ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

ADAPTIVE
ANTENNAS

TYPES OF
BEAMFORMING

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

1- Outlines
This chapter will introduce :
Essential terminologies for beamforming; BF

Demonstrating the function of the complex weights and

how the phase and amplitude weights impact on the
control of the array radiation pattern
Types of BF

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

2- BF terminologies
BF definition
Beam forming is a method used to create the radiation pattern of
an array antenna by adding constructively the weights of the
signals in the direction of SOI and nulling the pattern in the
direction of SNOI (interference)
This array can be antennas in the smart antennas context
, or any other types of sensors (radars, medical sensorsetc)
can be an array of microphones in the speech signal
processing context
Beamforming can be used at both the transmitting and receiving
ends in order to achieve spatial selectivity

i.e., an appropriate feeding allows antenna arrays to steer their beam

and nulls towards certain directions, this is often referred to as
spatial filtering

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

BF network

In transmission mode , the

majority of signal energy
transmitted from a group of
sensor array can be directed
in a chosen angular
direction

TX
BF network

In reception mode, you

can calibrate your group
of sensors when receiving
signals such that you
predominantly receive
from a chosen angular
direction

RX

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Weight vector
Is a vector of complex weights w, each element consists of real and
imaginary components, or alternatively, amplitude and phase components

wn n e j n
wn is the complex weight of the nth element, n is the amplitude weight of the
nth element and n is the phase weight of the nth element
-Amplitude components control the sidelobe level and main beam width
-Phase components control the angle of the main beam and nulls
Phase weights for narrowband arrays are applied by a phase shifter
Steering vector
If there are K transmitters, K received signal vectors can be determined as a
radar tracking multiple targets, or in a mobile communications system where
multiple users are active. The received signal vector of the kth signal is
frequently referred to as the steering vector a(k)
5

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

The array outputs y(k) at any snapshot of time k

The array output can be determined for any snapshot of an incoming signal plus noise
once the weight vector is calculated
-Let us start with a description of the array, the received signal, and the additive noise
-Consider D signals arriving from D directions
-They are received by an array of M elements with M potential weights

W w1 w2 . . wD

-Each received signal xm(k) includes additive

zero mean Gaussian noise. Time is represented
by the kth time sample

s1 (k )
s (k )
2
X (k ) a( 1) a( 2 ) . . a( D ) . n(k )

s D (k )

X (k ) A.S (k ) n(k )
-The array output y can be given in the
following form
T

y (k ) W . X (k )

vector of incident complex monochromatic signals at time

n(k) noise vector at each array element m, zero mean, variance
n2
a(k) M-element array steering vector for the k direction of arrival
A .is an M D matrix of steering vectors
S(k)

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

3- Impacts of complex weight function on the array pattern

Main beam steering
-A simple beamformer steers the main beam in a particular direction (,)
-In case of an environment consisting only of noise, i.e., no interfering signals, this
beamformer provides maximum SNR because of the antenna gain
-The weight vector for steering the main beam is

W ao / M

Example
A receiver equipped with an uniform amplitude four-element linear array of isotropic elements
oriented in y- direction required to enhance the wanted signal arriving from a transmitter
located at 20o from the array boresight in yz plane. Compute the required weight vector for the
array to perform this function given that the elements are spaced apart by d = /2
Solution
The electric field of the received signal is

N 1

For M=4, o

=20o,

=90o

En ( , ) e jnkd sin sin

n 0

Thus, So is computed to be

aTo =[ 1 e j1.0745 e j2.149 e j3.223 ]

Thus, Wo is computed to be

WoT =[ 1 e j1.0745 e j2.149 e j3.223 ]/4

Eo(t)

E1(t) E2(t) E3(t)

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Null steering
-The previous case will not produce a maximum SNR in the presence of
directional interference. Such scenarios are very common in radar
deployments where an intentional jammer may be targeting the area, as well
as mobile communications where other network users will create
unintentional interference
-Thus null steering is useful when it is necessary to attenuate unwanted
signals arriving at angles other than that of the main beam
Let ao
be the main beam steering vector
(K=M-1)
a1 , . . . , aK are k steering vectors for the K nulls

WH= [w1 w2 w3 wM]

