1. Propagation measurements• This chapter discusses common evaluation procedures for

propagation measurements of mobile terminals and the base stations.

– Mobile terminal measurements are carried out as final tests on newly developed hand phones following checks in an indoor measurement environment are also tested in a real environment while testing of base stations at the observation site generates field profiles.

– Base station measurements are the performance test of a newly built base station antenna for basic propagation characteristics during the construction of new mobile communication systems.

• The purpose of the two measurements is different, however identical procedures are used.

1

1.1 Overview of propagation measurement• Propagation measurements can be divided into narrowband and wideband.

• Narrowband measurements use the envelope of a continuous electromagnetic signal with no modulation.

– From the recorded level of the electric field strength both the propagation loss curves for macroscopic observations of the site may be evaluated and also the fading characteristics for microscopic observations.

• Recent mobile communication systems use ever increasing higher-speed data transmission rates. The delay profile of the site limits the maximum data transmission rate of the system.

– A delay profile measurement to be performed for wideband communication systems.– When the delay profile is known in advance, a circuit called an equalizer can be used to cancel

the unwanted signal delay.– A receiver called a rake receiver can be used to sum several delayed signals with the direct

propagation path signal and so increase the effective received signal level.

2

Propagation loss or attenuation• The propagation loss or attenuation between the transmission and reception sites depends on

the characteristics of the environment.

• If the system has a transmission power of Pt, a transmission antenna gain Gt, and a reception antenna gain Gr, then the received power Pr can be obtained by including the propagation loss L as:

(1-1)• When the signal travels through a direct path, of length d, between the transmitter and the

receiver, L is given as the free space propagation loss factor.

(1-2)

• In real environments, very often several propagation paths exist other than the direct one. – Interference between the signals along the various paths causes severe field variation and

multi-path fading. – The measured data includes this fading and often proves to be significantly larger than

the free-space propagation loss. – A propagation loss measurement of the site includes both of these factors.

ttrr PGLGP =

2

4

=

dL

πλ

3

Propagation Loss in Free Space

ttrr PGLGP =

2

4

=

dL

πλ

Pt

d 24 dPp t

d π=

Power density

Transmission power

Pｔ Pr

GrGt

d

Gi : Antenna gaini=t: Transmission, i=r: Reception

24 dPGP t

tr π= rG

πλ4

2

×

Effective receiving area

ttr PGGd

2

4

=

πλ

Propagation loss 4

Propagation loss and fading• Propagation loss is measured within a range from a few hundred meters to tens of thousand of

kilometers, however, the received signal level frequently drops due to the multi-path fading.

• In order for a wireless communication system to overcome this fading phenomenon higher power margins are required to guarantee the system quality.

• In terrestrial mobile communication systems, a fading margin of more than 20 dB is used. A knowledge of the fading characteristics of a site allows a suitable fading margin to be selected for the system design.

• Definitions of basic terminology for narrow band propagation measurements, such as– propagation loss– cross polarization– diversity

5

1.2 Field profile measurement• Measurement parameters of interest in a narrowband system are the propagation loss and the

fading characteristics.

• In order to find the electric field strength over an area of more than a few kilometers from the base station a propagation loss measurement is made.

• On the other hand, if the area of measurement has a radius from only a few meters to a few 10s of meters centered on the base station, then only the fading characteristics are determined.

• In this measurement, the average received electric field strength can be regarded as almost constant.

6

Outdoor propagation environments• Outdoor propagation environments are roughly classified into suburban areas and city areas

from the viewpoint of the fading characteristics.

• There are few tall buildings in a suburban environment and the propagation path consists of a strong direct path signal from the base station plus an array of scattered waves at the observation point. These fading characteristics are categorized as Rician fading.

• Inside a city, the scattering and reflections of the signal from the base station from multiple locations eliminates the direct signal. For this situation almost the same level of multi-path signals from many directions are received at the receiving point. The fading characteristics in this situation are classified as Rayleigh fading.

• To find a typical measurement environment, it is necessary to select the propagation site before field testing.

7

1-2-1 Site selection• The environment characteristics close to the observation point can seriously affect the

measurements. Sites discussed here are classified in terms of fading characteristics, field strength, and cross polarization ratio.

• The simplest way to classify a measurement site is to determine whether or not the transmitting station is directly observable from the observation point.

• If it can be directly observed, it is called a Line Of Sight (LOS) measurement site and, if not, a Non Line Of Sight (NLOS) site. This method of categorization is based on the geometrical theory of optics.

• The dominant fading characteristics for LOS and NLOS sites are called the Rician and Rayleigh distributions, respectively. It can be stated in general that Rayleigh fading can be observed inside a city while Rician fading tends to be found in the suburbs.

8

Long-range propagation measurement

nrp 1

∝ n=2: free space

LOS : Line-Of-SiteNLOS : Non-Line-Of-Site

LOS

NLOS

TX

RX

9

Site classification• When the measurement site is categorized in terms of the field strength, there are three types

of site classification.

• The average field strength level defined later is roughly divided into strong, medium, and weak regimes. The electric field strength discussed here is at the reception antenna output.

• The electric field strength level classification greatly depends on the operating frequency and system design.

– The strong field region has as a mean field strength level of more than 40 dBµV. – The medium field range is then 20 to 40 dBµV. – The weak field range is less than 20 dBµV

for current mobile communication systems operating in the UHF band (300 MHz-3 GHz). The base noise level is classified as being around 0 dBµV.

10

Propagation measurement

Strong

Medium

Weak

Electric field falls down often by a sum of several paths at reception point.

Multi-path fadingRX

Several paths with different phase and amplitude

0

2

46

8

10

0 1 2 3I

V

Pick-up probe I-V characteristics

nCIV =n≈2 (n=2.1)

11

Conversion of dBµV to dBm• The received electric field strength at the antenna input port is measured as a voltage Vµ in

units of dBµV. The unit conversion of received power Pr in dBm, is described below.

