Upload
others
View
9
Download
0
Embed Size (px)
Citation preview
出國報告(出國類別:其他)
3GPP RAN1 #72bis Meeting
會議報告
出國人員:郭秉衡、邱頌恩、謝佳妏、鄭延修、邱哲盛
派赴國家:美國/芝加哥
出國期間:102年 4月 13日至 102年 4月 22日
報告日期:102年 5月 17日
2
摘 要
此次 RAN1#72Bis的議程包括 New Carrier Type、eIMTA、Downlink MIMO
Enhancements、3D-MIMO、Network-Assisted Interference Cancellation (NAIC)、
Low-cost MTC、D2D和 Small Cell Enhancements,參加這次會議的目的在於持
續關注 LTE-A Rel-12標準制定的動向及各家公司提案方向,以利於及早布局我
方往後在標準制定工作上的策略方針。另一個主要目的在於更加深入了解針對
一些新技術的模擬環境假設,以便於我們未來模擬平台的建構與發展。
會議解說:
3GPP RAN1 Meeting #72bis會期於 2013年 4月 15日至 2013年 4月 19日
在美國芝加哥舉行。本團隊在本次會議著重於 Device to Device (D2D)
Communications 並關注新的工作項目如 Network Assisted Interference
Cancellation、Low-cost MTC、Small Cell Discovery、NCT。
3
目 錄
摘 要 .......................................................................................................... 2
一、會議名稱 ............................................................................................ 4
二、參加會議目的及效益 ........................................................................ 4
三、會議時間 ............................................................................................ 5
四、會議地點 ............................................................................................ 5
五、會議議程 ............................................................................................ 5
六、會議紀要 ............................................................................................ 7
七、心得與建議 ...................................................................................... 36
4
一、會議名稱
3GPP TSG RAN1 #72Bis Meeting
二、參加會議目的及效益
參與 3GPP RAN1並關注 Rel-12的議題包括 UMTS HetNet、New Carrier
Type、eIMTA、DL MIMO enhancements、3D-MIMO、Low-Cost MTC、
D2D、Network-Assisted Interference Cancellation and Suppression 和 Small
cell Enhancements等,並參與討論以追蹤各項技術項目之進展。
RAN 1目前所討論的議題如下:
New Carrier Type for LTE:
- Study of new DL DMRS patterns
- Standalone NCT: Evaluation of benefits and identification of scenarios
Further Downlink MIMO Enhancements for LTE-A :
- 4-tx codebook
- Other details of new aperiodic PUSCH feedback modeRemaining details of DCI
formats
Further Enhancements to LTE TDD for DL-UL Interference Management and Traffic
Adaptation:
- Interference Mitigation Schemes
- Signalling mechanisms for TDD UL-DL reconfiguration
Study on Provision of Low-cost MTC UEs based on LTE:
- PSS/SSS
- PBCH
- PRACH
- PDSCH/PUSCH
- (E)PDCCH
- PUCCH
5
Study on Small Cell Enhancements : Physical-Layer Aspects
- Remaining aspects of evaluation assumptions
- Enhancements for improved spectral efficiency
- Control Signaling
- Higher-order Modulation
Mechanism for Efficient Small Cells Operation
- Interference avoidance and coordination
- Small cell discovery
Study on 3D-channel model for elevation beamforming and FD-MIMO
- Remaining details of scenarios
- Modifications to the 3GPP evaluation methodology
Study on LTE Device to Device Proximity Services
Study on Network-assisted Interference Cancellation and Suppression for LTE
報告本團隊所發表的文章
發表系統實作所發現的相關議題,增進實作技術和系統概念的交流
與其他大廠接觸以討論合作項目
三、會議時間
15 Apr.- 19 Apr. 2013
四、會議地點
Palmer House Hilton Hotel (Chicago, IL, USA)
五、會議議程
RAN1 的會議議程如下:
6
7
六、會議紀要
會議資訊摘要:
本次會議由 North American Friends of 3GPP 主辦,共約有 324人參加。除
了 RAN1 #72bis會議之外,RAN2 #81bis、RAN3 #79bis及 RAN4 #66bis等
工作組(Working Group,WG)亦同時召開。
台灣派員參與此次會議之公司/機構有 ACER、ASUS、CHTTL、HTC、III、
ITRI及 MediaTek等 7家。
本次會議 UTRA Rel-12 WCDMA/HSPA+相關技術議題共有 5個議程,包括
1個工作項目(Work Item,WI)及 4個研究項目(Study Item,SI)),分別為:
1. Agenda Item 6.2: (WI) HSPA Signalling Enhancements for More Efficient
Resource Usage for LCR TDD
2. Agenda Item 6.3: (SI) Study on UMTS Heterogeneous Networks
3. Agenda Item 6.4: (SI) Study on DCH Enhancements for UMTS
4. Agenda Item 6.5: (SI) Study on Scalable UMTS
5. Agenda Item 6.6: (SI) Study on Further EUL Enhancements
本次會議 E-UTRA Rel-12 LTE/LTE-A 相關技術議題共有 8個議程,包括 3
個 WI及 5個 SI,分別為:
1. Agenda Item 7.2.1: (WI) New Carrier Type for LTE
2. Agenda Item 7.2.2: (WI) Further Downlink MIMO Enhancement for
LTE-AdvancedFurther Enhanced Non-CA-Based ICIC for LTE
3. Agenda Item 7.2.3: (WI) Further Enhancements to LTE TDD for DL-UL
Interference Management and Traffic Adaptation Network-Based
Positioning Support for LTE
4. Agenda Item 7.2.4: (SI) Study on Provision of Low-cost MTC UEs based
on LTE
5. Agenda Item 7.2.5: (SI) Study on Small Cell Enhancements –
Physical-layer Aspects
6. Agenda Item 7.2.6: (SI) Study on 3D-channel model for Elevation
8
Beamforming and FD-MIMO
7. Agenda Item 7.2.7: (SI) Study on LTE Device to Device Proximity Services
8. Agenda Item 7.2.8: (SI) Study on Network-Assisted Interference
Cancellation and Suppression for LTE
本次會期 Rel-12 LTE/LTE-A共有 8個WIs/SIs進行討論,技術議題數目與
前次 RAN1#72 會議相較增加「Study on Network-Assisted Interference
Cancellation and Suppression for LTE (NAICS)」一項 SI。此 SI由MediaTek
領銜於今年 2月底 RAN Plenary #59 Meeting 提出並獲通過,共有超過 30
家公司支持,其中支持的 Operator 有 Orange、Telefonica、CHTTL、US
Cellular、Verizon Wireless、China Telecom、China Mobile、T-Mobile USA、
Deutsche Telekom、Telecom Italia、Sprint、China Unicom、Lightsquared、
Softbank Mobile等 14家。由電信營運商關注程度看來,此 NAICS SI勢將
成為 Rel-12 LTE/LTE-A重要研究發展議題之一。
這次大會中,本團隊主要負責專注 DL MIMO Enhancemen、Elevation
Beamforming and FD-MIMO、D2D、Small Cell Enhacement、UMTS HetNet、
Network-assisted Interference Cancellation (NAIC)等議題的發展概況。本章節將
簡述這幾項議題的最新進度。
1. 『UMTS Heterogeneous Networks』研究項目發展摘要:
具同頻(co-channel)干擾時之評估結果
Huawei於「R1-131482 Summary of email discussions on the alignment of
simulation results」文件,彙整各家公司在 UMTS HetNet同頻佈建且基
於 full buffer假設時之 downlink模擬結果(如表一所示)。
由評估結果可發現,與 macro-only同質網路相較,UMTS HetNet 的確
具有顯著增益,且各家公司結果趨於收斂,因此會議通過將 full buffer
評估結果納入 TR25.800 文件,不過呈現方式尚須討論。Huawei 彙整
UMTS HetNet模擬結果(R1-131482) 如下表所示:
9
Downlink Throughput
Offload
[%]
LPN is
empty
[%]
LPN
serves
one UE
[%]
RX
Type
Mean
[kbps] Gain
Median
[kbps] Gain
5%
[kbps] Gain
Alcatel-Luc
ent
Baseline 261 263 140 Type3
HetNet 1083 316% 368 40% 199 42% 34% 21% 38%
CHTTL
Baseline 391 308 102 Type3i
HetNet 1580 304% 666 116% 135 32% 35% 25% 28%
Ericsson
Baseline 406 354 76 Type3
HetNet 1493 267% 743 110% 145 92% 34% 35% 36%
Huawei
Baseline 612 604 251 Type3i
HetNet 2130 248% 1092 81% 337 34% 31% 26% 35%
NSN
Baseline 365 293 105 Type3i
HetNet 1414 287% 556 90% 147 40% 27% 30% 41%
Qualcomm
Baseline 518 486 220 Type3i
HetNet 1707 230% 827 70% 318 45% 29% 28% 40%
Renesas
Baseline 374 329 120 Type3i
HetNet 1534 310% 646 96% 173 43% 29% 28% 37%
ZTE
Baseline 477 312 81 Type3
HetNet 1237 160% 545 75% 124 53% 28% 27% 26%
此次會議本團隊亦提出模擬結果及看法(R1-131214),我們的見解如下:
在同頻佈建時,HetNet不論在 full buffer 或 bursty情況皆於平均、
50%-tile、以及 5% -tile (i.e., cell edge)用戶流通量(user throughput)
呈現顯著增益。(此看法與大部分公司相同)
增加 LPN (low power node)發射功率與 CIO (cell individual offset)可
改進整體用戶流通量效能。(如下圖所示)
10
結合 CIO及 Multiflow SF-DC (single frequency dual cell)技術可進一
步提升細胞邊緣用戶體驗,同時系統容量亦能維持。(如下圖所示)
2. Further Downlink MIMO Enhancement for LTE-Advanced
關於 DL MIMO Enhancement 的 Working Item在上次會議已經正式開始討
論,主要方向可分為 4-TX 系統下的預編碼碼書(codebook)加強、以及新的
PUSCH反饋模式(feedback mode)以更進一步改善SU-MIMO和MU-MIMO下行
