System Level Performance Evaluation on Multiplexing of USCCH in IEEE...

advertisement
System Level Performance Evaluation on Multiplexing of USCCH in IEEE 802.16m
IEEE 802.16 Presentation Submission Template (Rev. 9)
Document Number:
IEEE C802.16m-08/671r1
Date Submitted:
2008-07-16
Source:
Hyunkyu Yu, Jaeweon Cho, Taeyoung Kim, Mihyun Lee, Jeongho Park,
Heewon Kang, Hokyu Choi , Rakesh Taori
Samsung Electronics
hk.yu@samsung.com, jaeweon.cho@samsung.com
Yi Hsuan, Ping Wang, Hujun Yin
Intel Corporation
yi.hsuan@intel.com, ping.wang@intel.com, hujun.yin@intel.com
Jun Yuan, Sophie Vrzic, Dongsheng Yu, Mo-Han Fong, Robert Novak,
Hosein Nikopourdeilami, Kathiravetpillai Sivanesan, Sang-Youb Kim
Nortel Networks
junyu@nortel.com, svrzic@nortel.com, mhfong@nortel.com
Sungcheol Chang, Hyun Lee
scchang@etri.re.kr
ETRI
Venue:
IEEE 802.16m-08/024, “Call for Comments and Contributions on Project 802.16m System Description Document (SDD)”.
Target topic: “DL Control Structure”.
Base Contribution:
None
Purpose:
To be discussed and adopted by TGm for the 802.16m SDD
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 >.
System Level Performance Evaluation on
Multiplexing of USCCH in IEEE 802.16m
Hyunkyu Yu, Jaeweon Cho, Taeyoung Kim, Mihyun Lee,
Jeongho Park, Heewon Kang, Hokyu Choi, Rakesh Taori
Samsung Electronics
Yi Hsuan, Ping Wang, Hujun Yin
Intel Corporation
Jun Yuan, Sophie Vrzic, Dongsheng Yu, Mo-Han Fong, Robert Novak,
Hosein Nikopourdeilami, Kathiravetpillai Sivanesan, Sang-Youb Kim
Nortel Networks
Sungcheol Chang, Hyun Lee
ETRI
July, 2008
Multiplexing of USCCH and Data
 Options
(1) FDM
(2) TDM
(3) Hybrid FDM/TDM
 Proposed Scheme  FDM
 Rationale: Gain over TDM
• Data throughput gain: 6~30%
• Capacity (# of supportable users) gain: > 80%
Annex 1, 2, 3
• Link level performance gain
 Proposed Change to SDD Text
• Within a sub-frame, control and data channels are multiplexed using FDM.
Both control and data channels are transmitted on logical resource units
(LRU) that span all OFDM symbols in a sub-frame.
Why FDM is better than TDM?
 More efficient Power Sharing
 Higher Throughput
Efficient
power
sharing
 Better Granularity
 Smaller Resource Loss
Control IE #1
Control IE #2
Control IE #3
TDM
Control
• TDM: Power sharing
within USCCH only
FDM
Small
Granularity
possible
IE #3 Control IE #2 Control IE #1
• FDM: Power sharing
between Data and USCCH
No power
sharing
• FDM: E.g. 108 tones (1 SubCH x 6 symbol)
• TDM: E.g. 864 tones (48 SubCHs x 1 symbol)
Resource
Large
loss
Granularity
Annex 1:
System Level Performance Evaluation on
Multiplexing of USCCH in IEEE 802.16m
Hyunkyu Yu, Taeyoung Kim, Mihyun Lee, Jeongho Park,
Jaeweon Cho, Heewon Kang, Hokyu Choi
Samsung Electronics Co., Ltd.
FDM vs. TDM
Performance Metrics
GRANULARITY of Ratio btw
FDM
TDM
Control and Data (1-D MAP
region)
 HIGHER
 Lower (especially for short-length sub-frame)
COVERAGE (Outage)
 BETTER
 Worse
SPECTRAL EFFICEINCY
 BETTER
 Worse
 BETTER
 Worse
(Sector Throughput)
CHANNEL ESTIMATION
 Shorter
Processing Time (Latency)
 Longer
 TRADE-OFF between CH. est. performance and
benefit of latency
 Support
Power Saving: Micro-sleep (in
one Sub-frame)
 Not support
 TRADE-OFF between CH. est. performance and
benefit of micro-sleep
 NOT significant gain within a sub-frame (<3%)
[Annex]
Comparisons btw TDM and FDM (1)
 Performance Metric
• With fixed orthogonal resource overhead, How many users can be
supported with satisfying MAP outage requirement (<3%)?
