IEEE C802.16m-07/244r1
Project
IEEE 802.16 Broadband Wireless Access Working Group <http://ieee802.org/16>
Title
Collaborative MIMO
Date
Submitted
2007-11-07
Source(s)
Yang Song, Liyu Cai, Keying Wu,
Hongwei Yang
Voice: + 86 21 58541240 Ext 7175
E-mail: Liyu.Cai@alcatel-sbell.com.cn
Alcatel-Lucent
Re:
IEEE 802.16m-07/040 Call for Contributions on Project 802.16m SDD
Abstract
System description of collaborative MIMO (Co-MIMO) technique
Purpose
To incorporate the proposals into the 802.16m SDD
Notice
Release
Patent
Policy
This document does not represent the agreed views of the IEEE 802.16 Working Group or any of its subgroups. It
represents only the views of the participants listed in the “Source(s)” field above. It is offered as a basis for
discussion. It is not binding on the contributor(s), who reserve(s) the right to add, amend or withdraw material
contained herein.
The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution,
and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name
any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole
discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The
contributor also acknowledges and accepts that this contribution may be made public by IEEE 802.16.
The contributor is familiar with the IEEE-SA Patent Policy and Procedures:
<http://standards.ieee.org/guides/bylaws/sect6-7.html#6> and
<http://standards.ieee.org/guides/opman/sect6.html#6.3>.
Further information is located at <http://standards.ieee.org/board/pat/pat-material.html> and
<http://standards.ieee.org/board/pat>.
Collaborative MIMO
Yang Song, Liyu Cai, Keying Wu, Hongwei Yang
Alcatel-Lucent
Background
IEEE 802.16m system requirement [1] shows great concerns on improved coverage, average user throughput
and cell edge user throughput. It is well known that the most significant factor limiting the coverage and cell
edge user throughput of a cellular wireless network is inter-cell interference (ICI), especially with low frequency
reuse factor.
As stated in [1], interference mitigation schemes and advanced antenna techniques shall be supported.
Collaborative MIMO (Co-MIMO) technique [2][3], is an advanced multiple-BS-based joint MIMO
transmission technique which aims at ICI mitigation through BS coordination. With performance improvement,
1
IEEE C802.16m-07/244r1
Co-MIMO provides a simple way to implement BS coordination and is of little implementation requirement and
complexity due to the fact that many existing techniques can be reused for Co-MIMO. Additionally, Co-MIMO
is a complement to the traditional single-user (SU-MIMO) and multi-user MIMO (MU-MIMO) in the future
cellular system and has the flexibility to cooperate with other MIMO modes.
Scheme Description of Collaborative MIMO
Co-MIMO is an extension of conventional single-BS based MIMO techniques. It allows multiple BSs to serve
one or multiple MSs simultaneously over the same radio resource through coordination among BSs. Co-MIMO
is characterized by the following three features:
1) In Co-MIMO, each MS is jointly served by multiple BSs, which is different from conventional MIMO
techniques where each MS is only served by a single BS. Through BS coordination, the ICI among these
coordinating BSs can be reduced greatly especially in the case of low frequency reuse.
2) In Co-MIMO, each BS can also serve several MSs simultaneously using the same radio resource in order
to further increase the system throughput. The co-channel interference (CCI) among these MSs can be
minimized by spatial division multiple access (SDMA) such as beamforming or multi-user MIMO (MU-MIMO)
techniques.
3) In Co-MIMO, SDMA processing is implemented independently within individual coordinating BS based
on the channel state information (CSI) between this BS and its served MSs. Therefore, Co-MIMO limits the
amount of information exchange between BSs, which is different from other BS coordination approaches
requiring massive information exchange.
Figure 1 gives an illustration of Co-MIMO where three neighboring BSs jointly serve three MSs in the
collaborative region denoted by the hexagonal shadowed area. The scheduler, a logical functional module in the
network, is responsible for collecting the information from coordinating BSs and determining the serving
relationship between BSs and MSs. Figure 1 shows an example of the serving relationship: MS1 is served by
BS1 and BS3, MS2 by BS1 and BS2, and MS3 by BS2 and BS3. Furthermore, more BSs can be involved into the
coordination for Co-MIMO transmission depending on the acceptable complexity and the backbone capability.
2
IEEE C802.16m-07/244r1
MS3
BS2
BS3
MS2
Signa l for MS1
Signa l for MS2
MS1
Signa l for MS3
BS1
Ba ckbone
Scheduler
Figure 1 An example of Collaborative MIMO
A conceptual Co-MIMO architecture is depicted in Figure 2. In this figure, M coordinating BSs are connected
with a scheduler via backbone. BSi (i =1,…, M) contains a precoding matrix generator which calculates its own
precoding matrix Wi to implement SDMA. At the BS side, Co-MIMO may work in the following steps:
1) Every BS reports to the scheduler certain channel information, such aschannel quality indicators (CQIs),
