Title: Smart Antennas for Wireless System
1Smart Antennas for Wireless System
ReferenceIEEE Personal Communications,1996.10
JACK H. WINTERS
ATT LABS-RESEARCH
ReporterYi -Liang Lin
Advisor Prof .Li-Chun Wang
2Smart Antennas for Wireless System
Multibeam antenna
Adaptive array antenna
3Impairment
Multipath fading
- Received signal amplitude and phase vary with
antenna location,direction and polarization and
with time
Delay spread
- Difference in propagation delay among multiple
path
- Intersymbol interference can occur
Co-channel interference
- Cellular system divide the available frequency
channels into channel sets,using one channel set
per cell with frequency reuse (frequency reuse
factor 7 )
- The number of channel set decreases ?
interference increases
- For given interference(channel set), cell size
?capacity
4Smart Antenna Techniques
- Permit greater coverage and capacity at each BS
- The signals received by multiple antenna elements
are weighted and combined to generate an output
signal
5Smart Antenna Techniques
- An array provides an increased antenna gain of M
plus a diversity gain against multipath fading
which depends on the correlation of the fading
among the antennas
- Define Antenna gain as the reduction in required
receive signal power for a given average output
signal-to-noise ratio(indep of environment)
- Define Diversity gain as the reduction the
required average output signal-to-noise ration
for a given BER with fading(only with multipath
fading)
6Diversity
Spatial Diversity
- Antennas are separated far enough for low fading
correlation
- Separation depends on angular spread
Handset , BS indoor ?angular spread
360quarter-wavelength
Outdoor system with high BS ? few degrees 10-20
wavelength
Polarization Diversity
horizontal
Low correlation
vertical
? Only double diversity , Small profile
Angle Diversity
7Diversity
- Adjacent narrow beams are used and have low
fading correlation
- With small angular spread,adjacent beams can have
received signal levels more than 10dB weaker than
the strongest beam,resulting in small diversity
gain
8Diversity
Four antenna diversity options with four antenna
elements
a) 7 ??spatial
b) 1/2??angularpolarization OR
7??spatialpolarization
c) 1/2??angular
4.2dB 2.9dB 4.4dB 1.1dB
10-2BER compare to two element array
9Diversity
- Diversity gain achieved in BS (uplink) to
compensate for the higher transmit power of the
BS on downlink
Selection diversity (Selecting the antenna with
the highest signal power)
using
Maximum ratio combining (weighting and combining
the received signal to maximize the signal
-to-noise ratio)
- The main limitation on the handset antennas is
typically not the handset size(diversity) , but
the cost and power consumption of the receiver
electronics for each antenna
10Smart Antenna
- Today cellular system,each BS uses three separate
sets of antennas - for each 120? sector ,with dual receive
diversity in each sector
Each sector uses a different frequency to reduce
co-channel interference, handoffs between sectors
are required
Performance ? narrower sector ? handoff
- This leads us to smart antennas without handoffs
between beams
Multibeam antenna
Smart antenna
Adaptive array antenna
Multibeam antenna ?different pattern
Antenna element
Adaptive array ?similar pattern
11Multibeam antenna
- Multiple fixed beam in a sector (four 30?beams
cover a 120?sector)
M-fold antenna gain
Some diversity gain by combining the received
from different beams(angle diversity)
Or achieve dual diversity by using a second
antenna array (orthogonal polarization or
space far from first)
- The same beam as on the uplink can be used for
the downlink , - thereby providing antenna gain (not diversity
gain) on the downlink
- These antennas have nonuniform gain with respect
to angle due to - scalloping , 2dB less gain
- ( decrease in gain between the beams due to the
beam pattern of each)
12Multibeam antenna
- Problem locking wrong beam due to multipath or
interference - and provide limited interference
suppression
Since they cannot suppress interference if it is
in the same beam as the desired signal
13Adaptive array
- The signals received by the multiple antennas are
weighted and combined to maximum the
signal-to-interference-plus-noise ratio
- Adaptive arrays have advantages
M-fold antenna gain without scalloping
M-fold diversity gain with sufficient low fading
correlation
- These array cancel N interference with M
antenna(MgtN) - and achieve M-N fold diversity gain
- (MltN is possible)
Requiring a receiver for each antenna
Tracking the antenna weight at fading at fading
rate versus beam switching every few seconds
with the multibeam antenna
14Adaptive array
Multipath LOS environment
15Adaptive array
- The adaptive antenna array weight and combines
the signal to - enhance desired signal reception and null
interference
A main beam in the direction of the desired
signal
It generate an antenna pattern
A null in the direction of interference
- Under this conditions, with the number of
antennas much greater than the number of arriving
signal rays , it is easier to express the array
response in terms of a small number of of angles
of arrival, rather than the received signal phase
at each antenna
MUSIC ESPRIT algorithms for improved
performance
- M array antennas can form up to M-1 nulls to
cancel up to M-1 - interference
- Such angular domain methods can be useful in some
wireless - situations with near-LOS( BS in flat rural with
high many antennas )
16Adaptive array
however
- The signal s arrive from each user via multiple
paths and angles of - arrival , it becomes impossible to form an
antenna pattern (above) - since the number of required nulls would be
much greater than the - number of antenna
- To provide diversity gain ,
the antennas at a BS can be spaced many
wavelengths apart , which results in many
grating lobes (many repetitions of the antenna
pattern in the field of view)
And with dual polarization antennas there is a
different pattern for each polarization
Antenna pattern is meaningless
17Adaptive array
- No matter how many paths each signal uses , the
result is a given phase and amplitude at each
antenna for each signal
- There is an array response for each signal , and
the performance of - the array depends on the number of signals ,not
