systems.
• However, breakthroughs in hardware, modulation, and coding
techniques have brought commercial modems of today very close to the
speeds predicted by Shannon in the 1950s.

2017/3/7 Lecture 2: Capacity of Wireless Channels 5

 Mutual Information
• The converse theorem shows that any code with rate R > C has a
probability of error bounded away from zero.
• For a memoryless time-invariant channel with random input x and
random output y, the channel’s mutual information is defined as
X ✓ ◆
p(x, y)
I(X; Y ) = p(x, y) log .
p(x)p(y)
x2X ,y2Y

where the sum is taken over all possible input and output pairs x 2 X and
y 2 Y for X and Y the input and output alphabets.
• Mutual information can also be written in terms of the entropy in the
channel output y and conditional output y|x as

where

2017/3/7 Lecture 2: Capacity of Wireless Channels 7

 The AWGN Channel
• The most important continuous alphabet channel is the Gaussian
channel depicted in the following figure. This is a time-discrete
channel with output Yi at time i, where Yi is the sum of the input Xi
and the noise Zi .
• The noise Zi is drawn i.i.d. from a Gaussian distribution with
variance N. Thus,
Yi = Xi + Z i , Zi ⇠ N (0, N )

• The noise Zi is assumed to be

independent of the signal Xi .
• If the noise variance is zero or the
input is unconstrained, the
capacity of the channel is infinite.
Figure: A Gaussian Channel

2017/3/7 Lecture 2: Capacity of Wireless Channels 8

 Capacity in AWGN
• Given E[Y 2 ] = P + N, the entropy of Y is bounded by 12 log 2⇡e(P + N )
(the Gaussian distribution maximizes the entropy for a given variance).
• Applying this result to bound the mutual information, we obtain

I(X; Y ) = h(Y ) h(Z)

1 1
 log 2⇡e(P + N ) log 2⇡eN
2 2
✓ ◆
1 P
= log 1 +
2 N
• Hence, the information capacity of the Gaussian channel is
✓ ◆
1 P
C = max I(X; Y ) = log 1 + ,
E[X ]P
2 2 N
and the maximum is attained when X ⇠ N (0, P ) .

2017/3/7 Lecture 2: Capacity of Wireless Channels 15

 Capacity of Bandlimited Channels
• Let the channel be used over the time interval [0,T]. In this case, the
energy per sample is P T /2BT = P/2B , the noise variance per sample
is N0 2B T = N0 /2 , and hence the capacity per sample is
2 2BT
! ✓ ◆
P
1 1 P
C = log2 1 + 2B = log2 1 + (bits per sample)
2 N0 /2 2 N0 B

• Since there are 2B samples each second, the capacity of the channel
can be rewritten as ✓ ◆
P
C = B log2 1 + (bits/s, bps)
N0 B
(This equation is one of the most famous formulas of information
theory. It gives the capacity of a bandlimited Gaussian channel with
noise spectral density N0 /2 watts/Hz and power P watts.)
• If we let B ! 1 in the above capacity formula, we obtain
P (for infinite bandwidth channels, the capacity grows
C= log2 e
N0 linearly with the power.)
2017/3/7 Lecture 2: Capacity of Wireless Channels 20
 Capacity of Flat-Fading Channels
• Shannon capacity of a fading channel with receiver CSI for an average
power constraint P can be obtained as
Z 1
C= B log2 (1 + )p( )d (4.4)
0

• By Jensen’s inequality,
Z
E[B log2 (1 + )]= B log2 (1 + )p( )d  B log2 (1 + E[ ])
= B log2 (1 + ),
where is the average SNR on the channel.
• Here we see that the Shannon capacity of a fading channel with receiver
CSI only is less than the Shannon capacity of an AWGN channel with the
same average SNR.
• In other words, fading reduces Shannon capacity when only the receiver
has CSI.

2017/3/7 Lecture 2: Capacity of Wireless Channels 25

 Capacity of Flat-Fading Channels
• Example 4.2: Consider a flat-fading channel with i.i.d. channel gain g[i]
which can take on three possible values: with probability ,
with probability with probability .
The transmit power is 10 mW, the noise spectral density is W/Hz,
and the channel bandwidth is 30 KHz. Assume the receiver has knowledge
of the instantaneous value of g[i] but the transmitter does not. Find the
Shannon capacity of this channel and compare with the capacity of an
AWGN channel with the same average SNR.

2017/3/7 Lecture 2: Capacity of Wireless Channels 26

 Capacity of Flat-Fading Channels

• Capacity with outage applies to slowly-varying channels, where the

instantaneous SNR γ is constant over a large number of transmissions (a
transmission burst) and then changes to a new value based on the fading
distribution.
• If the channel has received SNR γ during a burst, then data can be sent
over the channel at rate B log2 (1 + )with negligible probability of error
• Capacity with outage allows bits sent over a given transmission burst to
be decoded at the end of the burst with some probability that these bits
will be decoded incorrectly.
• Specifically, the transmitter fixes a minimum received SNR min and
encodes for a data rate C = B log2 (1 + min ) .
• The data is correctly received if the instantaneous received SNR is greater
than or equal to min .
2017/3/7 Lecture 2: Capacity of Wireless Channels 27