• When the impedances of both the antenna and the receiver input ports are matched to Zo, the impedance of the transmission line connecting the antenna and the receiver, Pr is expressed by V (V) as:

(A-1)

• The voltage is given as a unit of dBµV as:(A-2)

• The equation for conversion of V to Pr for Zo=50, is:(A-3)

• and for Zo=75:(A-4)

30log10log2010

14

log10)( 3

2

10 +−=

×= − o

or ZV

ZVdBmP

120log2010

log20)( 610 −== − VVVdBV µµ

01.113)()( −= VdBVdBmPr µµ

77.114)()( −= VdBVdBmPr µµ

12

1.2.2 Attenuation coefficient measurement• When the electric field is measured against the range, r, from the transmitting antenna, it is

found that the amplitude decreases as 1/r for the condition where there are no reflected or diffracted waves present. For practical measurement of the electric field, where the relative power of the field is measured in units of dB, the measured data decreases as 1/r2.

• Fig 1-1 shows an example of the measured electric field in the UHF band within a range from a few meters to a few kilometers. The measured electric field does not follow a 1/r2 decay law but in general a 1/rn (n>2) decay law.

• This approximation is quite rough and the measured data can be quite scattered. This macroscopic data treatment is commonly used for the design of a base station coverage area.

• The attenuation coefficient is a basic evaluation factor for propagation characteristic measurement. The received signal levels expressed in dBµV are plotted as a function of distance from the transmitting station. The measured data are varied over a wide range so the propagation decay profile is found using a statistical procedure.

13

Fig. 1.1 Long-range Propagation Measurement

MeasurementApprox. lineFree Space

Distance form source[m]

Rec

eive

d le

vel[d

Bm

]

1 10 100 1000 10000-120

-90

-60

-30

14

Short range measurement• Another propagation measurement can be carried out near the transmitting base station for the

purpose of indoor propagation measurements or micro/pico cellular system design.

• Fig. 1-2 plots the results of an example of a short range measurement of the received electric field in the vicinity of the transmitting station. As shown in the figure, severe fading can be observed.

• To find the average signal level at the measurement site, a short-term median is used.

• The definition of the “short term” varies, depending on the measurement rage. When the total measurement length is a few hundred meters, the median value of the collected data in a few meters is used as the short term median. However, for measurements of more than tens of kilometers the short term is defined as a few kilometers. The short term is classed as being roughly 1/100 of the total measurement length.

15

Fig. 1.2 Short Range Propagation Measurement

MeasurementShort term median

Distance form source[m]

Rec

eive

d le

vel[d

BV

]m

0 10 20 30 40

20

40

60

80

16

Measurement process• A communications receiver converts the free-space radio frequency (RF) signal to

intermediate frequencies (IF) using an envelope detector. When seeking to find the average received signal level from the measured data (normally recorded in units of dB) it is important to take into consideration the method of calculation used.

• The measured data must first be transformed into linear units before the mean value of the data is sought. A square law detector is used to pick up the electric field but has input-output characteristics that do not always exactly follow the square curve law. If calibration of the detector is carried out for every measurement, the measurement process can be time consuming and cumbersome.

• A median is used as an average value. The median for received data consisting of a total number of N data points may be estimated using the following procedure. The data is sorted by amplitude with the (N/2)th amplitude chosen as the median. The median value eliminates the deep fading in the measurement and gives an average value for the measurement site.

• In addition the cell edge is defined by the level of the median of the down link (transmitter to receiver) at the mobile terminal in practical cellular systems. The critical value is set at between 10 and 20 dBµV, depending on the system design.

17

Fading simulation

d(m)

dB

dB

d(m)

Sinusoidal signals are summed at 2GHz.

Data plotted at 0.1m spacing.

𝑓𝑓 𝑥𝑥 = �𝑛𝑛=1

𝑁𝑁

𝑎𝑎𝑛𝑛 cos2𝜋𝜋𝑥𝑥𝜆𝜆 + 𝜓𝜓𝑛𝑛

𝑁𝑁 = 5, 𝑎𝑎𝑛𝑛 =1, 𝜓𝜓𝑛𝑛 = 𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

𝑁𝑁 = 5, 𝑎𝑎𝑛𝑛 = 𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 , 𝜓𝜓𝑛𝑛 = 0

18

Fading simulation

d(m)

dB

dB

d(m)

Sinusoidal signals are summed at 2GHz.

𝑓𝑓 𝑥𝑥 = �𝑛𝑛=1

𝑁𝑁

𝑎𝑎𝑛𝑛 cos2𝜋𝜋𝑥𝑥𝜆𝜆

+ 𝜓𝜓𝑛𝑛

𝑁𝑁 = 5, 𝑎𝑎𝑛𝑛 = 𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟, 𝜓𝜓𝑛𝑛 = 𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

𝑁𝑁 = 3, 𝑎𝑎𝑛𝑛 = 𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟, 𝜓𝜓𝑛𝑛 = 𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

19

1.2.3 Fading structure• Electromagnetic waves radiated from a transmitting station reach the observation point

through multiple reflections and diffractions and these cause multi-path fading.

• The fading structure is modeled here from a statistical point of view.

• The signals arriving at the observation point are concentrated in a horizontal plane when the transmitting station is located far from the point of observation.

• Fig. 1.3 shows the case where the observation point is moving along the x-axis with a velocity of v meters per second (m/s).

v m/s

Φi

Ai

φi

Fig. 1-3 Mobile Terminal Signal Reception Parameters

Ai, ζi

20

Doppler shift (Hz)

Walking speed 80m/min.

