Beamforming Antennas for

Wireless Communications

Yikun Huang, Ph.D.

ECE/CCB
Yikun@cns.montana.edu
November 24 2003
 Outline
Introduction
Beamforming and its applications
Beamforming antennas vs. omnidirectional antennas

Phased Array Antennas

Direction of arrival (DOA) estimation
Beamforming
Basic configurations: fixed array and adaptive array
smart antenna systems:switched array and adaptive array

Vector Antennas
DOA and polarization
super CART
3-loop and 2-loop vector antenna array
Direction of arrival (DOA) estimation
Vector antenna vs. phased array antenna

Beamforming antennas for WLAN

Infrastructure mode
An indoor WLAN design
Ad hoc mode
Ad hoc WLAN for rural area

Conclusion
 Applications of beamforming technology

Applications Description

RADAR Phased array RADAR; air traffic control; synthetic

aperture RADAR

SONAR Source location and classification

Communications Smart antenna systems; Directional transmission and

reception; sector broadcast in satellite communications

Imaging Ultrasonic; optical; tomographic

Geophysical Exploration Earth crust mapping; oil exploration

Astrophysical Exploration High resolution imaging of universe

Biomedical Neuronal spike discrimination; fetal heart monitoring;

tissue hyperthermia; hearing aids

Source: B.D.Van Veen and K.M. Buckley, University of Michigan, “Beamforming: A

Versatile approach to spatial filtering”,1988
Phased array RADAR
 Phased array spike sorting

1
6
0.148

1
5
Rn( 15  t )

 0.534

Phased array spike sorting system

0 t 4
1.2 10

1
4
0.139

Rn( 13  t )

1
3
 0.534
0 t 4
1.210

1
2
0.183

Rn( 11  t ) 0.042

1 Sorted
1
Neuronal
Ey3n ( t )
 0.539
0 t 4
1.210  0.187
0 t 4
1.2 10

spikes Spike of
1
0
0.147

individual
Rn( 9  t )

recorded by
9

0.056

 0.534
0 t 4 Ey2n ( t )
1.210

electrode neurons.
 0.205
0 t 4
1.2 10
8

0.147

array Rn( 7  t )
7

 0.534 0.139
0 t 4
1.210
Ey1n ( t )

 0.544
6

0 t 1.2 10
4

0.183

Rn( 5  t )

 0.539
5

0 t 4
1.210
4

0.139

Rn( 3  t )

 0.534
3

0 t 4
1.210

0.14
2

Rn ( 1  t )

 0.534
1

0 t 4
1.210

Center for Computational Biology, MSU

 Patterns, beamwidth & Gain
top view(horizontal)

side lobes
Main lobe

φ1/ 2
nulls

Half-power
beam width

Half-power Half-power
beam width beam width
side view(vertical)

78° θ1/ 2

Isotropic dipole half-wave dipole beamformer

 Beamformers vs. omnidirectional antennas

1) Beamformers have much higher Gain than omnidirectional antennas:

Increase coverage and reduce number of antennas!
GN
Gain:  N2
G1
90
6
6
120 60

150 30

Field( 6  0  )

Field( 2  0  ) 7
180 0 0 9.96110
Field( 1  0  )

210 330

240 300

270

 Beamformers vs. omnidirectional antennas

2) Beamformers can reject interference while omnidirectional

antennas can’t: Improve SNR and system capacity!

interference null interference

user user

3) Beamformers directionally send down link information to the

users while omnidirectional antennas can’t: save energy!
 Beamformers vs. omnidirectional antennas

4) Beamformers provide N-fold diversity Gain of omnidirectional antennas:

increase system capacity(SDMA)

5) Beamformers suppress delay spread:improve signal quality

null

user user
multipath
 DOA estimation
Plane wave

……

φk

1 2 3 4 5 6 7
…… N-3 N-2 N-1 N

d
δ k  d sin φk phase delay

2πd
Δk  sin φk  β  kd sin φk  β
λ
 Beamforming

……

φk

1 2 3 4 5 6 7
…… N-3 N-2 N-1 N

1,,k 2,,k 3,,k 4,,k 5,,k 6,,k 7,,k N-3,,k N-2,,k N-1,,k N,,k phase shifters

……

Δ N ,k  (N  1)(kd sin φk  β )
 Basic phased array configurations

sN(k) Z-1 Z-1

sN(k) w*N w*N,0 w*N,1 w*N,k-1

.
.
.
.
. .
y (k ) .
. s2(k)
s2(k) w*2  Z-1 . Z-1
y (k )
w*2,0 w*2,1 w*2,k-1

.
.
.