between MSs and itself through backbone.
2) The scheduler determines the serving relationship between BSs and MSs, and transfers the scheduling
decision to each coordinating BS. Based upon the scheduling decision, user data streams shall be delivered by
the user data distributor to the corresponding BSs. In Figure 2, di denotes the user data streams that shall be
transmitted by BSi.
3) Based on the scheduling decision, BSi estimates the CSI, Hi, between the served MSs and itself. Then the
precoding matrix generator in BSi independently derives the precoding matrix Wi to implement e.g.
beamforming or MU-MIMO techniques for BSi itself.
As the precoding operation in Co-MIMO is performed individually within each BS, there is no need for
excessive CSI exchange among coordinating BSs, which leads to very limited information transfer over the
backbone as well as low computational complexity.
3
IEEE C802.16m-07/244r1
Channel
knowledge, e.g.
CQI
Data streams for
MS1 to MSK
Scheduling
decision
d1
…
…
MS1
MSK
BS1
…
channel knowledge,
e.g. CQI
Backbone
Scheduler
Precoder
…
d1
dM
Scheduling
decision
Precoding
matrix
generator
W1
…
User Data
distributor
H1
Channel
knowledge, e.g.
CQI
Scheduling
decision
HM
Precoding
matrix
generator
WM
dM
…
Precoder
BSM
Figure 2 Conceptual architecture of Collaborative MIMO
Advantages of Collaborative MIMO
Co-MIMO has the advantages of ICI mitigation and spectral efficiency improvement. ICI mitigation is realized
by turning interference from neighboring cells into useful signals and separating signals for different users via
multi-user MIMO or SDMA techniques. Spectral efficiency improvement comes from two factors, ICI
mitigation and more supported users at the same time-frequency resource.
Moreover, Co-MIMO provides a simple way to implement BS coordination and an acceptable tradeoff between
the performance gain and the stringent requirement on the backbone traffic and implementation complexity. On
the one hand, Co-MIMO requires negligible control signaling transfer and from equal to doubled data traffic
delivery over backbone depending on the MIMO mode for the user. On the other hand, Co-MIMO requires not
much implementation complexity because some existing techniques can be reused or helped. For instance, user
data traffic delivery to multiple coordinating BSs in Co-MIMO can be realized in the same manner as multicast
over backbone in enhanced multicast broadcast (EMBS); and there is no further computational complexity in
BSs supporting Co-MIMO compared with the BSs supporting multi-user MIMO techniques.
Simulation Assumptions and Results
In this section, we use numerical results to demonstrate the advantage of the proposed Co-MIMO technique. In
4
IEEE C802.16m-07/244r1
our system level simulation, the collaborative region is set to be three neighboring sectors belonging to different
BSs as shown in Figure 3. As the ICI is most severe in the cell edge, we focus on the evaluation of Co-MIMO
performance in the cell-edge area. The cell-edge area is determined by the cell-edge length in our simulation and
is marked green in Figure 3.
We assume a 19-collaborative-region configuration in the simulation as depicted in Figure 4. The users are
randomly dropped with a uniform distribution in the cell-edge area, one user per sector. The system
performances with different sizes of cell-edge area are evaluated in the system by setting the cell-edge length to
be 150m, 100m, and 50m. In each collaborative region, three coordinating BSs communicate with three MSs
located in its cell-edge area using the same time and frequency resource. Through scheduling, every MS in the
cell-edge area will be served by two coordinating BSs and every BS will serve two MSs simultaneously.
Precoding matrix for the two served MSs is calculated independently within each single sector of a coordinating
BS. Every coordinating BS transmits with full power and equal power allocation to the two MSs. A single
narrowband subcarrier is assumed in the system without loss of generality.
Traditional SU-MIMO is used as the benchmark for comparison where three MSs in the cell-edge area were
served independently by different BSs without any coordination in frequency reuse 1 deployment.
Cell radius
One collaborative region
BS2
BS3
Serving sector
Cell- edge
a rea
Interfering sectors
Cell- edge
length
One cell
BS1
Figure 3 Cell-edge area in a collaborative region
Figure 4 Cell configuration
The parameters used in the system-level simulation are listed in Table 1.
Table 1 Parameters in system-level simulation
Center frequency
Frequency reuse factor
Thermal noise spectral
density
Pathloss model
Wall penetration loss
Shadowing correlation
BS Tx power per Tx
antenna per subcarrier
BS Tx antenna
2.5 GHz
1
Cell radius
Number of drops
500 m
1000
-174.3 dBm/Hz
Noise figure
5 dB
COST-231
10 dB
0.5 (inter site)
1.0 (intra site)
Fast fading model
MS speed
SCM: Urban micro
30 km/h
Lognormal shadowing
8 dB
5.3 dBm
MS implantation loss
5 dB
4 Tx antennas per
MS Rx antenna
2 Rx antennas per
5
IEEE C802.16m-07/244r1
sector per BS with 4
wavelength spacing
17 dBi
30 m
BS antenna gain
BS height
   2