λ(2GHz)

Express train 300km/h

𝑓𝑓𝑟𝑟 =𝑣𝑣𝜆𝜆

λ= 0.32

= 0.15 𝑟𝑟

8060

=43𝑟𝑟/𝑠𝑠 𝑓𝑓𝑟𝑟 =

43

×1

0.15= 8.9 𝐻𝐻𝐻𝐻

𝑓𝑓𝑟𝑟 = 3.3 𝑘𝑘𝐻𝐻𝐻𝐻

21

• For this situation N waves arrive from the φi direction, each with an amplitude Ai and phase ζi. The time-varying electric field Er(t) at the observation point is expressed in terms of the angular frequency, ωc as:

(1-3)

• The last term in equation (1-3) is the frequency shift due to the Doppler effect caused by the movement of the observation point.

• Using standard trigonometric expressions to expand the above equation, we obtain:(1-4)

• When the amplitudes of the individual waves, Ai, are almost the same and the phase ζi changes randomly, the coefficients of and are given by the Gaussian distributions x(t) and y(t).

• The mean values of both x(t) and y(t) are 0, with the same variance, σ, and Er thus becomes:(1-5)

• where x(t) and y(t) are expressed by the following distribution functions:

(1-6)

∑=

++=N

iiicir tvtAtE

1}cos2cos{)( φ

λπζω

∑∑==

−=N

icii

N

iciir tAtAtE

11sinsincoscos)( ωψωψ tv

iii φλ

πζψ cos2+=

ttyttxtE ccr ωω sin)(cos)()( −=

−= 2

2

2exp

21)(

σσπxtx

−= 2

2

2exp

21)(

σσπyty

22

Normal (Gaussian) distribution

),( 2σmN

−

−=≡ 2

22

2)(exp

21)(),(

σσπσ mxxfmN

m: mean value (average), σ2: variance

)1,0(N Standard normal distribution

00.10.20.30.40.5

0 1 2 3 4

x

f(x)

Central limit theoremRandom variables x1, x2, …, xn are independent each other, its average become normal distributions in case of n being sufficiently large.

23

Rayleigh distribution• The Gaussian distributions x(t) and y(t) are replaced by the following probability density function:

(1-7)

• where the amplitude of the total electric field strength is denoted by , and the amplitude probability in the range r to r+dr is obtained by integrating (1-7) over the region 0≤θ≤2π after employing the transformations and . Then,

(1-8)

+−= 2

22

2 2exp

21),(

σπσyxyxp

22 yxr +=

−= 2

2

2 2exp

21)(

σπσrrp

x

y

f(x)

f(y)r φ

δr

−=== ∫ 2

2

2

2

0 2exp)(2)()(

σσπφ

π rrrrprdrprf

PDF of Rayleigh distribution

Probability of r existing in region of r∼r+δr, 0≤φ≤2π is

rrf δ)(

ttyttxtE ccr ωω sin)(cos)()( −= 24

Rician distribution• The Rayleigh distribution results from the summation of independent waves arriving with

almost the same amplitude.

• On the other hand, if the signal is composed of one strong wave r0 and several weak diffracted waves, then the received signal is given by a Gaussian distribution with independent variables X=x+r0 and Y=y and the following probability density function:

(1-9)

• After transformation, p(r) is given by an integral in terms of θ as:

(1-10)

• where I0 is a zero order modified Bessel function. This function is known as the Riciandistribution.

+−−= 2

220

2 2)(exp

21),(

σπσyrXYXp

+−−= 2

002

220

2 2)(exp1)(

σσσrrIyrXrp

r0

25

Nakagami-Rice (Rician) distribution

One strong wave r0 and several weak diffracted waves at reception point.

x

y

f(x)=N(r0,σ2)

f(y)rφ

δr f(y)=N(0,σ2)

r0

r’

+−−= 2

20

20

2 2cos2exp

2),(

σφ

πσφ rrrrrrf

Probability of Nakagami-Rice

+−== ∫ 2

002

220

2

2

0 2exp),()(

σσσφφ

π rrIrrrdrfrfr

PDF of Nakagami-Rice distribution

Zero-order modified Bessel function

r0

0

1

2

3

4

0 1 2 3 4

PDF

r

ro=1ro=1.5

σ=126

1.2.4 Cumulative probability• The preceding section defined the fading structure of a mobile propagation environment using

the probability density function.

• From a practical point of view, it is useful to express the received field strength as a probability function.

• The probability is useful to judge if the received signal strength is less than a critical level or not.

• The probability density function, p(R), of the Rayleigh distribution can be rewritten as follows by using the mean value of the square amplitude

(1-11)

• Then, the probability of a received signal level less than R is given as:

(1-12)

22σν =

−=

νν

2

exp2)( RRRp

−−== ∫ ν

2

0

exp1)()( RdxxpRPR

2ln,2ln,exp21,exp1

21 222

νννν

=−=−

−=

−−= RRRR

27

• In mobile propagation measurements, the median square of the field amplitude strength is used as the average level for short range measurements. The above probability can then be rewritten in terms of the amplitude level Rm for p(Rm)=50% as:

(1-13)

• This probability, normalized by Rm is shown in Fig. 1-4, and is called the cumulative probability distribution.

• For Rician fading, the cumulative probability function is given as,

(1-14)

• ν/2 is the total received power of the weak diffracted waves

• γ is the ratio between the power level of the strong direct wave and ν/2.

• Its fading depth is decreased by a factor γ.

−−=

2

2lnexp1)(mR

RRP 2lnν=mR

−−=

νγ

γνν

RIRRRp

2exp2)( 0

2

+0 +10

10Rayreigh

Rician

-10-20-30-40

10

10

10

10

2

1

0

-1

-2

Nomarized electric field strength (dB)

Cum

mul

ativ

epro

babi

lity

(%)

Fig. 1-4

28

1.2.5 Cross polarization• The electromagnetic wave radiated from the base station propagates in free space and consists

of electric and magnetic field component which are orthogonal to each other.