s1(k) w*1
s1(k) Z-1 Z-1

w*1,0 w*1,1 w*1,k-1

.
.
.
Narrowband broadband

phased array (fixed/adaptive) configurations-time domain

Basic phased array configurations

F I y (k )
F
sN(k) F w*N 

…
T F
. . T
. . -
. . +
F d (t ) MSE
s2(k) F …
w*2
T
F
F
T
F
F w*1
s1(k)
…

broadband

phased array (fixed/adaptive) configuration-frequency domain

 Smart antenna systems

Cellular Wireless
Military
communication local area
networks
networks networks
switched array switched array switched array
adaptive array adaptive array adaptive array

3G Data rate:100kbps Wi-Fi Data rate:11Mbps

 Smart antenna systems
top view(horizontal)

5 4
6 3
7 2
interference
8 1

9 16 user

10 15

11 14
12 13

Switched array (predetermined)

 Smart antenna systems

top view(horizontal)
Interference 1

user 1

user 2
Interference 2

Adaptive array
 Smart antenna system
Example: Vivato 2.4 GHz indoor & outdoor Wi-Fi Switches
(EIRP=44dBm;Gain=25 dBi;3-beam)

  11 Mbps: up to 300m
In door range 5.5 Mbps: up to 400m
(Mixed Office) 2 Mbps: up to 500m
1 Mbps: up to 600m
  11 Mbps: up to 1.00km
Out door range 5.5 Mbps: up to 1.25km
100
(outdoor to indoor) 2 Mbps: up to 2.00km
1 Mbps: up to 2.50km
11 Mbps: up to 4.20km 12
Out door range 5.5 Mbps: up to 5.10km
(outdoor to outdoor) 2 Mbps: up to 6.00km
1 Mbps: up to 7.20km
Active user per switch 100

www.vivato.net
 Polarization
circular ellipse linear
Ei

E i sin γe jη Z E E E
i
E cos γ

Y
   
E E E

’ X

=45 =0
=90
Super CART

SuperCART
Compact array radiolocation technology
Flam&Russell,Inc.,1990
U.S. Patent No., 5,300,885;1994
Frequency range: 2 – 30 MHz
 3-loop
V6
Y

V4
V0e   I (0) Z L
Ve   I ( ) Z L  X

V1 V2
b

V3
i
V5 I   zˆ  H 0

i
kb0.5 I   yˆ  E0
 2-loop
Blind point

E Steering vector
H
 e y  sin Φ0 cos Θ cos Φ0 
   
S e   sin Θ  sin γe 
jη
0
a40   z     
 cos Φ0 cos Θ cos γ 
 h x    sin Φ 0 
h  0 sin Θ 
 z 

Ei0
H 
i
0
ζ
ex2  ey2  ez2  1

hx2  hy2  hz2  1

Vector antennas vs. spatial array antennas

Vector antennas measure: ,,,, and power simultaneously,

no phase shift device, or synchronization is needed.

Phased array antennas with omnidirectional element measure:

,, and power
 Vector antennas vs. spatial array antennas

VA SA VA

SA

Source: Nehorai,A.,University of Illinois at Chicago

Vector antennas vs. spatial array antennas

Vector antenna: no ambiguities for DOA estimation

e x , e y , e z , h x , h y , hz  φ, θ,γ , η, P

Phased array antennas: spatial ambiguities exist

φk φ1

φ2

1 2 ……
3 4 5 6 7 1 2 ……
3 4 5 6 7

φk f1 sin φ1  f2 sin φ2
 Vector antennas Vs. phased array antennas

Disadvantages of vector antennas

Low profile?

f=2.4GHz,  =0.125m; vector antenna size: 0.0125m ~ 0.063m

Phased array:d /2=0.063m;L=(N-1)d: 0.188m-0.69m(N=4…12)

f=800MHz,  =0.375m; antenna size: 0.04m ~ 0.19m

Phased array:d /2=0.19m;L=(N-1)d: 0.56m-2.06m(N=4…12)

Cheap?

Can use hardware and software of existing communication

systems for performance?
 Working in scattering environment

source:M.R. Andrews et al., Nature, Vol. 409(6818), 18 Jan. 2001, pp 316-318.

Low profile antennas with polarization diversity
(a) 2-dipole(monopole)

(b) 2-loop

(c) dipole-loop
 Packet switching

AP1 AP2
A

user

Handoff between Aps

was not standardized
at the same time as
802.11b

TDD/TDMA
 Packet switching: 3 beam system
top view(horizontal)
Pi 1

Pi 1  Pi 1
d
Pi
Δφ
Pi
Δφ
φmax
i
 1 / d  2   ( Δφ / 2), d  1
 i
φˆ DOA  φmax  d  ( Δφ / 2), d 1
 i
φmax  1 / d  2   ( Δφ / 2), d  1

Pi 1

P. Sanchis, et al. 02
 An indoor WLAN design
A 4-story office building (including basement), high 30 m, wide 60m and long 100m. We
plan to install a Vivato switched array on the 3rd floor.