• MAP outage is controlled by ∆MARGIN
Availability (%) = 100 – MAP outage
100
100.0
100.0
98.61
97.47
MUX
Orthogonal Resource
Overhead
FDM
16.7%
TDM
16.7% (1 OFDMA symbol)
97%
*16.7%: Enable to support Maximum
DL8 UL8 assignment blocks
AVAILABILITY (%)
95
89.37
90
87.20
TDM
85
• Even if ∆MARGIN is increased, TDM
cannot support more than DL3, UL3
users
80
75
70
6
9
DELTAMARGIN (dB)
FDM: DL 8 UL 8 users
6
9
6
9
DELTAMARGIN (dB)
DELTAMARGIN (dB)
TDM: DL 3 UL 3 users
TDM: DL 4 UL 4 users
FDM
• Enable to support DL8, UL8 users
without change of resource OH
Comparisons btw TDM and FDM (2)
 Performance Metric
• Maximum SECTOR THROUGHPUT with
satisfying MAP outage requirement (<3%)
MUX
• GRANULALITY of resource ratio between
data and control
• FDM: Resource + Power (SOFT Separation)
FDM
SECTOR THROUGHPUT (Mbps)
• TDM: Resource (HARD Separation)
8.0
TDM
7.0
# of Users
(DL, UL)
Orthogonal
Resource
Overhead
∆MARGIN
(2, 2)
8.3%
6dB
(3, 3)
8.3%
5dB
(4, 4)
8.3%
5dB
(5, 5)
16.7%
5dB
(2, 2)
16.7%
4dB
(3, 3)
16.7%
4dB
(4, 4)
33.3 %
2dB
(5, 5)
33.3 %
4dB
6.0
5.0
(2, 2 )
(3, 3)
(4, 4)
NUM. of (DL, UL) TX. USERS
FDM
TDM
(5, 5)
FDM yields BETTER
throughput performance
than TDM
System Level Simulation (1)
 Performance Metrics
• Sector Throughput with satisfying MAP outage requirement
• MAP Outage requirement: Distribution of user whose BLER is larger than
1% < 3% of total users
 Per User Power Control
• PMAPIE[i] = SINRREQ– SINR(CQI)[i] + ∆MARGIN
• SINRREQ: SINR value required to satisfy 1% BLER
• SINR(CQI)[i]: i-th user SINR set by CQI feedback value
• ∆MARGIN: Margin value to accomplish required MAP outage
System Level Simulation (2)
 Major Assumptions
• Subframe structure
• 48 bits (including CRC) per assignment
block
• 1-D MAP region indication
Assign.
#1
Assign. Assign.
#N
Assign.
#2
...
...
Assign.
...
• Separate coding
Assign. Assign. Assign.
#1
#2
DATA
Traffic
DATA
Traffic
NRU,TOTAL
• Only assignment block in MAP region
1/12 or 1/6 ⅹNRU,TOTAL
• [IEEE C802.16m-08/062r1]
1 or 2 OFDMA symbols
Assign.