• The propagation characteristics are mainly described by the electric field component. The polarization describes the direction of the electric field component.

• In the early days, mobile communication systems used vertical polarization in the VHF (30 MHz-300 MHz) band because the horizontal polarization component becomes very weak at a height of 1 to 2 m above the mobile station ground level.

• Although recent terrestrial mobile communication systems in the UHF band (300 MHz-3 GHz) transmit vertically polarized waves from the base station, many obstacles in the propagation path reflect or diffract the transmitted waves and excite a proportion of the horizontally polarized component.

• It is therefore important to measure the cross polarization ratio (XPR) at a mobile terminal in advance of the propagation measurement.

29

• The XPR is defined as the ratio between the vertically and the horizontally polarized components as:

(1-15)

• where Pv and Ph are the short range medians for the vertically and horizontally polarized components respectively at the measured site.

• In the ideal case, the radiation pattern of the measuring antenna should be identical in both planes of measurement (vertical & horizontal). However, in real XPR measurements it is often convenient to use a standard half-wave dipole antenna as the measuring antenna.

• Figure 1-5 shows typical XPR measurement results around a tall building in the city and in the corridor between offices on the 7th floor of the building.

• The XPR is high in the line of sight region and is degraded in the shadow of the building.

• The XPR was found to be around 9 dB in a suburban area and from 3 to 6 dB in the city. The XPR decreased to 0 dB at sites surrounded by tall buildings.

h

v

PPXPR =

30

Cross-polarization (XPR)

0

5

10

15

20

25

30

0 1 2 3 4 5 6Site number

Ele

ctri

c fie

ld s

tren

ght (

dB)

V-polH-polXPR

dBµ

3 to 6 dBin the city

9 dB in suburban

0 dB at sites surroundedby tall buildings

h

v

PP

XPR =

31

Examples of XPR measurements

(a) Outdoor environment

Study room

Study room Study room

Office Office

Study room

Office

Office

Study room Study room

5 45m

20m Beacon

(b) Indoor environment

Site number

LOS

NLOS

32

1.2.6 Delay profile and delay spread• In wireless high-speed digital transmission, the desired symbol data can be corrupted by

interference effects caused by the interaction of the directly transmitted signal with a reflected path signal.

• When the data speed is increased, the previously transmitted symbol data can arrive at the receiving point simultaneously with currently transmitted symbol data.

• The use of diversity reception, adaptive antenna pattern shaping, or implementation of a signal processing system may eliminate these undesired signals.

• To select an appropriate method for designing a communication system, the characteristics of the delayed signal must be measured in advance at several sites.

• For the delayed signal measurement, the base station transmits a pulse signal and the mobile terminal receives the echo.

• Figure 1-6(a) shows an example of an actual delayed signal profile received at a fixed observation point.

33

Fig. 1.6 Received signal echo profile

-90

-80

-70

-60

-50

0 5 10 15 20Delay time (µs)

f(t)

dBm

L0

L1

t0

t1 t2

t3

(a) Measured delay profile (a) Parameters for delay profile

34

• To make it easier to understand the definition of the delay profile the actual received signal shown in Fig. 1.6(a) is modeled by an ideal received signal shown in Fig. 1.6(b).

• The distribution function of the received signal strength, as a function of time t, is denoted by the function f(t). Thus, the received power Pm, which includes the delayed pulse, is given by:

(1-16)• where t0 is the time at which f(t) first crosses the noise level L0 and t3 is the time at which f(t)

next crosses the noise level L0.

• A theoretical definition of t0 can be calculated using the direct line-of-sight (LOS) distance from the transmitter to the receiver.

• However, a definition also including the noise level L0 is useful for practical evaluation purposes. A second evaluation factor TD is defined as the mean of the delay time as:

(1-17)• from which the delay spread or standard deviation of f(t) may be defined as:

(1-18)

∫= 3

0

)(t

tm dttfP

∫ −= 3

0

)()(10

t

tm

D dttfttP

T

∫ −= 3

0

22 )(1 t

t Dm

TdttftP

S

35

Average delay and delay spread

Time delay

Rec

eive

d po

wer

Average delay

Delay spread

∫ −= 3

0

)()(10

t

tm

D dttfttP

T

∫ −= 3

0

22 )(1 t

t Dm

TdttftP

S

∫= 3

0

)(t

tm dttfPTD : Average delay

S : Delay spread or standard deviation of f(t)

Noise level Lo

to t3

f(t)

36

Other delay profile models

Delay time (µs)

Ele

ctri

c fie

ld s

tren

gth

(dB

)

τ1 τ1

{ })()(1)( 221121

ττδττδτ −+−+

= PPPP

f

Delay spread ?

Double Spike

Delay time (µs)

Ele

ctri

c fie

ld s

tren

gth

(dB

)

τ0 τ

)(exp1)( oo

o

o

Uf ττσ

ττσ

τ −

−−= U(x) = 1 : x ≥ 0,

0 : x < 0.

Delay spread ?

Exponential Function

τ2τ1

Question?

37

Multiple Exponential Function Model

Delay time (µs)

Ele

ctri

c fie

ld s

tren

gth

(dB

)

τ0 τ1 τ2

∑∑

−

=−

=

−

−−=

1

01

0

)(exp1)(n

ii

o

in

ii

Up ττσ

ττ

στ

n = 1 : exponential modeln = 2 and σi = 0 : double spike model

This profile can include several delayed large waves by high mountains and at some locations surrounded by tall buildings in a big city.