 , Am 
A    min 12

  3dB 

BS antenna pattern
θ3dB = 70 degree,
MS antenna gain
MS height
MS with 1/2
wavelength spacing
4.1 dBi
1.5 m
RF band filter loss
0.5 dB
Am = 20 dB
It can be seen from Figure 5, even though there remains certain amount of residual ICI, the interference power
from neighboring cells within the collaborative region is considerably reduced. Taken into account the
interference from cells outside the collaborative region, the reduction of overall interference power is less
significant. This can be further solved by extending the collaborative region to involve more coordinating BSs.
The cumulative distribution function (CDF) of cell-edge capacity per sector of traditional SU-MIMO and CoMIMO are plotted in Figure 6. As shown in Figure 7, the average capacity gain of Co-MIMO over SU-MIMO
are 31.7%, 50.2% and 71.9% with respect to cell-edge length of 150m, 100m and 50m. The capacity gain of CoMIMO over SU-MIMO increase with the shrinking of the cell-edge area. This implies that Co-MIMO is more
applicable to cell-edge users. So Co-MIMO is a complement to the traditional SU-MIMO and MU-MIMO in the
future cellular system.
Interference power
1
0.9
0.8
0.7
Single-user MIMO: 50m
Collaborative MIMO: 50m
Single-user MIMO: 100m
Collaborative MIMO: 100m
Single-user MIMO: 150m
Collaborative MIMO: 150m
CDF
0.6
0.5
0.4
Interference from cells
within collaborative
region
0.3
Overall interference
0.2
0.1
0
-150
-140
-130
-120
-110
-100
Interference power (dBm)
-90
-80
Figure 5 CDF of interference power of the cell-edge users
6
IEEE C802.16m-07/244r1
Capacity
1
100
0.9
90
0.8
80
71.9%
Average capacity gain (%)
0.7
CDF
0.6
0.5
0.4
Single-user MIMO: 50m
Single-user MIMO: 100m
Single-user MIMO: 150m
Collaborative MIMO: 50m
Collaborative MIMO: 100m
Collaborative MIMO: 150m
0.3
0.2
0.1
0
0
1
2
3
4
5
6
7
Capacity (bit/s/Hz/Sector)
8
9
70
60
50.2%
50
40
31.7%
30
20
10
0
10
Figure 6 CDF of cell-edge capacity
150
100
Cell-edge length (m)
50
Figure 7 Average capacity gain
Coexistence of Co-MIMO and Other MIMO techniques
One of the MIMO mode categorizations concerns the number of BSs and MSs involved in the transmission. In
this way, three MIMO modes shall be supported in IEEE 802.16m system:

SU-MIMO: each user is served by only one BS and it occupies the resource exclusively, including time,
frequency, and so on. SU-MIMO is a preferred option when a system wants to optimize single user
performance

MU-MIMO: several users are served by one common BS by sharing the same radio resource simultaneously.
MU-MIMO is most favorable for heavily loaded systems where the maximization of overall system
throughput is the primary concern.

