10 00

11 01

Mobilkommunikation - Mobile Communications

Lecture 2: Wireless Transmission

Prof. Dr.-Ing. Markus Fidler

Institute of Communications Technology

Leibniz Universität Hannover

April 21, 2023

© Markus Fidler | IKT LUH | 1/55

 10 00

Outline 11 01

Radio Propagation
Line-of-sight propagation
Non-line-of-sight and multipath propagation
Fading

Channel Capacity

Encoding and Modulation

Multiplexing

Spread Spectrum

© Markus Fidler | IKT LUH | 2/55

 10 00

Antennas 11 01

y y
▶ isotropic antenna
z z
▶ theoretic concept of
point antenna x x
▶ omni-directional in
x, y, and z
▶ directed antennas omni-directional directional

▶ directive effects in z z
all real antennas
▶ size of antennas is
in the order of the x x
wavelength
▶ sectorized antennas
top view, 3 sectors top view, 6 sectors

© Markus Fidler | IKT LUH | 3/55

 10 00

Antennas (2) 11 01

Simple antenna dipoles

▶ length in the order of the wavelength, e.g. λ/4 or λ/2
Marconi Hertzian
antenna dipole

radiation pattern of a Hertzian dipole

y y z

x z x

side view side view top view

Real antennas are not isotropic

▶ antenna gain: power in the direction of the main lobe relating
to the power of an isotropic radiator

© Markus Fidler | IKT LUH | 4/55

 10 00

Antenna diversity 11 01

Antenna diversity: use of 2 or more antennas, also called

multi-element antenna arrays
▶ diversity at
▶ sender
▶ receiver
▶ both
▶ switched or selection diversity n
▶ receiver selects the antenna
with the largest output
▶ diversity combining
+
▶ receiver combines the output
power of all antennas to
produce gain
▶ cophasing is used to avoid
cancellation of signals

© Markus Fidler | IKT LUH | 5/55

 10 00

Multiple-input multiple-output 11 01

Multiple-input multiple-output (MIMO) refers to multi antenna

systems at sender and receiver.

Given two antennas at the sender and receiver each, the signals at
the receiver yj depend on the respective channel condition hi,j and
the signals at the sender xi
    
y1 (t) h11 h12 x1 (t)
=
y2 (t) h21 h22 x2 (t)
MIMO can be used in different ways
▶ transmit and receive diversity
▶ beamforming
▶ spatial multiplex

© Markus Fidler | IKT LUH | 6/55

 10 00

Signal propagation 11 01

distance

▶ transmission range
low error rate at the
receiver
▶ detection range
transmitted power differs
from background noise
▶ interference range transmission

sender adds to background

noise detection

interference

© Markus Fidler | IKT LUH | 7/55

 10 00

Free-space path loss 11 01

Consider an isotropic antenna that emits an electromagnetical

signal with power Pt . The signal propagates at the speed of light c.
It forms a sphere with surface 4πd2 where d is the distance from
the sender resulting in the inverse square law (Friis)
1
S = Pt
4πd2
where S is the power per unit of area, i.e., Watt/m2 . Higher
exponents in the range of 2 . . . 5 apply if the assumption of free
space does not hold.

The receive power is Pr = S · A where A is the aperture area. For

isotropic antennas A = λ2 /(4π) where λ is the wave length, i.e.

λ2
Pr = S .
4π

© Markus Fidler | IKT LUH | 8/55

 10 00

Free-space path loss (2) 11 01

By insertion the received power for free space propagation is

Gt Gr
Pr = Pt
Lp
▶ Pr = received power
▶ Pt = transmitted power
▶ Gr = gain of the receiving antenna
▶ Gt = gain of the transmitting antenna
▶ Lp = free space path loss
 2  
4πd 4πd
Lp = or Lp = 20 log [dB]
λ λ
▶ λ = wave length = c/f
▶ c = speed of light, in free space c = 3 · 108 m/s
▶ f = carrier frequency