38

Power Series Function Model

Delay time (µs)

Ele

ctri

c fie

ld s

tren

gth

(dB

)

τ0 τ

)(1)( 1 oUpo

ττττατ α

ε −−

= −+−

micro-cellular, a coverage of 100 m to 500 m

The base station in this system is assumed to be as high as surrounding buildings and the constant α is 3 or 4.

Average power Delay spread021 τ

α −

2

0 21

31

−−

−

−−

αα

αατ

39

Measured delay spread

BER versus delay spread

Aver

age

BER

Normalized code period

10-1

10-2

10-3

10-4

10-2 10-1 10

MSKOQPSKQPSK

BPSK

S=0.085[µs]

S=1.348[µs]

NLOS

LOS

40

Reduction of delay spread by directional antenna

Dela

y sp

read

[nse

c]

Beam width (deg.)

Delay time [nsec]

(b)TX/RX:30[Deg]

[dB]

Delay time [nsec]

(a)TX/RX:omni

[dB]

Delay spread

Simulation model

41

6-sector antenna using PCTSA

Monopole

PCTSA

Ground Plane GapParasitic element

FRP

20mm

25mm

75mm

6mm

35mm

MonopoleGround Plane

PCTSA

Monopole

Parasitic element

Ground Plane

PCTSA

42

Antenna characteristics of PCTSA

90°-90°

0°

180°

-10-20-30

[dB]

90°-90°

0°

180°

-10-20-30

[dB]

f=8.45GHz

-30

-20

-10

0

-180 -120 -60 0 60 120 180Degree

7.0GHz 8.0GHz 9.0GHz

Horizontal plane Vertical plane

Radius=250mm, height=60mm43

Location of base station antenna

AB

C

F

6Fα β

Top

7F

7F

AB

C

D

EF

G

J

H Iα β

BS α BS β

44

Measured delay profile[d

B]

Delay Time [ s]µ0 2 4 6 8-200

-180

-160

-140

-120

[dB

]

Delay Time [ s]µ0 2 4 6 8-200

-180

-160

-140

-120

[dB

]

Delay Time [ s]µ0 2 4 6 8-200

-180

-160

-140

-120

[dB

]

Delay Time [ s]µ0 2 4 6 8-200

-180

-160

-140

-120

[dB

]

Delay Time [ s]µ0 2 4 6 8-200

-180

-160

-140

-120

[dB

]

Delay Time [ s]µ0 2 4 6 8-200

-180

-160

-140

-120

S=0.14[µs]

S=0.495[µs]

S=0.202[µs]

S=0.03[µs]

S=0.334[µs]

S=0.09[µs]

sector1

sector6

sector5 sector4

sector3

sector2

BS

Mobile terminal

45

1.3 Diversity measurement• Multi-path fading at the mobile receiver varies with the sum of the electric field components

incident on the receiving antenna from various directions and can range from about 20 dB to 40 dB.

• One technique to reduce the influence of deep fading is diversity reception, using several simultaneously received independent signals and synthesizes these to compensate for the fading.

• The diversity scheme often is employed as MIMO antennas. In this section, diversity schemes are discussed as well as suitable measurement techniques.

• Diversity reception is used for mobile terminals and base stations to help compensate for very weak signal reception caused by multi-path fading. The same increase in communication performance is also expected for the base station.

• The downlink signal from the base station to the mobile terminal is stronger than the up link from the mobile terminal to the base station. The base station transmits tens of watts, while the mobile terminal radiated power can be less than 1 W in current cellular systems.

• The main purpose of using diversity reception at the base station is therefore to compensate for the difference in signal levels between the uplink and the downlink. 46

1.3.1 Diversity scheme• Diversity schemes may be categorized into 3 major types: antenna (or space), frequency and time.

The latter two methods depend on the type of communication system while the antenna diversity method is independent of the system.

• Antenna diversity techniques use several methods to help reduce fading. One method is to use several independent signals from different antennas spaced apart. Another method is termed polarization diversity and takes advantage of the principal- and cross-polarization components of the received signal.

• Another technique called antenna pattern diversity utilizes the signals from antennas with non-overlapping antenna radiation patterns.

• Antenna pattern diversity may also be called angular diversity and has been used for overseas wireless communications in the HF band (3 MHz-30 MHz) to eliminate the fading caused by time variations from signals reflected from the ionosphere.

47

Various diversity reception schemes

1

2

(a)

12

(b) (c)

1

2

f1

f2

1

2

(d)

τ

1

2

(e)

(a) space diversity, (b) polarization diversity, (c) antenna pattern diversity,(d) frequency diversity, (e) time diversity

τ

48

• Space diversity systems are widely used at the cellular base station antenna. However, the antenna gain of the base station is approximately 10 dBi or more and this requires antennas to be spaced by more than 10 to 20 wavelengths apart in order to decrease the mutual interference between the two antennas.

• Recently, the drastic increase in the number of subscribers has forced wireless service providers to install an even larger number of base station antennas. Polarization diversity schemes using an inclined polarization scheme of ±45 degrees have been adopted in Europe and America to obtain a stable mobile up link communications.

• In Japan, a polarization scheme utilizing the horizontal and vertical polarization components has been used because the hand held mobile phone terminal is often used at a tilted angle.

• In addition to diversity antennas, the, so-called, sector antenna is widely used for base stations. This antenna divides the coverage area into 3 or more sectors. In high-speed digital data transmission the sector antenna helps to eliminate the delayed signal due to multiple reflections and may be used for applications such as wireless LANs or mobile communication systems.

• The main purpose of using diversity reception at the base station is therefore to compensate for the difference in signal levels between the uplink and the downlink.

49

Car mounted diversity antennaSpace diversity

Antenna pattern diversity

50

Flat diversity car mounted antenna

51

Diversity sector antenna arrangement

Space diversity antenna should have more than 10λ spacing.

