Communications and Network, 2013, 5, 217-222
http://dx.doi.org/10.4236/cn.2013.53B2041 Published Online September 2013 (http://www.scirp.org/journal/cn)
Coexistence Evaluation of LTE with Active Antenna
System
Fei Xue, Qimei Cui, Ting Fu, Jian Wang
Key Laboratory of Universal Wireless Communication, Ministry of Education, Beijing University of Posts and Telecommunications
Received June, 2013
ABSTRACT
Due to the new radio architecture of Active Antenna System (AAS), the LTE BS equipped with AAS will have more
flexible RF planning than that equipped with passive antenna. However, the new RF characteristic of AAS will have
potential impacts on the existing BS RF requirements, which is imperative to be evaluated. In this paper, general an-
tenna pattern and spatial Adjacent Channel Leakage Ratio (ACLR) are introduced considering the new RF features of
AAS, and the impacts of AAS on BS performance are analyzed based on the AAS fundamental applications. Sys-
tem-level coexistence simulation is conducted to evaluate the RF features of transmitter and receiver of AAS on BS
performance. Simulation results show that RF feature of spatial ACLR of AAS transmitter has little impact on thro ugh-
put loss in victim system and ACLR of 45 dB per transmitter of AAS can fully meet the coexistence requirement; RF
feature of in-band blocking level of individual receiver of AAS is higher than that of receiver of BS equipped with pas-
sive antenna around 1-4 dB, which imposes harder requirement to design the channel selection filter in the individual
receiver of AAS, thus in-band blocking requirement needs to be redefined for AAS BS in the existing 3GPP specifica-
tion.
Keywords: Coexistence Evaluation; AAS; LTE; Spatial Characteristics
1. Introduction
In the progress of the specification of 3GPP Release 11,
lots of new characteristics and studies are introduced to
further enhance the performance of LTE-Advanced [1].
Given that radio network are being required to support
multiple technologies and improve the system capacity in
the foreseeable future, there is a growing need to inte-
grate the antenna array and the transceivers in a typical
base station to minimize site footprint and lower costs [2].
The new radio architecture of AAS is raised in this de-
mand. Besides bringing in smaller site footprint and
lower costs, AAS can also support a host of advanced
electronic beam-tilt features that can enable improve-
ments in network capacity and coverage. A typical im-
plementation of advanced antennas tilt features of AAS
is vertical cell splitting in a single sector cell [6]. Thus,
AAS is introduced as an altern ative antenna system from
the one installed in the conventional BS. A study item
and work item have been approved in 3GPP TSG RAN
WG4 to study the characteristics of the AAS transmitter
and receiver, and investigate the impacts on the coexis-
tence performance with other systems based on un-coor-
dinated deployment [3].
Introducing of AAS may bring special changes in RF
performance. In the traditional passive antenna system,
there is only one transceiver connected with passive an-
tenna array through coaxial cable. However, as shown in
Figure 1, there are several individual transceivers in the
AAS, which is connected with individual antenna ele-
ment separately [2]. This modification in radio architec-
ture will lead the two RF features, ACLR and in-band
blocking being different from that in passive antenna
system. As the correlation level varies with different RF
signals (e.g. in channel signals and adjacent channel sig-
nals) induced on the antenna element in the physical de-
vice, composite antenna pattern for different RF signals
should be distinguished. ACLR of AAS will be spatially
distributed, which is different from flat distributed ACLR
of traditional passive antenna system [4,5]. Detailed
theoretical analysis of spatial ACLR is shown in Section
2. In the design of radio system, ACLR requirement is
indispensable for power amplifier in any radio transmit-
ter to guarantee an acceptable adjacent channel interfer-
ence caused by its nonlinear characteristic. Thus, it’s
imperative to investigate the spatial ACLR of AAS and
define the reasonable ACLR requirement for AAS.
Meanwhile, due to new radio architecture of AAS,
in-band blocking of RF receiver of AAS, which is used
to study the receiver’s ability to detect a wan ted sign al on
its assigned channel with an unwanted signal inside the
C
opyright © 2013 SciRes. CN