© Markus Fidler | IKT LUH | 9/55

 10 00

Free-space path loss (3) 11 01

Free space path loss after insertion of constants

Lp = −147.55 + 20 log f + 20 log d dB

▶ f carrier frequency in MHz

▶ d distance in m where d > 1 km

Useful rules to remember

▶ doubling the length of a path increases free space attenuation
by 6 dB
▶ doubling the frequency also increases free space attenuation
by 6 dB

© Markus Fidler | IKT LUH | 10/55

 10 00

Attributes of radio wave propagation 11 01

▶ reflection:
if a radio wave impinges
upon the boundary of two
materials with different
electromagnetic
properties; causes
shadowing
▶ diffraction:
radio waves bend in the
neighborhood of obstacles
▶ scattering:
if a radio wave hits a
rough surface or an object
having a size in the order
of the wavelength

© Markus Fidler | IKT LUH | 11/55

 10 00

Attributes of radio wave propagation (2) 11 01

▶ refraction: the velocity of

of electromagnetic waves
depends on the density of
the medium; at the
boundary to a denser
medium waves are bent
towards the medium

This effect causes that radio waves are bent towards the earth,
since the density of the atmosphere increases closer to the ground.

© Markus Fidler | IKT LUH | 12/55

 10 00

Propagation behavior at different frequencies 11 01

▶ ground wave < 2 Mhz: Low frequencies can follow the

curvature of the earth over long distances
▶ sky wave 2-30 MHz: Short waves are reflected at the
ionosphere and at the surface of the earth
▶ line-of-sight > 30 Mhz: Ultrashort waves and below closely
follow a straight line of sight

© Markus Fidler | IKT LUH | 13/55

 10 00

Earth curvature and line-of-sight propagation 11 01

Line-of-sight (LOS) propagation

▶ without any obstructions due to obstacles etc.
▶ Pythagoras (r + h)2 = r2 + (d/2)2
▶ earth radius r = 6370 km
▶ height of the antenna, e.g. h = 30 m
√
▶ distance to horizon d/2 = 2rh + h2 , e.g. d/2 = 20 km

© Markus Fidler | IKT LUH | 14/55

 10 00

Multipath propagation 11 01

direct wave

ground reflected wave

Signals propagate along different paths due to reflection,

diffraction, scattering, and refraction.
Since the speed of light is finite, different paths result in different
receive times resulting in delay spread at the receiver
▶ the signal is smeared or separated into several weaker impulses
▶ originally separated signals interfere with each other
E.g. in case of 5 km path difference the delay spread is 16 µs.
The problem is addressed using training sequences to tune an
equalizer at the receiver that seeks to compensate the distortion.

© Markus Fidler | IKT LUH | 15/55

 10 00

Delay spread and intersymbol interference 11 01

send power

receive power
delay spread

time time
send power

receive power
intersymbol
interference

time time

© Markus Fidler | IKT LUH | 16/55

 10 00

Doppler shift 11 01

In case of relative motion between sender and receiver a Doppler

shift fD , i.e. change in frequency, occurs:
v v
fD = f=
c λ
where
▶ c = speed of light
▶ v = relative speed
▶ f = carrier frequency
▶ λ = carrier wave length

Doppler shift is relevant in case of quickly moving transceivers.

© Markus Fidler | IKT LUH | 17/55

 10 00

Mobile radio signal fading 11 01

power

time

Variation of the attenuation of the radio signal at a mobile receiver

▶ short-term fading: usually over distances of about half a
wavelength; signals due to multipath propagation may
strengthen or cancel each other
▶ long-term fading: average over tens of wavelengths;
fluctuations due to distance, obstacles, shadowing

© Markus Fidler | IKT LUH | 18/55

 10 00

Fading models 11 01

Fading is frequently modeled by a random process:

▶ Rayleigh distribution: short-term fading in multipath
environments where there is no line-of-sight
▶ Rician distribution: short-term fading in multipath
environments with line-of-sight
▶ Lognormal distribution: long-term fading, shadowing