52

V-pol.

Redome

Sector 1

Sector 2, 3

H-pol.

VH polarization diversity base station antenna

53

Dielectric substrate

Feeding MS line(top side)

Parasitic element(top side)

Twin-dipole (back side)

Feeding probe

Printed twin-dipole antenna

54

1.3.2 Correlation histogram & coefficient• The performance of diversity antennas is evaluated by measuring the amplitude and phase

components of the radiation pattern, or by testing the antenna in a practical environment. These radiation patterns will hereafter be referred to as complex radiation patterns.

• In both cases, the evaluation factor is the correlation coefficient between different diversity branches of the antenna. Generally, the term diversity branch denotes the multiple output from the antenna after the modulated signal is detected, but it refers to the different antenna output ports.

• When there are two output branches for a particular diversity antenna, the correlation coefficient of the detected signal envelope of e1 and e2 is defined as:

(1-19)

• where <x> denotes the mean value of the measured parameter x, and (y)* denotes the complex conjugate of the measured parameter (y).

( ) ( )

( ) ( )222

211

22*

11

21

21

eeee

eeee

−−

−−=ρ

55

• In propagation measurements the correlation coefficient is calculated every 100 or 200 samples, and a correlation histogram is shown in Figure 1-8.

• In practical diversity measurements, more than tens of thousands of data samples are collected in one measurement. Thus, for a single measurement, hundreds of correlation coefficients must be calculated.

• Due to the large number of data samples required to be collected, it is best if the correlation coefficient is calculated in the region where the average received electric field level is almost constant, thus reducing the total number of calculations. The correlation coefficient is used for the macroscopic evaluation of diversity antennas.

0

5

10

15

20

25

30

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

The

num

ber o

f tim

es

Correlation Coefficient

Diversityscheme

e1 e2

Diversity branchei : signal envelope

56

1.3.3 Diversity gain• The effectiveness of diversity reception is evaluated by the correlation coefficient presented in

the previous section. Another evaluation factor is the diversity gain. Using diversity reception in a multi-path propagation environment reduces the fading depth of the received signal.

• In a propagation environment exhibiting Rayleigh distribution fading, the probability P of the received signal to noise ratio (SNR) less than γ is given for n branches of the diversity antenna as denoting the average signal to noise ratio as γ /Γ.

(1-20)

• where selection diversity is assumed after detection of the received signals and the correlation coefficient between branches is zero. In selection diversity, selecting the branch with highest power combines the data sequences representing the envelope power received by each branch.

• In Fig. 1-9 the diversity gain is defined by the difference of the reception level at P(γ) = 1 % between the curve n = 1 and N.

Γ−=

n

P γγ exp1)(

57

Fig. 1.9 Cumulative Probability Distribution

1

1

1

1

1 2

- 4 0 - 3 5 - 3 0 - 2 5 - 2 0 - 1 5 - 1 0 - 5 0 5 1 0 0

- 2

0 - 1

0 0

0 1

0

Rayleigh Antenna 1 Antenna 2 Diversity

Normalized electric field (dB)

Diversity gain

Cum

mul

ativ

e pr

obab

ility

(%)

Diversityantenna

e1

e2

eo =max(e1, e2)

Selection diversity

Depth of fading

eo

Looking at CPD of 1%, fading depth is decreased by several dB after selection diversity.

This value is called “diversity gain”.

58

• Diversity gain indicates the degree by which diversity reception reduces the fading depth.

• In order to evaluate the selection process statistically from the measured propagation data, the envelope of the received signal for each diversity branch is combined by selecting the branch with the highest level of max(Ei

1, Ei2).

• In this case, Ei1 and Ei

2 represent the received signal levels aft r the process of envelope detection by branches 1 and 2 of the diversity antenna and i represents the discrete time when the sample was taken.

1.4 Definition of the Diversity Gain

59

1.5 Propagation Measurement System• Measurement systems may be classified into four types.

• The most basic system measures the propagation loss factor for both short and long ranges to find the site attenuation coefficient.

• A second type of measurement system is used to investigate the fading structure and diversity performance of the antenna used in narrowband mobile communication systems.

• A third type of measurement system finds the delay spread to evaluate of the wide band characteristics of the system.

• Lastly, for maintenance and field checking of cellular systems currently in service, fully automated field measurement systems are required.

60

1.4.1.1 Long range & short range propagation loss measurement

• A standard measurement system in Fig. 1-10 consists of a signal generator connected to a low gain antenna and a receiver with a standard antenna which is a half wavelength dipole antenna with a known radiation pattern and antenna gain.

• The level of the envelope signal at the receiver output after detection is sampled using an analog-to-digital converter (ADC). For long range measurements, the sampling rate of the ADC is chosen to sample the data at less than a quarter wavelength spacing. This is because the period of the standing wave created between the transmitting and receiving antenna has a period of one half wavelength, as is shown in Figure 1-11.

• Thus, the sampling frequency fs is determined by the relation fs = 4v/λ Hz, where the receiving system is moving at a rate of v m/s. In order to record the fading structure in detail without using any interpolation techniques, the sampling spacing should be less than λ/20 for short range measurements .

• Finely spaced sampling is required to obtain more than a -30 dB fading depth accuracy in the measured Rayleigh distribution data.

61

Fig. 1.10 A standard measurement system

RECEIVER PERSONALCOMPUTER

A/DCONVERTER

DOWNCONVERTER

ANTENNA

ENVELOPEDETEVTOR

SIGNALGENERATOR

ANTENNA

(a) Transmitting system (b) Receiving system

λ/2

Sampling interval

λ/2

Fig: 1.11 Receive signal standing wave and the sampling interval

62

1.4.1.2 Diversity performance measurement• The only difference between the previous measurement system and that used for diversity

performance evaluation is the multi-channel receiver/ADC shown in Fig. 1-12.