F. XUE ET AL.
218
operating band, should be evaluated at the interface be-
tween antenna element and individual receiver to define
its filter specification of channel selection filter, not at
the interface between receiver and coaxial cable in the
traditional passive antenna system. Thus, in-band block-
ing requirement of the individual receiver is also impera-
tive to be evaluated to define the filter specification to
guarantee the receiver function well. However, until right
now, there are no research papers analyzing the RF re-
quirements of AAS in detail to ensure practical applica-
tion.
In this paper, based on the new radio architecture of
AAS, general antenna pattern and spatial ACLR is in-
troduced considering the correlation level of different
signals induced at the antenna element. Then we ana-
lyzed the impacts of AAS on BS performance based on
the AAS fundamental applications. System-level coexis-
tence simulation is conducted to evaluate the spatial
characteristic of transmitter and receiver of AAS on BS
performance, extending our work in [11]. Simulation
results show that spatial ACLR of AAS transmitter with
different correlation levels has little impacts on through-
put loss in the victim system, especially for individual
transmitter of AAS with high ACLR. ACLR of 45 dB per
transmitter of AAS can fully meet the coexistence re-
quirement, and in-band blocking level of individual re-
ceiver of AAS is higher than that of receiver of BS
equipped with passive antenna around 1-4 dB, thus
in-band blocking requirement for AAS individual re-
ceiver needs to be redefined in the existing 3GPP speci-
fication [3].
The rest of this paper is organized as follows: the gen-
eral antenna pattern and the spatial ACLR of AAS is
introduced in Section 2. The impacts of AAS on BS per-
formance and evaluation results are described and ana-
lyzed in detail separately in Section 3 and 4. Con clusions
are drawn in Section 5.
2. General Antenna Pattern and Spatial
ACLR of AAS
2.1. General Antenna Pattern
Radio architecture of AAS is shown in Figure 1. Each
radiation element of AAS is connected with an individual
transceiver unit. However, active components connected
with each antenna element shall make the RF signals
induced on the antenna element from several transceivers
unit in the transceiver array partially correlated. For ex-
ample, the identical sign als are applied to the inpu t of the
transceiver array, unwanted signals generated on the ra-
diation element is implementation specific and hence
unknown in nature (e.g. amplified thermal signal noise is
random and hence will have no similarity in different
transceivers) [2,9]. Given this consideration, the compos-
ite antenna pattern of AAS should include both radiation
element pattern and correlation of the RF signals from
different transceivers.
The signal from the direction U is acting on the an-
tenna array, and elevation angle of the signal direction is
denoted as

and azimuth angle of the signal direction
is denoted as

. The composite antenna pattern of AAS
consists of three parts: radiation element pattern, array
factor for antenna array and correlation matrix of differ-
ent transceiver paths, which will be described as the fol-
lowing.
1) Antenna Element pattern: The horizontal pattern of
radiation element is described as [2]
2
3
()[12() ,],
m
dB
A
min A



E,H
(1)
where 3 is the horizontal 3dB bandwidth of
antenna element, and
65
dB
 
30
m
A
dB

is the front to back
ratio. The vertical pattern of radiation element is de-
scribed as [2]
2
3
90
()[12() ,],
v
dB
A
min SLA




E,V
(2)
where 3 is the vertical 3dB bandwidth of an-
tenna element, and is the attenuation
lower limit of side lobe. The antenna model of antenna
element can be modeled as [1]
65
dB
 
30dB
v
SLA 
10
,,,
20log((,))
{[()()], }
E
E
MaxEHEVm
P
GminAA A

   (3)
where (,)
E
P

 is denoted as the radiation gain of an-
tenna element, ,8
EMax iGdB

is the maximum direc-
tional gain of antenna element.
2) Array factor for single column: Due to specific ar-
ray placement and electrical steering down-tilt, the signal
radiated from each radiation element will experience
different phase shift before it arrived at the User Equip-
ment (UE) and the power strength of received signal will
vary with the specific phase shift. Here V is the phase
shift due to the array placement, as shown in Figure 2.
Figure 1. Radio architecture of AAS.

