Pilot specification for OL rank-1 region

Document Number: IEEE C80216m-09_1915

Date Submitted: 2009-08-29

Source:Kiran Kuchi, J. Klutto Milleth, Padmanabhan M S Voice:CEWiT, India E-mail:

kkuchi@cewit.org.in, klutto@cewit.org.in

Venue: Comments on P802.16m/D1

Base Contribution: None

Purpose: To discuss and adopt the proposed text in section 15.3.5.4.1

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 >.

Background

For a reuse 1 system, the strongest

interference for a cell edge user comes

from those sectors with sector numbers

different from its desired one

The SINR seen by the pilot symbols can

be significantly improved by avoiding

pilot to pilot collisions between sectors

Improves Channel Estimation

Improves Covariance Estimation

Interlaced Pilots

Interlaced pilots use pilot patterns which are cyclically shifted, resulting in

the collisions between pilot and data.

But, the use of interlaced pilots (with power boosting) has the following

disadvantages :

It reduces the data SINR.

Interference suppression receivers can not function smoothly, because of the mismatch between

the estimated covariance and the actual covariance in the data locations that do not have power

boosting

With CDR encoding, conjugate data pairs collide with pilots. Therefore, the number of samples

available for interference covariance estimation is half – compared to pilot-on-pilot case.

In general, poor CDR performance with interlaced pilots

Pilot on Pilot

An option for CDR is to use 2-stream pilots in pilot-on-pilot mode: Total 12 pilot tones

Good quality interference covariance estimation requires very high quality channel estimates

The pilot sequence used by different base stations determine the quality of channel and

interference covariance estimates

Poor CDR performance with PRBS pilots in pilot-on-pilot mode. Need to use pilot sequencers

with low-cross correlation properties.

Re-use 1 needs 21 pilot sequences with low-cross correlation

12-pilots tones not enough to design 21 pilot codes

The number of active interferers may vary from RB to RB.

Pilots from neighboring RBS cannot be used improve interference

covariance estimates

Collision Free Interlaced Pilots(CoFIP)

Interference to the pilot tones from the data tones of the other sectors is completely avoided.

A sector transmits null tones at the pilot locations of the other sectors.

The accuracy of channel/interference covariance estimation improves because:

the pilot tones are almost interference free (only weak higher tier interference)

The signal received on the null tones can be used

to estimate the covariance or to explicitly

Also, possible to estimate two dominant interferers

Implementation Complexity

Reliable channel estimation and interference covariance estimation with single RB processing

Channel estimation complexity same as that of conventional 2D-MMSE

Interference covariance estimation using null tones and pilot tones

Very high CDR performance

Sector 0

0 P 0 0 P 0

P 0 0 P 0 0

0 0 P 0 0 P

P denotes pilot tone and 0 denotes null tone

P 0 0 P 0 0

0 0 P 0 0 P

0 P 0 0 P 0

Previously proposed CoFIP Pattern

0 0 P 0 0 P

0 P 0 0 P 0

P 0 0 P 0 0

Sector 1

Sector 2

•CDR requires subcarriers in pair, which is not possible with this pilot structure• Needs even number of data subcarriers per OFDM symbol

Sector 0

0 P 0 0 P 0

0 P 0P 0 0

0 0 P 0 0 P

P denotes pilot tone and 0 denotes null tone

P 0 0 P 0 0

0 0 P0 P 0

0 P 0 0 P 0

The new CoFIP Pattern for irregular sub frames with 6 OFDM symbols

0 0 P 0 0 P

P 0 00 0 P

P 0 0 P 0 0

Sector 1

Sector 2

•Even number of data subcarriers in every OFDM symbols in each PRU• Enables simple subcarrier mapping for CDR with all advantages of CoFIP

Sector 0

0 P 0 0 P

0 P 0 0P 0 P 0

0 0 P 0 0

P denotes pilot tone and 0 denotes null tone

P 0 0 P 0

0 0 0 P0 P 0 0

0 P 0 0 P

The new CoFIP Pattern for irregular sub frames with 5 OFDM symbols

0 0 P 0 0

P 0 P 0

0 0 0 P

P 0 0 P 0

Sector 1

Sector 2

•For 18 X 7 RBs, add the first OFDM OFDM symbol after 6th OFDM

Estimation of interference covariance

The interference covariance matrix is estimated using all

the symbols received at the pilot/null locations

The covariance is estimated from the pilot tones

corresponding to the same sector number, and from the

null tones of different sector numbers

The CQI estimates are very stable

Channel estimation and interference covariance estimation

confined to 1 RB processing

Pilot sequences for CoFIP

The pilot sequences assigned to different cells should

have low cross correlation within the space of an RB.

Some of the possibilities are:

Use of short pilot sequences, which have low cross-correlation within a

localized resource unit.