• For a two-branch diversity system, two identical receiving subsystems are required for simultaneous measurement of the received signals from the two antennas. A portable spectrum analyzer can be used as a receiver for the measurement system and a commercially available wide band receiver can also be used for the same purpose.

• The two independent receiver subsystems should also be calibrated using a known power level standard as an input to each channel. There can often be a discrepancy of 1 to 2 dB between the two receivers which has to be calibrated.

• When a mobile terminal moves at a speed of 80 km/h, the required sampling rate is 0.68 ms to satisfy the λ/20 sampling interval at 1 GHz. A multi-channel ADC usually has a sampling rate of less than a few ms and therefore does not affect the overall sampling rate of the system.

• To simplify the measurement system, one receiver can be used with multiple antennas. For this measurement, the technique of switching between different antennas at a specified switching rate is used.

63

Fig. 1.12 Multichannel receiver system

PERSONALCOMPUTER

(b) Receiving system

RECEIVER ENVELOPEDETECTOR

MULTI-CHANNEL

ADCONVERTERRECEIVER ENVELOPE

DETECTOR

64

1.4.1.3 Delay profile measurement• The delay profile measurement system is shown in Fig. 1-13. In order to perform this type of

measurement a precise timing signal is essential for both the transmitter and the receiver.

• If the receiving points are located close to the transmitter, a common signal generator source can supply the timing signal to both the transmitter and the receiver, using coaxial cable.

• For these measurements, however, the use of cables is not feasible. A rubidium oscillator in a standard signal generator synchronizes the oscillators used in the transmitter and the receiver.

• When performing a delay profile measurement indoors, the transmitting and receiving antennas can both be connected to a vector network analyzer to obtain the received signal time domain characteristics. After performing the measurement over a wide frequency range, the transmission coefficient (S21) frequency data is transformed from the frequency domain into the time domain using Fourier transformations.

• This is the simplest method for obtaining the time domain delay profile. However, it should be noted that the measurement resolution is dependent upon the frequency bandwidth of the antennas used.

65

Fig. 1.13 Transmitting system and receiving system

Rubidium oscillator

TransmittingAntenna

Up converter

RF amplifier

Base band modulator

Rubidium oscillator

Receivingantenna

ADAD converter GPSGPS

Down converter

Demodulator

Oscilloscope

Data recorder

(a) Transmitting system (b) Receiving system

Rubidium oscillator synchronizes the oscillators in TX and RX.

66

Rubidium

Frequency stability 3×10-11 /s

5×10-11 /month

Crystal 10-6 ppm (parts per million)

Rubidium 10-9 ppb (parts per billion)

Cesium 10-12 ppt (parts per trillion)

24 hours or 1 year

Example

1GHz 1×109×10-6 ×365=365×103 = 365 kHz

67

1.4.1.4 Base station maintenance and testing• A propagation measurement is also required when checking the performance of the base

station antenna and the corresponding coverage area.

• This measurement is made using a receiver placed on a vehicle to reduce the time required to perform a measurement over a wide coverage area.

• Fig. 1-14 shows a typical on-board receiver configuration used to measure the received signal strength. The system consists of the following sub-systems: a location detector, a data processor and an electric field strength measuring device.

• The location detection sub-system is used to estimate the precise position of the vehicle and consists of a directional sensor, a distance sensor, route history data and a road map stored in a CD-ROM to help generate the measurement route history.

• Multiple measuring receivers are also placed on-board to allow the simultaneous measurement of multiple radio channels.

68

Fig. 1.14 Field strength measurement system

GPS Reciever

Controller

Gyro sensor

Directionsensor

RPM sensor

Display

CDROM

PC Data storage

Divider

MeasuringRX

MeasuringRX

Mullti-channelmeasurement

system

69

1-4-2 Calibration• For precise propagation measurements the calibration procedure is very important. For this

type of measurement two types of calibration procedure are generally used.

• The first involves the linearity compensation of the receiver in such systems used in short range measurement. The second type of calibration procedure is that of time standard calibration and is generally used in the delay measurement system.

• The calibration of the RF system is carried out by connecting the signal generator output directly to the antenna input port of the RF receiver. Fig. 1-15 shows a calibration chart of the detector system output power versus the detected signal level. This chart may be used to help compensate for the effects of the detector non-linearity in the measured received signal data. This detector system calibration is best carried out in advance of the actual measurement.

• In order to synchronize the phase of the signal between the transmission and reception systems, the two rubidium oscillators are synchronized before the actual measurement. However, this synchronization may last for a period of only a few days and for a precise measurement the synchronization procedure must be repeated every two to three hours.

70

Fig. 1.15 Detector calibration curve

0 20 40 60 80 100 1200

1

2

3

71

1-4-3 Antenna switching measurement system• The conventional propagation measurement system uses a diversity antenna in conjunction

with the corresponding number of calibrated receivers to detect several input signals simultaneously.

• A simple calibration-free measurement procedure using a single receiver with an antenna switching unit, since it forms the basis for all propagation measurement systems.

• Fig. 1-16 shows the block diagram of the measurement system. An antenna switch consisting of 3 GaAs chips with dummy loads selects the received signals alternatively from 2 antennas.

• When the switch selects one antenna the other antenna is terminated by a 50 ohm dummy load. This provides an isolation of 60 dB and an insertion loss of 1.3 dB below 1 GHz. This type of switch is inexpensive compared with a pin diode switch and is ideal for low cost measurement systems.

• The received signal is amplified by 15.1 dB and is then converted to an intermediate frequency (IF) signal (at 10.7 MHz) by a receiver IC chip (IC-R9000). The IF signal is then detected by a standard envelope detector of a type used in FM IF transceivers and portable phones.