the number of the - path, with analysis in the signal space domain
rather than the angular - domain
- The hold true as long as the delay spread is
small if not delayed - version s of the signals must be considered as
separate signal (below)
18Adaptive array
- An important feature of adaptive arrays in
multipath environments - the ability to cancel interferers independent of
the angle of arrival
- With multipath , objects around the antennas act
as a huge reflecting - antenna (with the actual antennas acting as
feeds ) - which permit the receiving array to separate
the signals
- In particular , if the receiving antennas are
spaced far enough apart - such that beams can be formed which are smaller
than the angular - spread, the signals from two closely spaced
antennas can usually be - separated using adaptive array combining
techniques
- The number of signals that can be separated
increases with - the number of receive antennas , the angular
spread and - the density of the multipath reflections within
the angular spread
?multipath can be beneficial
19Adaptive array
- With delay spread , the array treats delayed
versions of the - signals as separate signals
- An adaptive array with M antennas can eliminate
delay spread - over (M-1)/2 symbols or cancel M-1 delayed
signals over any delay
20Application
IS-136 TDMA
- 3users/channel,162symbols/timeslotDQPSK
modulation ?48.6kb/s
- Equalizer is neededhandle delay spread up to one
symbol
- Synchronization sequence14symbol/timeslot
Equalizer training Determine adaptive array
weight
Used to
- With rapid fading ,the channel change
significantly across a timeslot
? the adaptive array weights must be adjusted
across the time slot the equalizer is
relatively simple(using MLSE)
GSM TDMA
- 8users/channel,156.25symbols/timeslotMSK
modulation?270.833
- Because higher data rate ?delay spread over
several symbols
21Application
- Synchronization sequence26symbol/timeslot
- Channel does not change significantly over a time
slot - weight need only be calculated per frame
- Equalizer is more complex
IS95 CDMA
- Multiple simultaneous users in each 1.25MHz
channel - with 8kb/s per user, spreading gain 128
- RAKE receiver
- which combines delayed version of the CDMA
signal
Provide diversity gain
Overcome the delay spread problem
- The CDMA spreading codes can provide the
reference signal for adaptive array
weight calculation
22Range Increase
- With small angular spread
Adaptive array
M element
Provide M fold increase in antenna gain
Multibeam antenna
This increase range M1/? and reduce the number of
BS by M1/? for given area ?propagation loss
exponent
Adaptive array with spatial diversity for given
array size diversity gain increases with angular
spread ?greater range (decreases fading
correlation)
Multibeam (angel diversity ) provides only small
diversity gain ?diversity gain is limited
23Range Increase
The antenna gain of the multibeam antenna is
limited by the angular spread
The multibeam antenna cannot provide additional
antenna gain when the beamwidth is less than the
angular spread because smaller beamwidths
exclude signal energy outside the beam
24Range Increase
TDMA systems
25Range Increase
The range is limited to the predicted range
limitation
The range improvement is degraded due to the
angular spread for M less than the theoretical
value corresponding to the range limitation
The range exceed the no-diversity theoretical
range for all angular spreads ,due to antenna
diversity
The diversity gain increases with M,angular
spread, antenna spacing which decreases fading
correlation but does not increase for angular
spread greater than about 20
26Range Increase
- Range increase applies to upink
-
For IS136, GSM ,IS95
Since downlink freq is different from the uplink
freq the same adaptive array techniques cannot be
used for transmission by the BS
Multibeam antenna can be used
to achieve diversity gain transmit diversity must
be used
But
Or handset have multiple antennas
These techniques may provide less gain on the
downlink than on the uplink ,this may be
compensated for by the higher transmit power of
the BS as compared to the handset
TDMA systems
27Range Increase
Additional diversity gain of adaptive is much
smaller
Antenna gain limitation is much less
?adaptive array provide only a slightly larger
range increase than multibeam antennas
- Multibeam require less complexity
?preferable
Multibeam antenna
CDMA
may be preferable for
adaptive array
TDMA
(particularly large angular spread)
28Capacity Increase
CDMA
- Capacity(bit per sec,hertz,base station)depends
on
Spreading gain
Corresponding number of equal power co-channel
interference
- Multibeam antenna with M beams
- reduces the number of interference per beam by a
factor of M - and thereby increases the capacity M-fold
- Adaptive array provide only limited additional
interference - suppression because the number of interference
gt antennas
- Multibeam are less complex ?preferred in CDMA
29Capacity Increase
TDMA
- TDMA are limited in capacity by a few dominant
interference
- Multibeam antenna reduces the probability of the
interference - being in the same beam as the desired
?permits higher capacity through greater
frequency reuse (particularly small angular
spread)
- Adaptive array can cancel the dominate
interference with just - a few antenna
?M-element array permits greater than M-fold
increase (independent of angular spread)
4 element adaptive array ?permit frequency reuse
in every cell
for 7fold increase in capacity
4-beam antenna ?permit a reuse of three or four
for doubling of
capacity (small angular)
30Capacity Increase
Adaptive array ? uplink only
Multibeam antenna ? downlink
(less effective in reducing interference)
- This problem is worse in I-136
Because the handsets require a continuous
downlink, the same beam pattern must be used for
all three user in a channel ?reduce the
effectiveness of multibeam antenna against
interference
- Interference on uplink is typically worse than
downlink
The signal from interfering mobile could be
stranger than desired mobile mobile At the
mobile the signal from an interfering BS should
not be stronger since the mobile chooses the BS
with the strongest signal
BS are typically more uniformly spaced than the
mobiles