#N
• Period: Semi-static
MUX
Orthogonal Resource
Overhead
FDM
8.3 or 16.7 %
TDM
16.7 or 33.3 %
Subframe (6 symbols)
FDM
Zero insertion
TDM
* 8.3%: Maximum DL4 UL4 assignment blocks
16.7%: Maximum DL8 UL8 assignment blocks
33.3%: Maximum DL16 UL16 assignment blocks
System Level Simulation (3)
 Simulation Environments/Assumptions
Index
Deployment Scenario
MCS for MAP
HARQ
Scheduler
Value
EMD baseline [IEEE 802.16m-07/037r2 ]
QPSK, 1/2
Synchronous
(No assignment message for retransmission)
Proportional fairness
# of Users per Sector
10
# of Scheduled Users
2, 3, 4, 5 per mini-frame
(4, 6, 8, 10 for both DL and UL)
MAP Error Effects
Antenna Configuration
Channel Model
Channel Estimation
Other Simulation Assumptions
Resource loss for MAX retransmission
SIMO 1x2
Mixed (Ped B-3kmph-60%,
Veh A-30kmph-30%, Veh A-120kmph-10%)
Real channel estimation
(Equal impairment for both TDM and FDM)
EMD baseline
Power Saving (1)
 Micro-Sleep (within a sub-frame)
• Symbol level power saving
 Power Saving Gain
3.5  Max( p, q )  xi
 zi  Gi %
6
i 0
2
PSG  
= 2.04 %
Power saving gain by Micro-sleep is
NOT significant
Required Time
FFT
1 symbol
MAP Region
p symbols
p=1
Pilot Region
q symbols
q=2
CH. Est. Delay
0.5 symbol
Baseband Modem
z0
MAP Decoding
Minimum 1 symbol
RF Parts
Turn-off + Turn-on
x0, x1, x2
Display Device
Parts
Portion of
Power
Consumption
Time for turnoff + turn-on
Power
Saving
Gain (%)
0.1
x0
1 symbols
G1
50
z1
0.65
x1
1 symbols
G2
30
z2
0.25
x2
-
G3
-
Power Saving (2)
 Default Subframe Concept
• Sub-frame level power saving
• Power saving gain can be much larger than Micro-sleep
• One of sub-frame is pre-assigned to a MS as a default sub-frame,
then the MS may go sleep mode during other sub-frame
DL subframe
UL subframe
DL MAP
UL MAP
DL
Subframe
#2
ACK for UL
DL
Subframe
#3
DL
Subframe
#4
UL
Subframe
#0
DL Data
MSi
Power Off
DL Data
UL
Subframe
#2
ACK for DL
DL MAP
UL MAP
ACK for UL
DL
Subframe
#3
UL
Subframe
#1
UL Data
Power Off
UL
Subframe
#1
ACK for DL
Power Off
RTG
DL
Subframe
#1
TTG
BS
DL
Subframe
#0
UL Data
Annex 2:
Evaluation Results of 802.16m
DL Control Structure
Yi Hsuan, Ping Wang, Hujun Yin
Intel Corporation
Multiplexing Criteria II: Coverage Analysis
866m Cell (EMD Baseline Configuration)
Test Scenarios
QPSK ½ co QPSK ¼ co QPSK 1/8 c QPSK 1/12
verage
verage
overage
coverage
Outage
FDM,
Ped B
Reuse 1, CDD
46.8%
67.2%
93%
98.4%
1.6%
TDM,
Ped B
Reuse 1, CDD
44.5%
64.4%
89.3%
96.5%
3.5%
FDM,
Veh A
Reuse 1, CDD
44.2%
62.8%
89.4%
96.2%
3.8%
TDM,
Veh A
Reuse 1, CDD
43.1%
61.6%
85.4%
94.7%
5.3%
Multiplexing Criteria II: Coverage Analysis
5000m Cell with Open Rural Macrocell Pathloss
Model
Test Scenarios
QPSK ½
coverage
QPSK ¼
coverage
QPSK 1/8
coverage
QPSK 1/12
coverage
Outage
FDM,
Ped B
Reuse 1, CDD
21.6%
43.9%
73.8%
87.2%
12.8%
FFR, CDD
45.7%
80.2%
96.5%
100%
0
TDM,
Ped B
Reuse 1, CDD
19.2%
41.1%
66.3%
81.3%
18.7%
FFR, CDD
42%
79.6%
96.5%
99.7%
0.3%
FDM,
Veh A
Reuse 1, CDD
18.9%
38.9%
66.7%
79.9%
20.1%
FFR, CDD
43.5%
77.6%
96.4%
99.6%
0.4%
TDM,
Veh A
Reuse 1, CDD
18.4%
36.6%
60.5%
77%
23%
FFR, CDD
44.8%
78.1%
94.6%
99.2%
0.8%
Multiplexing Criteria II: Coverage Analysis
5000m Cell with Baseline Pathloss Model
Test Scenarios
QPSK ¼ co
verage
QPSK 1/8 c
overage
QPSK 1/12
coverage
Outage
FFR, CDD
24.7%
No power boost
38.9%
54.2%
66.2%
33.8%
FFR, CDD
3 dB boost
31.5%
48.7%
65.6%
78.9%
21.1%
TDM,
Ped B
FFR, CDD
23.6%
37.8%
51.7%
63%
37%
FDM,
Veh A
FFR, CDD
23.6%
No power boost
35.9%
52%
62.3%
37.7%
FFR, CDD
3 dB boost
30.4%
45.7%
64.3%
74.7%
25.3%
FFR, CDD
23%
35.3%
48.4%
59.9%
40.1%
FDM,
Ped B
TDM,
Veh A
QPSK ½ co
verage
Multiplexing Summary
 FDM has better link level performance and coverage than TDM because
of channel estimation advantage for FDM.