© Markus Fidler | IKT LUH | 19/55

 10 00

Coherence Bandwidth 11 01

Signals received over multiple paths can strengthen or cancel each

other, depending on delay spread and frequency
▶ the channel becomes frequency selective
▶ the reciprocal of delay spread is called coherence bandwidth

Frequency components of a signal that are separated by

▶ less than the coherence bandwidth (e.g., narrowband
channel), will fade together, called flat fading
▶ more than the coherence bandwidth (e.g., wideband channel),
will fade independently, called frequency-selective fading

© Markus Fidler | IKT LUH | 20/55

 10 00

Coherence Time 11 01

In case of mobility, signals received over multiple paths can have

different Doppler shifts causing Doppler spread
▶ the channel gain varies over time
▶ the reciprocal of Doppler spread is called coherence time

A symbol interval (inverse of symbol rate) that is

▶ smaller than the coherence time, will not experience a
changing channel, referred to as slow fading
▶ larger than the coherence time, will experience a changing
channel, referred to as fast fading

© Markus Fidler | IKT LUH | 21/55

 10 00

Outline 11 01

Radio Propagation
Line-of-sight propagation
Non-line-of-sight and multipath propagation
Fading

Channel Capacity

Encoding and Modulation

Multiplexing

Spread Spectrum

© Markus Fidler | IKT LUH | 22/55

 10 00

Noisy channel 11 01

transmitter noisy channel receiver

Radio transmission over a wireless channel is subject to noise.

Useful and frequently used is the is additive white Gaussian noise
(AWGN) model
▶ linear addition of noise
▶ noise is white i.e.
▶ it applies to all frequency bands
▶ it has a constant spectral density independent of frequency
▶ noise samples have a Gaussian distribution
Gaussian noise has many natural causes such as thermal noise,
radiation from the earth and sun, etc. It is, however, not a model
for fading, interference, dispersion etc.

© Markus Fidler | IKT LUH | 23/55

 10 00

Shannon limit 11 01

Shannon’s limit for AWGN channels states that

 
S
C = B log2 1 +
N

where
▶ C = theoretical limit of the data rate [bit/s]
▶ B = channel bandwidth [Hz]
▶ S = signal power [W]
▶ N = noise power [W]
▶ S/N = signal-to-noise ratio

© Markus Fidler | IKT LUH | 24/55

 10 00

Shannon limit (2) 11 01

Given bits of duration T = 1/C the energy per bit is Eb = S/C.

The noise can be expressed as N = N0 · B where N0 is the

spectral noise power density [W/Hz] and B is the bandwidth [Hz].

By insertion  
C Eb C
= log2 1 +
B N0 B
where
▶ Eb /N0 = signal-to-noise density ratio
▶ C/B = η spectral efficiency [bit/(s·Hz)]

For η → 0 it follows that Eb /N0 = ln(2) ≈ 0.7 ≈ −1.6 dB, i.e.

information transfer is impossible if Eb /N0 < 0.7 even if the
bandwidth is arbitrarily large.

© Markus Fidler | IKT LUH | 25/55

 10 00

Shannon limit (3) 11 01

1
10
spectral efficiency η [bit/(s Hz)]

bandwidth limited regime

0
10

power limited regime

−1
10
−1.6 0 2 4 6 8 10
signal−to−noise density ratio Eb/N0 [db]

© Markus Fidler | IKT LUH | 26/55

 10 00

Impact of the signal power 11 01

Low SNR regime → small η

 
S S
η = log2 1 + ≈ · const
N N

Doubling S doubles η.

High SNR regime → large η

 
S
η ≈ log2
N

Doubling S increases η by 1 bit/(s Hz).

Generally, it follows that doubling S results in an increase of η

approximately by min{1, η}.