Co-MIMO: each MS is jointly served by multiple BSs and at the same time each BS can serve several MSs
simultaneously. Co-MIMO can greatly reduce the inter-cell interference through BS coordination in the
cellular network with low frequency reuse factor where severe inter-cell interference may degrade link
quality a lot to very lower SNR especially at the cell edge.
SU-MIMO and MU-MIMO can be referred to as the single-BS based MIMO, while Co-MIMO is multi-BS
based MIMO. In order to fully exploit the potential MIMO gain and meet the diversified system requirements,
the three complementary MIMO modes should coexist in IEEE 802.16m with the support of a practical lowcomplexity adaptive switching mechanism.
When the inter-cell interference becomes severe, such as in the cell edge for low frequency reuse deployment,
Co-MIMO should be exploited to reduce the interference and optimize the system coverage and throughput
performance at the cost of limited information exchange over backbone. Otherwise, SU-MIMO and MU-MIMO
can be selected.
7
IEEE C802.16m-07/244r1
Figure 8 A coexistence example of three MIMO modes
As a result of MIMO mode selection, over the same radio resource a BS may:

Serve one MS with its all transmit antennas (SU-MIMO)

Serve multiple MSs by SDMA (MU-MIMO or Co-MIMO)

Serve one MS while mitigating interference to other MSs by interference nulling
For any MS, it may be:

Served by one BS (SU-MIMO or MU-MIMO)

Served by multiple BSs (Co-MIMO)

Interference mitigated by one or more BSs
Besides, any single-user MIMO (SU-MIMO) modes (spatial multiplexing, space-time coding, etc.) can be
applied to one MS over the equivalent MIMO channel generated by Co-MIMO. It follows that one or multiple
data streams can be transmitted from multiple coordinating BSs in Co-MIMO. The coordinating BSs may
perform link adaptation, including the desired SU-MIMO mode and modulation and coding, that the MS
recommended based on the equivalent MIMO channel.
Text Proposal to SDD
------------------------------------Start text proposal------------------------------------
Collaborative MIMO based on multiple BS coordination
Severe inter-cell interference of cell edge users can be mitigated by collaborative MIMO (Co-MIMO) technique
with moderate requirement on BS coordination. Co-MIMO allows multiple BSs to serve one or multiple MSs
simultaneously over the same radio resource through coordination among BSs. Co-MIMO is characterized by
8
IEEE C802.16m-07/244r1
the following three features:
1) In Co-MIMO, each MS is jointly served by multiple BSs, which is different from conventional MIMO
techniques where each MS is only served by a single BS.
2) In Co-MIMO, each BS can also serve several MSs simultaneously using the same radio resource in order to
further increase the system throughput. The co-channel interference (CCI) among these MSs can be
minimized by spatial division multiple access (SDMA) such as beamforming or multi-user MIMO (MUMIMO) techniques.
3) In Co-MIMO, SDMA processing is implemented independently within individual coordinating BS based on
the channel state information (CSI) between this BS and its served MSs.
Co-MIMO and single-BS based MIMO modes such as SU-MIMO and MU-MIMO shall coexist in IEEE
802.16m system with the support of a practical low-complexity adaptive switching mechanism.
------------------------------------End text proposal------------------------------------
References
[1] IEEE 802.16m-07/002r4, IEEE 802.16m System Requirements, 2007-10-19
[2] IEEE C802.16m-07/162, Collaborative MIMO Based on Multiple Base Station Coordination, 2007-8-29.
[3] IEEE C802.16m-07/163, Coexistence of Collaborative MIMO with Single-Base-Station Based MIMO,
2007-8-29.
9