Use of long PN sequences

Results in increased cross correlation

Short PN sequence

Seven PN sequences are generated using an LFSR with the characteristic polynomial X3+X +1

The first 6 samples of a sequence is used for deriving the 6 pilots with in a PRU

Two pilot allocation methods are suggested:

Method 1: The same sequence is used in different PRUs of a cell

Method 2: The sequence used in different PRUs of a cell are different

Definitions:

A cell is a collection of 3 sectors and is assigned unique id = CellID

PRU index ‘s’ in frequency and sub-frame number ‘t’ in time

Method 1

The pilot code used in a cell is C( CellID%7 ), where C(i) is the i th code of the set of codes C, where i=0,1,..6

Some minimal pilot planning is required

c6

c1

c0

c5

c4

c3

c2

Note that the sectorization in the hexagon cell is not shown due to the fact that same PN sequence is used in all sectors in a cell

Method 2

Pilot sequence cycling in time and/or frequency is implemented to

exploit the advantage of interference averaging on the pilots. The

pilot code used in a cell with CellID = k is C( (s+t+k)%7), where C(i) is

the ith code of the set of codes C. i=0,1,…6

cell_ID = 0 cell_ID=1 cell_ID=2

C0 C1 C2 C3 C4

C1 C2 C3 C4 C5

C2 C3 C4 C5 C6

C3 C4 C5 C6 C0

C1 C2 C3 C4 C5

C2 C3 C4 C5 C6

C3 C4 C5 C6 C0

C4 C5 C6 C0 C1

C2 C3 C4 C5 C6

C3 C4 C5 C6 C0

C4 C5 C6 C0 C1

C5 C6 C0 C1 C2

s

t -->

Long PN sequence

Long PN sequences are generated using an LFSR with the

characteristic polynomial X11+X9 +1

The LFSR is initialized as follows:

b0...b6 = (CellID % 128 ) + 1

b7...b10 = (symbol index % 16)

For the ith OFDM symbol, a PN sequence PNi(k) , k=0,1,2,....

(FFTSI-1) is generated

The pilot tone (i,k) ( ith OFDM symbol, kth sub-carrier ) is modulated

with PNi(k)

Pilot planning is not necessary

Comparison of Various Pilot Modes

Pilot-on-Data Pilot-on-Pilot COFiP

Total overhead 5.5% 11.1% 16.6%

Channel estimation complexity

Low Medium Low

Performance Very Poor Poor High

1) Pilot-on-Data is not suitable for CDR2) COFiP needs 5.5% more overhead compared to COFiP 3) In spite of additional overhead, COFiP outperforms pilot-on-pilot significantly

Conjugate Data Repetition (CDR) with CoFIP

Simulation parameters

2Tx antenna – 2Rx antenna

Channel: Modified Ped B, 7 Hz doppler

3 interferers of equal power. One interferer colliding with pilot tones and the remaining two

interferers on null tones (Some what worst case scenario. In practice, the CCI on pilots is much

weaker)

Thermal noise level is 10 dB below the power of one of the interferers.

Twenty 18x6 RBs, 4 in freq. and 5 in time. All estimation is done within a single RB

2D-MMSE did not exploit knowledge of pilot sequence used by the CCI which collides with pilot

tones

The new CoFIP pilot pattern is used.

Performance with Short and long PN pilot sequences are compared.

CDR performance with different short PRBS

CDR Performance comparison between short and log PRBS

Conclusions

COFiP Low implementation complexity Additional overhead gives robustness in design,

reduces implementation complexity, higher overall system performance

Recommends COFiP with short PRBS 7- pilot codes with low-cross correlation with minimal pilot

planning

Pilot-on-Pilot with long PRBS significantly worse than COFiP

Proposed Text OL rank-1 uses collision free interlaced pilots. The index of the COFiP PRU type

used by a particular BS with Cell_ID=k is denoted by pk. The index of the used CoFIP type PRU is determined by the Cell_ID according to the following equation: pk

= mod(k,3)

In the Fig xxxx ‘X’ donates a null tone i.e., no data or pilot tone is transmitted in that location

COFiP Type 0 PRU

COFiP Type 1 PRU

COFiP Type 2 PRU

0 P 0 0 P 0

0 P 0P 0 0

0 0 P 0 0 P

P 0 0 P 0 0

0 0 P0 P 0

0 P 0 0 P 0

0 0 P 0 0 P

P 0 00 0 P

P 0 0 P 0 0

Proposed Text

Pilot Sequence Definition: Define 7- codes: C(j), j=0..6. [TBD] The pilot code used in a cell with Cell_ID = k is

P=C( (s+t+k)%7), where C(i) is the ith code of the set of codes C(i). i=0,1,…6

P=[p(0) p(1) … p(5)] where p(i), i=0,..,5 is the ith pilot symbol in Fig xxx