© Markus Fidler | IKT LUH | 27/55

 10 00

Outline 11 01

Radio Propagation
Line-of-sight propagation
Non-line-of-sight and multipath propagation
Fading

Channel Capacity

Encoding and Modulation

Multiplexing

Spread Spectrum

© Markus Fidler | IKT LUH | 28/55

 10 00

Digital communication 11 01

quantizer source
sampling
analog compander encoder

channel
interleaver modulator antenna
encoder
digital

digital
channel
antenna demodulator deinterleaver
decoding

source dequantizer analog

low pass
decoder compander

© Markus Fidler | IKT LUH | 29/55

 10 00

Modulation schemes 11 01

f(t)

s(t) = a · cos(2πf t + ϕ)
A basic cosine (sine) function has three parameters that can be
modulated to encode data:
▶ a = amplitude,
▶ f = frequency,
▶ ϕ = phase.

© Markus Fidler | IKT LUH | 30/55

 10 00

Send and receive path 11 01

digital digital baseband analog carrier

antenna
01001110 modulation modulation

carrier analog baseband synchronization digital

antenna
demodulation decision 01001110

Radio send and inverse receive path

▶ digital modulation modulates digital (binary) data onto a
baseband signal
▶ analog modulation takes a baseband signal and modulates it
onto the radio carrier, e.g. usually shifts the baseband signal
into a different frequency band

© Markus Fidler | IKT LUH | 31/55

 10 00

Amplitude shift keying 11 01

f(t)

1 0 1

Amplitude shift keying (ASK)

▶ different amplitudes represent different symbols
▶ simple scheme with low bandwidth requirement
▶ susceptible to interference that influences the amplitude
▶ therefore usually not used for wireless transmission
▶ but used for optical transmission, i.e. light pulses

© Markus Fidler | IKT LUH | 32/55

 10 00

Frequency shift keying 11 01

f(t)

1 0 1

Frequency shift keying (FSK)

▶ different frequencies represent different symbols
▶ usually sudden changes in the phase are avoided, so-called
continuous phase modulation
▶ demodulation can simply be performed using different
bandpass filters and a comparator
▶ frequently used for wireless transmission

© Markus Fidler | IKT LUH | 33/55

 10 00

Phase shift keying 11 01

f(t)

1 0 1

Phase shift keying (PSK)

▶ shifts in the phase of the signal represent different symbols
▶ to decode the signal the receiver has to synchronize in
frequency and phase using a so-called phase lock loop
▶ more resistent to interference than FSK but also more
complex transmitter and receiver

© Markus Fidler | IKT LUH | 34/55

 10 00

Advanced frequency shift keying 11 01

Minimum shift keying (MSK) is a widely used binary FSK scheme

▶ continuous phase modulation
▶ frequency changes occur at the carrier zero crossings
▶ binary scheme with two frequencies
▶ difference between low and high frequency is half the data rate
▶ minimum frequency spacing for continuous phase

Gaussian MSK (GMSK) basically adds a Gaussian lowpass filter to

reduce spectral side lobes. It is used e.g. for GSM and DECT.

© Markus Fidler | IKT LUH | 35/55

 10 00

Minimum shift keying 11 01

MSK modulation procedure

▶ bits are separated in odd and even positions
▶ each bit is doubled in time
▶ depending on the pair (even,odd) the frequency is chosen
▶ (-1, 1): low frequency
▶ ( 1, 1): high frequency
▶ ( 1,-1): low frequency inverted phase
▶ (-1,-1): high frequency inverted phase
▶ that is
▶ low frequency if even · odd = −1 and high otherwise
▶ phase inversion if odd = −1 and no inversion otherwise

© Markus Fidler | IKT LUH | 36/55

 10 00

Minimum shift keying (2) 11 01

1
data odd even odd even odd even odd even
0
t ▶ T = bit
even
1
duration
-1 t
▶ R = 1/T
1
odd
-1 t
data rate
▶ fH = 1/T
low
high freq.
t
by choice
▶ fL =
high
t fH − R/2
= 1/(2T )
low freq.
MSK
t