F. XUE ET AL. 219
X
Z

2
3
4
1
v
2
v
3
v
4
v
5
v

1
(,,)Pxyz
d


d


d


d


X
Z

2
3
4
3
w
4
w
5
w
2
(,,)
P
xyz
d


d


d


d


1
w
2
w
Figure 2. Geometry for calculation of the phase shift for
each dipole in the far field.
12 1,2,...,,[,,...,],
T
NnVvv vN (4)
(-2 ),(1) ()
v
n
d
v expin cos


 (5)
W is the steering factor, which is used to control the
basic physical characteristics of composite antenna pat-
tern, for example 3dB beam-width, side lobe attenuation,
electrical down-tilt [10]. Steering vector W mainly pro-
vides the minimum power level of side lobe, the electri-
cal down-tilt and side lobe attenuation for composite
antenna pattern, which can be denoted as
12 1,2,...,,[,,...,],
T
NnWwwwN (6)
1(-2 ),
(1)()
v
n etilt
d
w exp
Nin sin
   (7)
the electrical down-tilt etilt

is set to be that opti-
mizes the system throughput [7]. The combined phase
shift for each dipole in the far field can be denoted as
9
,WWV
 (8)
3) Correlation matrix of RF signals for single column:
The RF signals induced on all antenna elements is
12
()[ (),(),...,()],
N
Sts tstst (9)
The complex output of antenna array at the far field
becomes
1
(,,) ()
()
(,)
(,) ,
N
nnn
n
E
yt st
St
wE
PW









(10)
where (,)
n
E

 is the radiation gain of antenna element
together with phase shift due to array placement at the far
field and we assume that radiation gain of different an-
tenna elements is identical. The combined radiation pat-
tern is the mean output power of antenna array at the far
field which can be obtained by calculating conditional
expectation over 2 (,,)yt  [9].
2
2
2
(, )(,,)
(()()) (,)
(,) ,
HH
E
H
E
PEyt
ESt St
R
PW
PWW
 



W

 





(11)
R is the correlation matrix of the RF signals in different
transceiver paths represented as
11 121
21 222
12
(()())
N
N
H
nm
NN NN
RR R
RR R
REStSt R
RR R

,











 





(12)
Correlation of RF signals in different transceiv er paths
is denoted as a correlation coefficient, , de-
fined as the similarity of the unwanted signals generated
in different transceiver paths when an identical signal is
applied at the input of individual transceiver unit. Un-
wanted signals under different circumstances can be re-
garded as correlated, (e.g. unwanted signals
generated by Crest Factor Reduction will be generated
digitally and hence identical in each path [10]), or un-
correlated,
01
nm
R
1
nm
R
0
nm
R

, (e.g. amplified thermal noise in
each amplifier [2]). Given that the fast fading between
different antenna elements is spatially correlated, corre-
lation coefficient can be denoted as [2]
,,
1
22
,,
11
,
K
nk mk
k
nm KK
nk mk
kk
s
R
ss
s







(13)
where K is the number of sampling point in a distin-
guished time slot. For the sake of complexity and suffi-
cient correlation level analyzed in the coexistence study,
the same level correlation level

is assumed, which is
a value between 0 and 1 [2].
1
1
(()()) 1
1
H
REStSt .