 FFR improves coverage significantly in all cell sizes
 FDM allows power boosting for USCCH, which can improve the cell
coverage in some cases.
Annex 3:
Performance Evaluation for IEEE 802.16m
Downlink Control Structure
Jun Yuan, Sophie Vrzic, Dongsheng Yu, Mo-Han Fong,
Robert Novak, Hosein Nikopourdeilami,
Kathiravetpillai Sivanesan, Sang-Youb Kim
Nortel Networks
Link Level Simulation Assumptions
Table 1.1: Link Level Simulation Parameters
Parameters
Values
Bandwidth
10 MHz
FFT size
1024
Carrier Frequency
2.5 GHz
Channel Model
Pedestrian B 3 km/hr,
ITU-Vehicular A 120 km/hr,
DL Tx scheme
2 Tx antenna, STBC
DL Rx scheme
2 Rx antenna
Permutation and
symbol structure
16e PUSC (baseline permutation in EMD)
Channel Coding
16e CTC
MCS
QPSK ½ with repetition 0, 2, 4 and 6.
Channel Estimation
MMSE based on all pilots in 2 symbols for TDM and 6 symbols f
or FDM
TDM
FDM
Figure 1: Pilot Design
Pilot for antenna 1
Pilot for antenna 2
Link Level Performance Results (1/2)
10
TDM QPSK, STTD 2x2, PB 3km/h
0
10
1/12 ideal
1/12 est
1/8 ideal
1/8 est
1/4 ideal
1/4 est
1/3 ideal
1/3 est
1/2 ideal
1/2 est
2/3 ideal
2/3 est
-1
BLER
10
-2
-15
-10
-5
0
5
10
10
10
TDM QPSK, STTD 2x2, VA 120km/h
0
1/12 ideal
1/12 est
1/8 ideal
1/8 est
1/4 ideal
1/4 est
1/3 ideal
1/3 est
1/2 ideal
1/2 est
2/3 ideal
2/3 est
-1
-2
-15
-10
-5
SNR
10
10
FDM QPSK, STTD 2x2, PB 3km/h
0
10
1/12 ideal
1/12 est
1/8 ideal
1/8 est
1/4 ideal
1/4 est
1/3 ideal
1/3 est
1/2 ideal
1/2 est
2/3 ideal
2/3 est
-1
10
-2
-15
-10
-5
0
SNR
0
5
10
SNR
BLER
10
BLER
BLER
10
5
10
10
FDM QPSK, STTD 2x2, VA 120km/h
0
1/12 ideal
1/12 est
1/8 ideal
1/8 est
1/4 ideal
1/4 est
1/3 ideal
1/3 est
1/2 ideal
1/2 est
2/3 ideal
2/3 est
-1
-2
-15
-10
-5
0
SNR
5
10
Link Level Performance Results (2/2)
Table 2: MCS SNR at 1% BLER
Code Rate
TDM PB3km/h
TDM VA120km/h
FDM PB3km/h
FDM VA120km/h
QPSK ½ rep 6
-2.3188 (dB)
-1.5222 (dB)
-4.1482 (dB)
-3.794 (dB)
QPSK ½ rep 4
-1.7772 (dB)
-0.9984 (dB)
-2.8395 (dB)
-2.8067 (dB)
QPSK ½ rep 2
0.3323 (dB)
0.5476 (dB)
-0.5509 (dB)
-0.5688 (dB)
QPSK ½
3.5882 (dB)
3.4901 (dB)
3.0496 (dB)
2.9522 (dB)
System Level Simulation Assumptions for
Coverage Evaluation
Table 1.2: System Level Simulation Parameters
Parameters
Values
BS-to-BS distance
1.5km urban
5.0km open rural microcell NLOS
Frequency reuse
Reuse-1
Transmission power/sector
46 dBm
BS height
32 m
Tx antenna pattern
70o (-3dB) with 20 dB front-to-back ratio
Tx antenna gain
17 dBi
MS height
1.5 m
Rx antenna pattern
Omni directional
Rx antenna gain
0 dBi
MS Noise Figure
7 dB
Penetration loss
10 dB
Hardware losses (Cable, implementation, etc.)