© Markus Fidler | IKT LUH | 37/55

 10 00

Advanced phase shift keying 11 01

Phase shift keying (PSK) is used in many ways

▶ Binary PSK (BPSK) only uses two phases with 180◦ shift to
encode one bit
▶ Quadrature PSK (QPSK) uses four phases to encode two bit
into one shift
The receiver uses a reference to detect phase shifts
▶ reference signal at the receiver, requires synchronization
▶ differential encoding, i.e. phase shifts relative to the previous
phase, e.g. DQPSK
Combination of PSK with ASK
▶ Qudrature amplitude modulation (QAM) superposes two
signals: in-phase (cos) plus quadrature (sin) each with
▶ 4 different amplitudes ⇒ 16 different symbols (16-QAM)
▶ 8 different amplitudes ⇒ 64 different symbols (64-QAM)

© Markus Fidler | IKT LUH | 38/55

 10 00

Multi-carrier modulation 11 01

High bit-rate transmission is vulnerable to inter-symbol

interference (ISI). To this end, Multi-carrier modulation (MCM)
performs a parallel transmission using several subcarriers.
▶ a data stream with rate R has bit duration 1/R
▶ splitting the stream into N substreams the bit duration
increases to N/R thus reducing ISI
▶ the bandwidth of the radio carrier f is split into N adjacent
subcarriers of f /N width
▶ each of the substreams is assigned to one subchannel
An additional benefit is that frequency selective fading only
influences certain subcarriers.

Orthogonal frequency division multiplexing (OFDM) is an efficient

implementation of MCM that is used e.g. in IEEE 802.11a.

© Markus Fidler | IKT LUH | 39/55

 10 00

Outline 11 01

Radio Propagation
Line-of-sight propagation
Non-line-of-sight and multipath propagation
Fading

Channel Capacity

Encoding and Modulation

Multiplexing

Spread Spectrum

© Markus Fidler | IKT LUH | 40/55

 10 00

Multiplexing 11 01

The goal is to divide radio resources and to assign space, time,

frequency, and code to individual pairs of senders and receivers. To
this end, multiplexing can use any combination of
▶ space division multiplexing (SDM)
▶ time division multiplexing (TDM)
▶ frequency division multiplexing (FDM)
▶ code division multiplexing (CDM)
At the same time the target is to
▶ minimize interference
▶ maximize utilization

© Markus Fidler | IKT LUH | 41/55

 10 00

Space division multiplexing 11 01

interference
▶ threedimensional space
transmission
▶ e.g. omnidirectional
guard space
propagation
▶ transmission range
z
▶ interference range
▶ guard spaces x

▶ spatial reuse

© Markus Fidler | IKT LUH | 42/55

 10 00

Frequency division multiplexing 11 01

▶ non-overlapping frequency code

bands
▶ adjacent channel interference
▶ guard spaces
▶ simple scheme without need for
complex coordination
▶ the receiver basically tunes to
the frequency of the sender frequency

▶ the static frequency allocation

is, however, inflexible
time guard spaces

© Markus Fidler | IKT LUH | 43/55

 10 00

Time division multiplexing 11 01

▶ all senders use the same

frequency but at different times
▶ receivers have to listen
code
exactly at the right time
▶ synchronization of different
senders is needed
▶ exact clocks or
synchronization signals
▶ co-channel interference
▶ guard spaces
frequency
▶ flexible since timeslots can be
reassigned to heavily loaded guard spaces

senders
time
▶ frequently combined with FDM,
e.g. in GSM

© Markus Fidler | IKT LUH | 44/55

 10 00

Code division multiplexing 11 01

▶ each channel uses a specific,

individual code
▶ codes have to be sufficiently code

orthogonal to realize guard

spaces
▶ secret codes achieve security
▶ good protection against guard spaces

interference
▶ high complexity of the receiver frequency

▶ need precise synchronization

▶ need sophisticated power
control, other codes appear time

as background noise

© Markus Fidler | IKT LUH | 45/55

 10 00

Outline 11 01

Radio Propagation
Line-of-sight propagation
Non-line-of-sight and multipath propagation
Fading

Channel Capacity

Encoding and Modulation

Multiplexing

Spread Spectrum

© Markus Fidler | IKT LUH | 46/55

 10 00

Spread spectrum 11 01

Spread spectrum techniques spread the frequency band that is

used
▶ narrowband signals are converted into broadband signals
▶ achieves robustness against narrowband interference
▶ interference is frequency selective and can vary quickly in time
▶ assuming the same energy, the power of the spread signal can
be much lower
▶ signals can look like background noise and may be hard to
detect
E.g. a narrowband FDM system separates frequency bands while a
spread spectrum system combines these bands and uses CDM.