1

e jkd sin1 sin1

.
a1

.
j ( M 1) kd sin1 sin1
e

e jkd sin k sin k

.
ak

.
j ( M 1) kd sin k sin k
e

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

-Conditions:
For Ao , the desired signal is received without any modification
WHao = 1

yo w1 w2e jkd sino sino ...... wN 1e j

( M 1) kd sin o sino

For Sk , the interference signals are rejected

WH ak = 0

w1 w2e jkd sin1 sin1 ...... wN 1e j

( M 1) kd sin1 sin1

w1 w2e jkd sin K sinK ...... wN 1e j

( M 1) kd sin K sinK

Which could be written in matrix form as

WH [ao a1 . . . aK]= [1 0 . . . 0 ]=C
WHA = C
WH = CA-1
9

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Example
Compute the weight vector required to steer the main beam of a uniform two isotropic element
array oriented in y-direction towards the required signal in the plane =90o, at 45o from the array
boresight and to minimize the interfering signal that appears at - 10o from the array boresight.
Assume the element spacing d = /2. Such a scenario is typical of a TV receiver situated on a tall
building and is therefore able to receive signals from both a local transmitter and a distant cochannel transmitter. A null is then steered towards the co-channel transmitter to provide
attenuation of this unwanted signal
Solution
N=2, o=45o,k=1=1 = -10o, =90o

N 1

The electric field of the received signal is

En ( , ) e jnkd sin sin

n 0

aTo = [ 1 e j2.221 ]

a1T = [ 1 e -j0.5455 ]

1
1
A j 2.221 j 0.5455
e
e

WH=CA-1

C=[1 0]

Eo(t) E1(t)

WH= [ (0.495

j0.096)

(-0.376- j0.343) ]

W= [ (0.495

-j0.096)

(-0.376+j0.343) ] T

The null steering technique described here jointly steers the main beam and nulls to
the desired angles. Modifying the vector C enables the existence of nulls and beams (or
signal minima and maxima) to be specified according to the prevailing requirements

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

-When the number of required nulls is less than K, K= M 1 , M is the

number of array elements. Matrix A is not a square matrix (the matrix
inversion will be singular and can not be inverted)
-Under such conditions suitable weights may be given by:
For a solution does not minimize the uncorrelated noise at the array output
W H CAH ( AA H ) 1

For a formulation requires noise with variance n be added in the system

W H CAH ( AA H n I ) 1
2

11

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Example
A uniform array of 5 isotropic elements oriented in y direction with spacing d =
/2, the desired signal arriving at = 0 and one interferer arrives at 15 while
the other interferer arrives at +25. If the noise variance is n2= 0.001, find the
weights and plot the array factor considering the reference element is the
middle one
1 e j 2 sin 1 e j 2 sin 2
ao = [ 1 1 1 1 1 ]

j sin 1
j sin 2
1
e
e
-j2sin1
-jsin1
jsin1
jsin1

a1= [e
e
1
e
e
]

A 1
1
1
a2= [e-j2sin2 e-jsin2 1
ejsin2 ejsin2]

j sin 1
j sin 2
1
e
e

C=[1 0 0]
1 e j 2 sin 1 e j 2 sin 2

Using the MATLAB to find the weight vector

0.26 j 0.11
0.17 j 0.08

W
0.13

0
.
17

j
0
.
08

0.26 j 0.11
12

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Onset of grating lobes

So far the analysis of the behaviour of arrays has considered elements with
spacing of d = /2. What happens when other element spacing are used? Such
a scenario occurs when certain physical design constraints are imposed on
the array size which impose to increase the spacing

An array of two-element with unity

amplitude and zero-phase weights,
oriented on y-axis, the radiation
pattern for =90o; azimuth plane is
shown for :
d = 0.5

d = 0.6

d = 2.0

The wider spacing the

narrower the main lobe and
the on setting of additional
lobes (grating lobes)

max sin 1 (n / d ) n 0,1,2,...

