Transcript IEEE C802.16m-09/0020r3

Interference Mitigation using Conjugate Data Repetition for Cell Edge Users
Document Number: IEEE C80216m-09_0020r3
Date Submitted: 2009-01-12
Source:
Kiran Kuchi, J. Klutto Milleth, Vinod R, Dileep M K,
Divagar , Padmanabhan M S,
Bhaskar R, Giridhar K
CEWiT, India
Voice:
E-mail:
[email protected],in,
[email protected]
Venue:
San Diego, USA. In response to the TGm Call for Contributions and Comments IEEE 802.16m-08/052 for
Session 59
Topic TGm SDD - section 20 (Interference Mitigation)
Base Contribution: IEEE C80216m-09_0020
Purpose: Presentation associated with comment tgmsdd_Kuchi_Kiran.cmtb associated with section 20 in SDD
To discuss in TGm for appropriate action. Notice:
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.
Release:
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.
Patent Policy:
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 >.
Presentation Outline
• Motivation
• Conjugate Data Repetition (CDR)
– Improving cell edge performance
• For both data and common control channels
• Conclusions
• Proposed SDD Text
Motivation
• Limited Spectrum
– Frequency reuse-1 is the most likely deployment mode
– Typical urban cell size has 200-300 m radius
– Interference limited in both uplink and downlink
– Very low cell edge data rates
• Provide uniform data rates throughout the cell
• Improve control channel coverage
• High Spectrum Efficiency
Interference Mitigation
• Poor cell edge SINR in reuse-1 networks
– Severe co-channel interference (CCI)
– Multiple CCI up to 4-dominant interferers
– Typically cell edge SINR is in the [-6 0] dB range
– Nearly 30% of users in the sector are cell edge
users
• Re-use 1:3 is spectrally inefficient
• Gain of soft re-use is not adequate for cell-edge
users
Interference Suppression in Legacy
WiMax Systems
•
•
•
•
•
2-antenna MMSE receiver nulls a single interferer
Pre-whitened MLD can suppress two interferers
Use single stream transmission i.e., POD/CL-div
However, cell edge users typically have 3-4 dominant interferers
Existing solutions
– For low SINR cell edge users
– Rate ½ QPSK and bit-level data repetition up to 6-times
• Bit level data repetition is spectrally inefficient
Conjugate Data Repetition (CDR)
•
•
•
•
BS-1
Signal repetition in time/frequency
Allow interference and suppress it
Conjugate symbol repetition across
adjacent subcarriers or adjacent
OFDM symbols
All cells synchronously transmit
data in conjugate symbol pairs for
select users
–
•
•
•
The network assigns a CDR frequency
partition in which the RBs from different
BSs overlap
BS co-ordination not required
MMSE filtering of complex, and
complex-conjugate signals provides
a high IC gain
MMSE IC for each RB independently
BS-2
F1
X1
F1
Y1
F2
X1*
F2
Y1*
F3
X2
F3
Y2
F4
X2*
F4
Y2*
F5
P1
F5
P2
F6
X3
F6
Y3
F7
X3*
F7
Y3*
F8
X4
F8
Y4
F9
X4*
F9
Y4*
Conjugate Data Repetition
•
Basic Idea
– Each BS transmits data in conjugate symbol pairs on a pair of subcarriers
– 1st BS transmits
y1  hx  gxˆ1  n1
( x , x* )
y2  hx*  gxˆ1*  n2
– 2nd BS transmits ( xˆ1 , xˆ1 )
*
– Complex-conjugation on
y2
y2*  h* x  g * xˆ1  n2*
– After complex de-conjugation, the receiver has two copies of signal and interference with
different channel states
•
Signal
( h , h* )
•
Interference
(g , g*)
*
– Receiver jointly filters ( y1 , y2 )
– Receiver does not estimate interference channel. MMSE needs interference covariance only
CDR-MMSE Receiver
•
CDR signal model
M
y  h x d   g l xl  n
 ,
 l 1
Signal
 Noise
Interferen ce
•
CDR-MMSE filtering,
 h1 
h * 
h 1
 h2 
 *
h2 
 g l ,1 
g* 
l ,1
gl   
 g l ,2 
 * 
 g l ,2 
z  wy
w  (1  h * R i1n h ) 1 h * R i1n
•
•
•
In Rayleigh fading channels, conjugation ensures that, the signal and interference
channel vectors h , g1 , g 2 ,.., g M  are linearly independent with probability 1
Linear independence ensures that MMSE provides full IC up to 2N-1 interferers
MMSE filter is applied to each RB independently
• In CDR mode, allow interference and suppress it
CDR Frequency Partition
CDR Partition
Sector 1
Sector 2
Sector 3
Low SINR users
•
•
•
•
•
High SINR users
In CDR frequency partition
• Allow interference
• Exploit the structure in the interference and suppress it using the noise+plus interference
covariance
The CDR region is dedicated for cell edge users
CDR implementation
•
In a frequency reuse-1 system, the SINR of all the users in a cell are ranked in descending order
•
All cells in the network synchronously allocate the bottom x% users to a CDR frequency partition
which is fixed for the entire network
CDR can be implemented using the existing FFR framework
Operator can define a CDR frequency partition
CDR-MMSE Pilot Processing
•
Use BPSK pilots and preferably pilot-on-pilot mode.
– Step 1: Collect complex, and complex conjugate copies of the received pilots
M
g x
y p  hx p 

l
p ,l
l 1



Pilot tones
 np

Noise
Pilot tones of CCI
– Step 2: Estimate channel coefficients of desired signal using 2D-MMSE
• Assume uniform power-delay-profile
• Knowledge of interferer pilot sequences improves 2D-MMSE
– Step 3: Subtract signal contribution from the received samples
M
e p   g l x p ,l  nˆ p
l 1
– Step 4: Estimate noise-plus-interference covariance matrix