72

Fig.1.16 Antenna switching measurement system

Antenna 1

Antenna 2

GaAsSwitch AMP

FreqeuncySynthesizer

ReceiverIC-R9000

fo=10.7MHzBW=230kHz

AMP

LocalOscillator

10.245MHz

Envelope detector

ClockGenerator

SginalGenerator

ADConverter

PersonalComputer

SignalClock

fo=455kHzBW=30kHz

The switch provides an isolation of 60 dB and an insertion loss of 1.3 dB below 1 GHz.

73

• To obtain a rapid response, it is necessary to use a wide bandwidth filter. The bandwidths selected for this particular system are 230 kHz for the first filter and 30 kHz for the second.

• The detected envelope signal level is recorded in a digital data format using a 2- channel A/D converter. The measurement system is placed on a portable trolley for convenience.

• To characterize diversity antennas, several antenna signals are received simultaneously. When the sampling interval is small enough to neglect the time deviation between samples, they may be considered to be sampled simultaneously. The maximum sampling rate gives an estimate of the short range measurement parameters.

• An interval of λ/40 is sufficiently short that the received signal envelope can be considered constant between measurements. This result is given by switching two output ports of a space diversity antenna consisting of two λ/4 monopoles with a 3/4 wavelength spacing.

• Its switching rate of 66 Hz is short enough to apply to a stationary environment, because the velocity of people moving in the building is less than 1 m/s. For this particular diversity antenna system, with a sampling interval of λ/40 and the antenna switching rate indicated above means that the sampling rate for a single antenna is λ /20.

• This is the reason for choosing the envelope signal level as described in short range measurement without the use of interpolation.

74

Data Sampling Rate

Sampling interval

How to determine d?

d

257mm

86mm Ground plane

Monopole antenna #1 #2

GaAs Switch

#1

output

GaAs Switch

#1

output

#2

Type A Type B

T-junction

≅ 1λ @ 900MHz

75/50

Cumulative Probability and Histogram

When the switching rate is fast enough to neglect time deviation between adjacentsamples, correlation is 1.0 for type A and 0.1 for type B.The sampling interval is λ/200, switching frequency is 330Hz and 25,000 data.

76/50

Histogram of Correlation Coefficient

d is λ/40 d is λ/20

The interval of λ/40 corresponds to 66 Hz.

77/50

1.4.4 Tips in propagation measurement• A problem in propagation measurement is how to remove the electrical effect of supporting

structure of handset antennas and feeding cables. The tested handsets are usually held by low dielectric constant materials such as polyethylene foam, wood and corrugated fiberboard.

• In the measurements, we need receivers and data recording devices, which are made by metal frame. To minimize the pattern distortions by electrical obstacles, the equipment should be placed away from tested devices. In the experience in early days’ propagation measurements at 900MHz band, tested devices are at least 1m away from the conducting materials. It is 3λin free space.

• A theoretical background is given by a concept of antenna clearance. An antenna needs some space to be operated in stable. It is called as “antenna clearance” and is calculated by placing an electrical obstacle in the vicinity of tested antenna.

• A simple obstacle is a λ/2 conducting wire to disturb antenna input characteristics and radiation patterns. The input impedance is changed seriously by this wire when the wire distance is less than λ, while the radiation patterns’ deviation is not converged for this spacing.

78

1.4.4 Tips in propagation measurement• For example, an H-plane radiation pattern of λ/2 dipole antenna is seriously affected by the

presence of λ/2 conducting wire arrayed in parallel.

• A ripple of H-plane radiation pattern is ±2.4dB for the distance d=λ between dipole and wire, and is ±0.8dB for d=3λ. The pattern deviation less than ±1dB is guaranteed for the tested devices at least 3λ away from the electrical obstacles.

79

d

d=0.5λ

d=0.75λ

d=1.0λ

d=2.0λ 80

d=3.0λ d=4.0λ d=5.0λ

0

2

4

6

8

10

0 1 2 3 4 5

δG(dB)

d/λ81

References1. Jakes, W. C. Jr., Microwave Mobile Communications, John Wiley & Sons, New York, 1974.2. Clarke, R. H., “A Statistical Theory of Mobile-radio Reception,” Bell Syst. Tech., J., Vol. 47, No. 6, Jul.-Aug., 1968, pp.957-1000.3. Hata, M. “Empirical Formula for Propagation Loss in Land Mobile Radio Services,” IEEE Trans. Vehicular. Technology., Vol. VT-29, no. 3,

1980, pp. 317-325.4. Taga, T., “Analysis for Mean Effective Gain of Mobile Antennas in Land Mobile Radio Environments,” IEEE Trans. Vehicular Technology,

Vol. VT-39, No. 2, May 1990, pp. 117-131. 5. Lee, W. C. Y., Mobile Communication Engineering, John Wiley & Sons, New York, 1982.6. Nakano M., et al., "Up-Link Polarization Diversity Measurement for Cellular Communication Systems Using Hand-Held Terminal" Conf. Dig.,

IEEE AP-S, Int, Symp, Montreal, Canada, Jul. 13-18, 1997, pp.1360-1363.7. Murase, M., K. Tanaka, and H. Arai, "Propagation and Antenna Measurements Using Antenna Switching and Random Field Measurements,"

IEEE Trans. Veh. Technol., vol.43, no.3. 1994, pp.537-541.8. Komizu, T., et al., "Development and Field Test Results of a Digital Cellular System in Japan" Proc. of IEEE Vehicular Tech. Conf., Stockholm,

Sweden, 1994, pp. 302-305.9. Arai, H. “Antenna Clearance, an Index Factor for Electrical Size of Small Antennas,” Conf. Dig., iWAT2008, Chiba University, Japan, IT45,

March 4-6, 2008.

82