 



 (14)
The radiation pattern of composited antenna array can
be simplified as [2]
2
10 1
10
(,) 10log11
20log( ,)) (
N
An
n
E
Aw
P
v  

n





  








 (15)
2.2. Spatial ACLR Model
In the existing specification [3], ACLR is defined as the
ratio of the filtered mean power on the assigned channel
frequency to the filtered mean power at an adjacent

F. XUE ET AL.
220
channel frequency. Due to the fact that the different RF
signals (e.g. in channel signals and adjacent channel sig-
nals) in different transceiver paths induced on the an-
tenna element have different correlation level in AAS BS,
antenna pattern should be applied separately for different
RF signals. Spatial ACLR of AAS can be described as
_, _,
() ()(),
iInChiAdjChi
s
ts tst (16)
2
_,
10 2
_,
(())
10log ,
(()
InCh i
Element
AdjCh i
Ey t
ACLR Eyt )









(17)
where _,
()
InCh i
s
t
()
represents RF signal induced at the
antenna element i on the assigned channel frequency, and
_,AdjCh i
i
s
t represents RF signals induced at the antenna
element i on the adjacent channel. We assume the RF
signals in different transceiver on the assigned channel
frequency are fully correlated, namely 1


1.
, and the RF
signals in different transceiver on an adjacent channel are
partially correlated, namely 0


 The spatial ACLR
in dB can be denoted as
2
1
10 2
1
() 10log)
11
,
N
ii
i
N
ii
i
Element
wv
ACLR
wv
ACLR




























(18)
3. Impacts of AAS on BS Performance
3.1. Impacts of Transmitter Characteristics
As shown in Figure 3, ACLR of AAS BS is spatially
distributed in the vertical plane and ACLR of traditional
passive antenna system is flat. Compared with the tradi-
tional passive antenna system, some of the areas covered
by adjacent system may suffer more interference from
020 40 60 80100 120 140 160 180
0
10
20
30
40
50
60
T het a ( de g)
ACLR (d B)
ro=1
ro=0 .8
ro=0 .6
ro=0 .4
ro=0 .2
ro=0
Figure 3. Spatial ACLR of AAS.
AAS BS while some others may suffer less. Therefore
spatial ACLR of AAS will have potential impacts on cell
average and cell edge throughput loss.
3.2. Impacts of Receiver Characteristics
As mentioned in Section 1, there is only one receiver
connected with antenna array through coaxial cable in
traditional passive antenna system. Therefore, in-band
blocking, which is the RX power level from UEs within
the systems at the adjacent channel, should be evaluated
at the interface between receiver and coaxial cable.
In-band blocking for the receiver of traditional passive
antenna system experienced the composite antenna pat-
tern gain and cable loss. However in the AAS, there are
several individual receivers, which is connected with
individual antenna element separately. Thus, in-band
blocking for individual receiver should be evaluated at
the interface between antenna element and individual
receiver. In-band blocking will just experience the an-
tenna element pattern gain, which is quite different from
that of traditional passive antenna system.
4. Performance Evaluation
4.1. Coexistence Scenario
In this paper, E-UTRA Macro to E-UTRA Macro coex-
istence scenario is evaluated by system-level simulation
for the purpose of investigating the spatial characteristics
of AAS BS, as shown in Table 1 and Table 2. The
channel frequency of the aggressive system is located
tightly besides that of the victim system.
Table 1. Coexistence scenario for ACLR.
Case Aggressor Victim
1a AAS E-UTRA Macro
system Legacy E-UTRA Macro system
1b AAS E-UTRA Macro
system AAS E-UTRA Macro system
1c(Baseline) Legacy E-UTRA Macro
system Legacy E-UTRA Macro system
Table 2. Coexistence scenario for in-band blocking.
Case Aggressor Victim
1a Legacy E-UTRA Macro
system AAS E-UTRA Macro system
1b AAS E-UTRA Macro
system AAS E-UTRA Macro system
1c(Baseline) Legacy E-UTRA Macro
system Legacy E-UTRA Macro system