© Markus Fidler | IKT LUH | 47/55

 10 00

Impact of narrowband noise 11 01

SNR

SNR
narrowband signal
guard space

spread spectrum signal

frequency frequency

In spread spectrum systems narrowband noise affects all signals,

but to a much lesser extend as opposed to narrowband signals.

© Markus Fidler | IKT LUH | 48/55

 10 00

Direct sequence spread spectrum 11 01

Direct sequence spread spectrum

▶ signals are spread before modulation
▶ bits are XORed with a so-called chipping sequence
▶ bits have a duration Tb
▶ chips are small pulses of duration Tc
▶ the spreading factor is s = Tb /Tc
▶ given the bandwidth of the original signal is w the bandwidth
of the spread signal is s · w
▶ chipping sequences use specific codes that appear as random
noise, also called pseudo-noise
▶ e.g. Barker codes
▶ spreading factors typically range from 10 up to 10000 if secret
codes are needed

© Markus Fidler | IKT LUH | 49/55

 10 00

Direct sequence spread spectrum (2) 11 01

user bits
1

0
time
Tb
XOR
chip sequence Tc

0
time

=
signal
1

0
time

© Markus Fidler | IKT LUH | 50/55

 10 00

Direct sequence spread spectrum (3) 11 01

Decoding spread spectrum signals is much more complex

▶ basically another XOR with exactly the same chipping
sequence at the receiver
▶ requires, however, exact synchronization at bit and chip level
▶ signals may be distorted at the receiver, i.e. chips are tilted
▶ consequently the result after XOR (ideally a single bit) may
alternate between zero and one
▶ the receiver uses an integrator that generates a sort of
”majority” decision

Multi-path propagation and different delays per path are a

challenge for spread spectrum. So-called rake receivers use several
decoders that are synchronized to different delays resp. paths.

© Markus Fidler | IKT LUH | 51/55

 10 00

Frequency hopping spread spectrum 11 01

Spread spectrum can be implemented using frequency hopping

▶ several frequency bands are used according to a fixed hopping
sequence, i.e. combination of FDM and TDM
▶ frequency hopping is implemented as a part of analog
modulation, i.e. after digital modulation
In case of several transmitters the hopping sequences have to be
coordinated to avoid that transmitters use the same frequency at
the same time
▶ coordination of all transmitters or
▶ pseudo-random hopping with different hopping sequences that
are sufficiently orthogonal

© Markus Fidler | IKT LUH | 52/55

 10 00

Frequency hopping spread spectrum (2) 11 01

The time spent on a channel is called the dwell time Td .

▶ slow hopping
▶ speed of hopping is slower than the bit duration Td > Tb , i.e.
several bits are transmitted before a hop is performed
▶ used e.g. in GSM and Bluetooth
▶ fast hopping
▶ speed of hopping is faster than the bit duration Td < Tb , i.e.
several hops are performed while a single bit is transmitted
▶ better robustness against narrowband interference and
frequency selective fading
▶ more complex due to high synchronization requirements

© Markus Fidler | IKT LUH | 53/55

 10 00

Frequency hopping spread spectrum (3) 11 01

user bits

time
Tb
slow hopping Td

f3

f2

f1

time

fast hopping Td

f3

f2

f1

time

© Markus Fidler | IKT LUH | 54/55

 10 00

Literature 11 01

▶ Jochen Schiller, Mobile Communications, Second Edition,

Addison-Wesley, 2003.
▶ Vijay Garg, Wireless Communications & Networking, Morgan
Kaufmann, 2007.

© Markus Fidler | IKT LUH | 55/55