2 dB
Lognormal shadowing
=0 dB, σSF =8 dB
Shadowing correlation
100% inter-sector, 50% inter-BS
Coverage Performance
1
0.9
0.8
0.7
CDF
0.6
1.5km urban: 95% SNR = -3.70dB
5.0km rural: 95% SNR = -5.88dB
0.5
0.4
0.3
0.2
0.1
0
-10
geo 1.5km Bs2Bs (sub)urban micro-cell
geo 5.0km Bs2BS rural macro-cell
-5
0
5
Geometry (dB)
10
15
20
Table 3: MCS for 95% cell-edge users
TDM
FDM w/o power boost
FDM with 3dB power boost
1.5km PB3
Not Supportable
QPSK ½ rep 6
QPSK ½ rep 4
1.5km VA120
Not Supportable
QPSK ½ rep 6
QPSK ½ rep 4
5.0km PB3
Not Supportable
Not Supportable
QPSK ½ rep 6
5.0km VA120
Not Supportable
Not Supportable
QPSK ½ rep 6
System Level Simulation Assumptions
for Capacity Evaluation
Table 4: Simulation Assumptions
Parameters
Values
Resource budget
30 slots
TDM USCCH region
30 subchannels by 2 symbols
FDM USCCH region
10 subchannels by 6 symbols
Other hybrid schemes
Total data subcarriers in the region should be the same as the above TDM or FDM
scheme.
Mininum resource unit
Per MAP IE size
Power budget
46 dBm for TDM
41.2 dBm for unboosted FDM (10 out of 30 subchannels are used so power budget
should be 46-10log10(3)=41.2)
44.2 dBm for 3 dB boosted FDM
Possible MCS
QPSK ½, QPSK ½ repetition 2, QPSK ½ repetition 4, QPSK ½ repetition 6
Code scheme
Separate encoding
CID size
0 bit (masked by CRC)
Start RB index
6 bits (or proposal specific value)
Allocated RB
5 bits (or proposal specific value)
Other L1/L2 information (data MCS etc.)
x (5, 21,37)
CRC
16 bits
Total MAP IE sizes
32, 48, 64 bits (including CID, RB allocation and other L1/L2 information, and CRC)
Power Sharing
Yes for TDM and FDM
BS-to-BS distance
1.5km
Capacity Results
Table 5: Number of Supportable Users
PB 3km/h channel, 30 slots resources , at BLER = 1% realistic channel estimation
1.5km TDM
1.5km FDM
(gain)
1.5km FDM Power Boost 3
dB
(gain)
32
24
27 (13%)
41 (70%)
48
14
19 (35%)
28 (100%)
64
11
14 (27%)
20 (81%)
sum
49
60 (22%)
89 (81%)
MAP IE size
Table 6: Number of Supportable Users
VA 120km/h channel, 30 slots resources , at BLER = 1% realistic channel estimation
1.5km TDM
1.5km FDM
(gain)
1.5km FDM Power Boost 3
dB
(gain)
32
20
27 (35%)
41 (105%)
48
14
18 (28%)
28 (100%)
64
10
14 (40%)
20 (100%)
sum
44
59 (34%)
89 (102%)
MAP IE size
Summary of Multiplex Schemes
 FDM outperforms TDM
• From coverage perspective
• Cell-edge users are supportable by FDM with power boost
• Cell-edge users are not supportable by TDM
• From capacity perspective
• FDM (with power boost) achieves more than 20% (80%) capacity gain over TDM
 Reasons
• FDM has ~2dB link level gain due to time-direction de-noising.
• FDM has 3dB power boost gain.
Download