13

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

For the case considered, the array factor of the two elements is
1
AFn ( , ) cos( kd sin sin )
2
d
o
Substitute for =90 in azimuth plane
Fn ( , ) cos( sin )
1

d
The maxima occurs when ( sin max ) n max sin (n / d ) n 0,1,2,...

n=0 is for the main lobe, and the others for grating lobes
Example
Compute the angles at which the grating lobes appear in the plane =90o for an array of two
isotropic elements oriented on y-axis and steered for maximum at =60o. consider the element
spacing of d = 0.5, , 0.536 ,0.7, ,2

1
AFn ( , ) cos[( kd )(sin sin sin o sin o )]
2
1
AFn ( , ) cos[( kd )(sin sin o )]
For d/=0.536
2
sin g 0.866 1.866n
AFn ( , ) cos[(d / )(sin sin o )]
For scanning array

(d / )(sin g sin o ) n
(sin g sin o ) n / d
For d/=0.5 no grating lobes

For n=1 g =90o

(from

the ve sign)

For n=2 no grating lobes

For d/=0.536

g =90o
14

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Amplitude weights
The amplitude components of the weight vector control the sidelobe level
and main lobe beam width
An array of four element with
zero-phase and different
amplitude weights as:
[1111]

[0.5 1 1 0.5]

[0.05 1 1 0.05] [0110]

The more tapering along the

array edges the wider the main
lobe and the lower SLL
It can be seem that the radiation pattern of the four-element array approaches that of
the two-element array as the amplitude weights of the outside elements are reduced
15

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Window functions
Window functions enable the amplitude weights to be controlled with
certain constraints. The constraints are related to the characteristics
of the choice of function, such as desired sidelobe level or main
beam width, or a combination of both
(1)Rectangular window
(3)Blackman window

W(n)=1
(2)Bartlett window
For odd N

(4)Triangular window
For odd N

For even N
For even N

16

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

(6)Hanning window

(5)Hamming window

(7)Kaiser window

Io is a modified Bessel function

of order zero
large gives lower sidelobe levels
at the expense of main lobe width
This type of beam steering or BF is suitable for the case of fixed or
switched beam, i.e., it is a conventional non adaptive narrowband BF
17

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Example

Rectangular

Bartlett

Blackman

Triangular

18

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Hamming

=5
=3
=1

Hanning

SLL=-40 dB
SLL=-30 dB
SLL=-20 dB

Kaiser

Chebyshev

19

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4- Types of beamforming
Scanning
RADAR either
tracking or
searching or both

-The beam is to be scanned across the aperture

in which a radar system is searching for new
targets. Once the scanning beamformer has
acquired a target it will be handed to a
conventional beamformer in order to track the
new target. Scanning will then resume

W1A

USER
A

W1B

W2A

2 B

USER
B
W2B

-It is implemented by incrementing the weight vector W,

mechanically or electronically, to steer the angle of the main
beam across the array aperture. It could be either azimuth or
elevation scanning or in both dimensions

-It is not classified as adaptive because they follow a programmed sequence instead of
dynamically adapting to a certain scenario. It is conventional array or conventional BF

20

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4-Types of beamforming
Scanning

Phased array

-It utilizes the DoA information concerning the

desired user or target to steer the beam toward
SOI
-In mobile communication system, the direction
of the mobile is created through weighting
received signal by each antenna element ( in
case of phased array antenna) or through the
reflector antenna system. Then the beam is
launched to communicate with the user
-In this type of beamforming the amplitude of
the weights are held constant while the only
phases are changed
-It is a conventional BF

21

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4-Types of beamforming
Scanning

Phased array

Switched beam forming

Butler
matrix

Finite number of fixed beams

with high sensitivity in
particular directions i.e., select
one of the predetermined fixed
beam pattern

Bless
array

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4-Types of beamforming
Switched approach
The SNR is measured for each beam and the
maximum is determined. The beam associated
with the largest SNR is chosen for further
processing, i.e., at any time , all channels
assigned to the cell that is serving by this array
are available to all users. Therefore a
particular beam , at certain times, may serve
several users
All channels are available to all beams
in the cell
More efficient and can serve up to N
users (N is the number of available
channels) but:
-More complex to implement
-Requires measurement of SNR,
determine the maximum one and an RF
switch should choose the appropriate
beam
-This process should be repeated for
each user