F. XUE ET AL. 221
4.2. Network Layout
The layout of the victim and aggressor network is identi-
cal and aggressor network’s sites are located at the victim
network’s cell edge with worst site shifts. The inter-site
distance is 750 m. Detailed network layout can be found
in [3].
4.3. Large Scale Channel Model
The channel path loss model is defined as [3]
PathLoss = max{L(R), Free_Space_Loss} + shadow-
fading,
where Free Space Loss is defined as
Free_Space_Loss = 98.46 + 20*log10(R) (R in kilome-
tre),
L(R) is defined as
L(R) = 128.1 + 37.6 log10(R),
The final coupling loss is defined as
Coupl_Loss_Macro
= max{PathLoss,Free_Space_Loss}-G_Tx-G_Rx,
where G_Tx is the transmitter antenna gain and G_Rx is
the receiver antenna gain.
4.4. Simulation Parameters
General simulation parameters for coexistence study of
AAS BS are listed in Table 3 [8]. The Power Control
(PC) scheme in LTE system is performed for the evalua-
tion of in-band blocking, and detailed power control pa-
rameters and link to system throughput mapping are de-
scribed in [3]. Both PC set 1 and PC set 2 are adopted for
studying the spatial receiver characteristics of AAS BS.
Table 3. General simulation parameters.
Parameters Values
Carrier freque nc y 2 GHz
System bandwidth 10 MHz
Minimum dist ance U E<->BS 35 m
Log normal sha d ow i n g Standard Deviation of 10 d B
Shadow correlation coefficient 0.5 ( inter site) / 1.0 (intra site)
Number of active UEs UL: 3UEs(16RBs /UE)
DL:1UE(50RBs/UE)
UE max Tx Power 23 dBm
UE min Tx Power -40 dBm
BS max Tx Power 46 dBm
Scheduling algorithm Round Robin, Full buffer
Antenna configuration at UE Omni-dire ctional
The height of BS 30 m
The height of UE 1.5 m
Number of antenna elements 10
ACS of LTE UE 33 dB
Cable Loss (Legacy/AAS) 1/0 dB
4.5. Simulation Results
Simulation results of Case 1a and Case 1b in Figure 4
and Figure 5 show that different correlation levels of
spatial ACLR have little impact on cell average and cell
edge throughput loss. The reason is that the adjacent
channel interference in the downlink mainly depends on
adjacent channel selectivity of UE, especially for AAS
BS with high ACLR. Comparing Case 1a Case 1b and
Case 1c, the throughput loss is almost consistent with the
same ACLR assumption. The small gap between Case 1a
and Case 1b is due to the cable loss of passive antenna in
Case 1a resulting in lower transmit power and hence
higher throughput loss in victim system. For AAS BS,
simulation results indicate that ACLR of 45 dB per trans-
ceiver for specific coexistence scenario studied is suffi-
cient to fulfill the coexistence requirement. No matter the
victim system is equipped with passive antenna system
or AAS, the throughput loss of victim system in Case 1a
and Case 1b is lower than 5%.
30 35 40 45 50
0. 025
0.03
0. 035
0.04
0. 045
0.05
0. 055
0.06
0. 065
0.07
ACLR Element [ dB]
Do wn lin k ce ll av er age t hr ou ghp ut lo ss [ % ]
case1a-0.0
case1a-0.2
case1a-0.4
case1a-0.6
case1a-0.8
case1a-1.0
case1b-0.0
case1b-0.2
case1b-0.4
case1b-0.6
case1b-0.8
case1b-1.0
case1c-1.0
Cas e 1a
Case1 c
Case1b
Figure 4. Downlink cell ave r a ge throughput loss.
30 35404550
0. 06
0. 08
0.1
0. 12
0. 14
0. 16
0. 18
0.2
0. 22
ACLR Eleme n t [d B]
Downlink cell edge througput loss [%]
case1a-0.0
case1a-0.2
case1a-0.4
case1a-0.6
case1a-0.8
case1a-1.0
case1b-0.0
case1b-0.2
case1b-0.4
case1b-0.6
case1b-0.8
case1b-1.0
case1c-1.0
case1a
case1c
case1b
Figure 5. Downlink cell edge throughput loss.