Switched array

2
h=30m
1

Basement
w=60m

L=100m
 An indoor WLAN design

Data rate 1Mbps, 2Mbps, 5.5Mbps, 11Mbps

AP’s EIEP 44dBm
AP’s antenna Gain GA 25 dBi
PC antenna Gain GP 0 dBi
Shadowing 8dB
AP’s antenna receiving sensitivity S min -95dBm ,-92dBm, ,-89dBm, -86dBm
AP’s Noise floor -178dBm/Hz
Body/orientation loss 2dB
Soft partition attenuate factor (p= number) p1.39 dB

Concrete-wall attenuate factor(q= number) q2.38 dB

Average floor attenuation(floor number) 14.0dB(1),19.0dB(2),23.0dB(3),26.0dB(4)
Frequency 2.4GHz
Reference pathloss PL0 (LOS/NLS, r=1m) 45.9dB/ 50.3dB
Pathloss exponent  (LOS/NLS, r=1m) 2.1/3.0
Pathloss standard deviation  (LOS/NLS) 2.3dB/4.1dB

Average floor attenuation(floor number) 14.0dB(1),19.0dB(2),23.0dB(3),26.0dB(4)

Data of AP’s antenna is from www.vivato.net

 An indoor WLAN design
Mean pathloss with smin: L  EIRP  Smin  GP

Allowable pathloss: PLallowable  L  Lw  Lsm  Lfl  Lsd  Lo

Path loss model: PL(r )  PL0  10γ log( r )

r0

PL(r )  PLal

Case 1: user is on the 3rd floor: 3 concrete walls, 3 soft partitions

The coverage ranges are: r=176m,140m,111m and 88m for date rate at 1Mbps,
2Mbps, 5.5Mbps and 11Mbps respectively .

Case 2: user is in the basement : 3 floors; 2 concrete walls, 3 soft partitions

The coverage ranges are:r=36m,29m,23m and 18m for date rate at 1Mbps, 2Mbps,
5.5Mbps and 11Mbps respectively
 Beamforming antennas in ad hoc networks

 W  ?
~ 
 n log n 
 
throughput obtained by each node

Beam-
new new forming
routing channel antennas
protocol access
scheme

P.Gupta and P.R. Kumar,00

 Beamforming antennas in ad hoc networks

Z0=50,L/2 Z0=25,L/2
Z0=50

Series resonant patch array

interference
Phased patch
antenna

target

Phased patch array

D.Lu and D.Rutledge,Caltech,02

Beamforming antennas in ad hoc networks

Medium Access Control Protocol(CSMA/CA)

CSMA/CA:carrier sense multiple access/collision avoidance
( for omnidirectional antennas)
 No standard MAC protocols for directional antenna
 No obvious improvement for throughput using beamforming antennas

Neighbor discovery
 Neighbor discovery become more complex using beamforming antennas.

Packet routing (Scheduled/On-demand)

 Ad hoc networks may achieve better performance in some cases
using beamforming antennas.
 Beamforming antennas can significantly increasing node and
network lifetime in ad hoc networks.
 Channel access

1) traditional exposed node 2) Omnidirectional and

problem for omnidirectional directional antennas solve
antennas the exposed node problem

A B C D E A B C D E

RTS
RTS RTS
CTS CTS
CTS CTS RTS
DATA DATA DATA
CTS CTS
DATA

The nodes DATA DATA DATA

DATA
are ACK
The node ACK
ACK
prohibit to is free to ACK
transmit or The node is
transmit or
receive blocked to
receive
signals communicat
signals
e with C

1) No coverage change. May save power.

Source:Y Ko et al., 00 2) B may not know the location of C.
 Channel access
3) beamforming antennas create new problems

A B C D E A B C D E

RTS RTS

CTS CTS
CTS RTS DATA
RTS
DATA

collision DATA

collision deaf
 Neighbor discovery
“Hello” Nt

t
B
A

E
D
A

AP Neighbors
A B,C
B A,C  
C A,B,E
D E
E C,D
Ad hoc WLAN for rural area
 Conclusion
Beamforming antenna systems improve wireless
network performance
-increase system capacity
-improve signal quality
-suppress interference and noise
-save power
Beamforming antennas improve infrastructure
networks performance. They may improve ad hoc
networks performance. New MAC protocol
standards are needed.
Vector antennas may replace spatial arrays to
further improve beamforming performance