*
ˆ
R
in  E e p e p

• Each RB can be processed independently to obtain high IC gain
Conjugate Data Repetition
• With N rx antennas, CDR provides 2N observations
– CDR doubles the number of copies of signal
• Full interference suppression up to 2N-1 interferers
• Typically 3-4 dominant interferers in re-use 1:1 systems
– Use POD, CDR, MMSE for cell edge users
• With 2-receiver antennas CDR-MMSE nulls 3-interferers
• CDR does not require active cooperation between BSs
• CDR does not require channel estimates of interferers
Conjugate Data Repetition Contd.
• In 16m DL, basic unit for transmission is 1 RB
• In the CDR region
– Data is repeated in conjugate pairs in each RB
– CDR RBs can be allocated in localized, block
distributed or tone-wise distributed modes
– Different sectors can use different permutations
– CDR receiver processes each RB independently
Conjugate Data Repetition Contd.
• Compatibility with other cell edge features
– Works with CL-diversity
– Best band scheduling feasible
– Restrict MIMO modes to
• Rank-1 pre-coding (2D-POD) and CL-beam forming
CDR Requirements
• Define a CDR resource block structure
• Pilot support
– 16 pilots per RB
– Pilot-on-Pilot
• MIMO support
– OL Rank-1 precoding
– CL-diversity
• CDR can be implemented using the FFR framework
– The operator can define a CDR frequency partition
CDR Implementation
• Define a CDR frequency partition
• Allocate common control channels in CDR partition
• Allocate low SINR cell edge users in the remaining CDR
partition
• CDR frequency partition size can be decided by the operator
based on QOS requirements
Link Level Performance
Evaluation of CDR
CDR-POD with MMSE, max: 4-interferers (12 pilots/RB)
•Simulation Assumptions
•PED-Bchannel
•Localized allocation
•RB size 18x6
•½ QPSK with 360 bits
•2Tx 2Rx
•2D-POD transmit diversity
•Total pilot density 11.11%
•Quasi-orthogonal pilots
•Interference power profile
•[0 -3 -6 -9] dB
•Rest of the interference is
modeled as AWGN
•Conjugate repetition
factor=2
•SNR is defined as signal to rest
of the interference power
•Total SINR=-3 dB with 4-interferers
•Good suppression up to 3-interferers, partial suppression
with4-interferers
Link Level Performance
Comparison of CDR and Bit-level
Data Repetition
Simulation Assumptions
•
•
•
•
•
•
Antenna scheme : 4 x 2, POD
Resource : Distributed LRU
Channel Model : eITU PED-B
Interference Profile
– Typical reuse-1 C/I profile = [0 3 6 9 12 14 14 15] ==> -2.75dB
– Typical reuse-1 Cell edge C/I profile=[0 0.6 2 3.4 4.6 6.7 8.3] ==> -5.5dB
– The profile consists of Signal to Interference Power Ratio up to 8-dominant interferers in
ascending order
BLER is evaluated as a function of SRIR which is defined as:
SRIR= Signal Power/(Noise+Residual interference power excluding the dominant 8-interferers)
*In interference limited networks, operating SRIR is usually in 12-15 dB range
SINR= Signal Power/(Noise+Total interference power)
Channel Estimation – 2DMMSE within RB (PRU)
Typical reuse-1 Interference Profile
Typical reuse-1 cell edge interference profile
System Level Performance
Evaluation of CDR
System Level Simulation for Cell Edge Users
CDF of Spectrum Efficiency
•Simulation Assumptions
•PED-A channel
•Localized allocation
•RB size 18x6
•2Tx 2Rx
•Adaptive MCS selection,
includes all MCS modes defined
in 16e
•Results are obtained with a system
level simulator in which actual link
simulation is run for 1000 frames for
each user. The CDF is obtained by
randomly selecting 150 cell edge
users
51% gain
75% gain
•Cell edge throughput gain of CDR 50-75%
System Level Simulation for Cell Edge Users
CDF of BLER for ½ QPSK with repetition factor =4
•BLER of CDR < 0.1 for 96% of users
•BLER of conventional system is < 0.1 for 48% users
•Doubles the control channel coverage
System Level Simulation for Cell Edge Users
CDF of BLER for ½ QPSK with repetition factor =6
•BLER of CDR < 0.05 for 98% of users
•BLER of conventional system is < 0.05 for 40% users
•The gain in control channel coverage is significant
Conclusions
 Due to limited spectrum, frequency re-use 1:1 is the expected deployment
mode in India
 Very important to improve cell edge performance
 CDR and rank-1 transmission (POD/CL-div) ensures high cell edge
performance
 CDR can be implemented along with FFR in a frequency partition
 High cell edge throughput
 Reliable control channel decoding
 Use CDR for both common control channel and cell edge data users
 Allocate cell edge users in a CDR frequency partition
 Define a CDR RB structure
 MIMO mode: Use OL or CL rank-1 precoding
Proposed SDD Text
• Add section 20.5
• 20.5 Interference mitigation region for cell
edge users
– An interference mitigation region for cell edge
users where schemes such as conjugate-datarepetition can be used