Butler
matrix

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4-Types of beamforming
Sectored approach
It subdivides the sector into many narrow beams
Each beam can be treated as an individual sector
serving an individual user or a group of users
Multiple antenna arrays are implemented
It combines the outputs of multiple antennas in such a
way as to form finely sectorized (directional) beams
with more spatial selectivity than those could be
achieved with conventional single-element approaches
The channels available in the cell are divided
equally between the beams
As a user (mobile, target,) moves through the
sector, the processor detects the signal strength
and switch from one beam to another according to
the strongest one. The beam is switched to
communicate with the user
If the mobile moves into the area covered by a
different beam a handoff must occur

It suffers from greater number of handoffs

Processor

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4-Types of beamforming
Scanning

Phased array

Switched beam forming

Butler
matrix

Bless
array

It is implemented using:
1- A number of fixed, independent, directional antennas
2- Antenna array and analogue beamformer such as Butler matrix (narrowband
BF) or Bless matrix ( wideband BF ). Both matrices are explained in narrowband
and wideband BF
3- Grid Of Beams (GOB) can be used with digital beamforming systems. It
selects the best weights from a stored set of weights. It however, leads to a more
complex implementation due to the drawbacks associated with digital
beamforming. It does not require the beams to be orthogonal and an arbitrary
number of beams can be formed

Phased array is more smarter than switching beam because it generates more
beams which could reach infinity

25

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4- Types of beamforming
Scanning

Phased array

Switched beam forming

Butler
matrix

Adaptive beam forming

Bless
array

An infinite number of patterns (scenario-based) that are adjusted in real time

If the desired signal and interference occupy the same frequency band , then
temporal filtering often cannot be used to separate signals from interference
However, the desired and interfering signals often originate from different angles
This spatial separation can be exploited to separate signals from interference
using a spatial filter at the receiver
A full adaptive beamformer is a processor used in conjunction with an array of
antenna elements, to provide a versatile form of spatial filtering
i.e., adaptive beamforming is able to automatically adapt the radiated power
from an antenna array to different situations and communication scenarios
26

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4- Types of beamforming
Scanning

Phased array

Switched beam forming

Butler
matrix

Bless
array

Adaptive beam forming

Temporal
reference

Spatial
reference

Adaptive
algorithm

DOA
algorithm

The weights for such a spatial filter can be computed in the time (temporal reference
beamforming) or spatial (spatial reference beamforming) domain or without the use
of a reference signal, which is referred to as blind beamforming
-Temporal reference BF:
The weights are computed in time with training sequence (Weiner Hopf equations+
optimization algorithms MMSE, MSINR,)
-Spatial reference BF:
Does not use of embedded training sequences. DoA is used to steer the wanted signal
and nulls directed toward the interference ( needs DoA algorithms, MUSIC, SPIRIT,
27

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

4-Types of beamforming
Scanning

Phased array

Switched beam forming

Butler
matrix

Bless
array

Adaptive beam forming

Temporal
reference

Spatial
reference

Adaptive
algorithm

DOA
algorithm

Blind

Blind BF: weights are computed without reference. It is maximizing the output
and interference suppression( needs DoA algorithms in semi blind BF)

In general, adaptive BF; ABF has different modes,

desired signal maximization mode,
interference signal minimization and cancellation mode
28

ADAPTIVE ANTENNAS.

Switched
scheme
Select the strongest
beam pattern

PROF. A.M.ALLAM

Adaptive
scheme
The beam pattern is adjusted in real time according to:
either a part of the desired signal is known through the
training sequence ( Temporal BF )

or the transmitting DOA is set so that the beam is

maximized toward the desired angle ( Spatial BF )

29

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Comparison between
Switched and Adaptive BF

Adaptive

Switched

High cost
Low cost

Easy to implement
Less

Difficult to distinguish
between the desired signal
and interference
Does not react to the
movement of interference

Integration

Coverage

Interference rejection

Complex
Less hardware redundancy
More

Focusing is narrower
Capable of nulling the
interference signal

ADAPTIVE ANTENNAS.

PROF. A.M.ALLAM

Example
3-5 isotropic elements in
linear array with
different phase

One element
33elements,
phase
elements,
phase0.5
0

31

ADAPTIVE ANTENNAS.

2/21/2015

PROF. A.M.ALLAM

LECTURES

32

ADAPTIVE ANTENNAS.

2/21/2015

PROF. A.M.ALLAM

LECTURES

33