F. XUE ET AL.
222
Case 1aCas e 1 bCas e 1 c
-60
-55
-50
-45
-40
-30
-20
-10
0
Uplink in-band blocking [dBm]
PC1
PC2
-42. 27
-51.47
-42.61
-51.79
-43. 66
-55.61
Figure 6. Uplink in-band blocking.
Comparing Case 1a and Case 1c in Figure 6, the up-
link in-band blocking interference signals presented at
the AAS individual receiver are higher than that of the
receiver of BS equipped with traditional passive antenna
for both PC set 1 and PC set 2 around 1-4 dB. The block-
ing interference signals of Case1a are a little higher than
that of Case 1b. The reason is that the aggressive BS in
Case 1a equipped with traditional passive antenna has
higher cable loss than AAS, which leads to higher trans-
mit power of aggressive UE and hence higher blocking
interference signals in victim system. The simulation
results indicate that it is necessary to redefine the block-
ing requirement for AAS individual receiver to design
the channel selection filter ensuring performance and
stability under different coexistence scenarios.
5. Conclusions
In this paper, the spatial transmitter and receiver charac-
teristics of AAS are investigated and evaluated through
system level coexistence simulation. Simulation result
show that different correlation level has little impact on
throughput loss caused by spatial ACLR of AAS, and
ACLR of 45 dB per transmitter of AAS is sufficient to
meet the coexistence requirement. And in-band blocking
level of AAS individual receiver is higher than that of the
receiver of BS equipped with traditional passive antenna
around 1-4 dB, thus it is necessary to redefine in-band
blocking requirement for the individual receiver of AAS
in 3GPP LTE existing requirement.
6. Acknowledgements
The research work is supported by National Science and
Technology Major Project of China (2012ZX03001039,
2013ZX03001018), Beijing City Science and Technol-
ogy Project (D121100002112002).
REFERENCES
[1] 3GPP, “Overview of 3GPP Release 11 V0.1.4”,
Mar.2013.
[2] 3GPP TR 37.840 V0.3.0, “Study of AAS Base Station”,
Dec.2012.
[3] 3GPP TR 36.942 V10.3.0, “E-UTRA Radio Frequency
(RF) System Scenario,” June.2012.
[4] 3GPP TR 25.816 V8.0.0, “UMTS 900 MHZ Work Item
Technical Report,” Sept.2009.
[5] 3GPP TR 36.814 V9.0.0, “Further Advancements for
E-UTRA Physical Layer Aspect,” Mar.2010.
[6] O. N. C. Yilmaz, S. Hamalainen and J. Hamalainen,
“System Level Analysis of Vertical Sectorization for
3GPP LTE,” IEEE 6 International Symposium On Wire-
less Communication System (ISSWCS), pp. 453-457,
Oct.2009.
[7] 3GPP R4-124171, “On the Down-tilt for 3D Coexistence
Simulation,” 3GPP TSG RAN WG4 Meeting#64.
[8] 3GPP R4-122397, “Text Proposal for Simulation As-
sumptions for AAS,” 3GPP TSG RAN WG4 Meeting#63.
[9] 3GPP R4-123245, “Correction on Composite Array Ra-
diation Pattern for AAS,” 3GPP TSG WG RAN4 Meet-
ing#64bis.
[10] 3GPP R4-123925, “Discussion for Composite Radiation
Pattern of AAS”, 3GPP TSG RAN WG4 Meeting#64.
[11] P. C. Kang, Q. M. Cui, S. Chen and Y. J. Liu, “Perform-
ance Evaluation on Coexistence of LTE with Active An-
tenna Array System,” IEEE23rd International Symposium
on Personal Indoor and Mobile Radio Communication
(PIMRC), pp. 1066-1070, Sept. 2012.