鏈路的傳輸效能。上次會議所達成的Working Assumption為:
Working Assumptions:
• A new aperiodic PUSCH feedback mode is supported in Rel.12 with following feedback
– CQI and rank feedback bit size as in PUSCH Mode 3-1 in Rel 10
– A wideband PMI
• 2 Tx: 0 bit
11
• 4 Tx: FFS with the consideration of codebook enhancement in Rel. 12 not
excluding 0 bit wideband PMI
• 8 Tx: 4/4/2/2/2/2/2/0 bits for rank 1-8 respectively
– Per subband PMI(s)
• 2Tx: 2/1 bits for rank 1 – 2
• 4Tx: FFS with the consideration of codebook enhancement in Rel.12
• 8Tx: 4/4/4/3/0/0/0/0 bits for rank 1 – 8 respectively
• Use Rel 10 W=W1W2 codebook structure for 4 antenna feedback for DMRS based TMs
• The following are for further study and evaluation:
– Subband size
– Detailed W1 and W2 structures, e.g. W1 corresponds to a long term and/or
wideband channel properties and W2 corresponds to a short-term and
narrowband channel
– Additional information in the CSI reports for this new feedback mode
For example CSI feedback enhancements targeted at improving MU
performance
(Note that the current Rel-10 4tx codebook has W=W1W2 structure).
在此 Working Item 之下,細分兩個 agenda items:
4-TX codebook
Other details of new aperiodic PUSCH feedback mode
在 codebook 的部分,此次會議前多家公司已透過 email discussion 分享各
自的 codebook提案,以方便各方模擬並比較其效能。因此,這次會議中大多數
的 contributions都是以呈現模擬結果為主。討論中較有爭議的問題在於:
1. Rel-8 codebook should be included in the new codebook for rank 1-2:
Yes: TI, Intel, ZTE,
No: Broadcom, AT&T, Samsung, Ericsson, ST-Ericsson, Alcatel-Lucent, Alcatel-Lucent
Shanghai Bell, Qualcomm, CATT, Nokia, NSN, LGE,
2. The new codebook should use constant modulus or Non-constant modulus:
Constant modulus: Samsung, Alcatel-Lucent, Alcatel-Lucent Shanghai Bell, AT&T,
Broadcom, NSN, Nokia, Renesas, Ericsson, ST-Ericsson, Fujitsu, CATT
Non-constant modulus: Docomo, Motorola, Huawei, HiSilicon, Intel, NEC
12
最後討論基於由 Alcatel-Lucent 主導的 Way Forward (R1-131719),對於
codebook的設計結構大會達到了以下的 agreement:
Agreement:
• Adopt a new 4 tx codebook at least for rank 1-2
Working assumption for rank 1-4:
• New 4 tx codebook is constant modulus
– Companies can provide further evaluations for Scenario C1 until RAN1#73 if
desired
Working assumption:
• At least for rank 1-2:
– 4Tx codebook enhancement is based on a grid of beams design similar to the Rel
10 8Tx codebook
• (this does not preclude that some of the Rel-8 codewords might be
included in the new codebook)
– a same W1 codebook is assumed for rank 1 and rank 2 feedback
• FFS until RAN1#73 whether rank 3-4 is:
– same as Rel-8 or
– enhanced or
– not supported in the new codebook
• If rank 3-4 is supported in the new codebook, a same W1 codebook is assumed for rank 3
and rank 4 feedback
Working assumption for structure of W1:
• For each rank R =1,2 (and 3 and 4, if 4Tx codebook enhancement of rank 3 and 4 is
supported in Rel 12), W1 is defined as below
1,..,1,0,0
011
Nn
n
n
X
XW
nRC,n,n aaanqqq ,21
111
111
X
Where Xn is a 2xCR matrix with DFT columns and 12
1
πj/Qeq
• To be determined:
– N1: total number entries of W1 per rank
– Q1: granularity of beam of W1
– For each block matrix Xn
• CR: total number of beams
• a1,n, …, aCR,n : beam coefficients determining specific directions within
13
the block
Working assumption for structure of W2:
• For each rank R =1,2 (and 3 and 4, if 4Tx codebook enhancement of rank 3 and 4 is
supported in Rel 12), W2 is a 2CR x R matrix defined as below
RaaW 12
2
2
1
/2
2121
22
12 and ,...,, where Qj
crr
r
m
r
m
r eqq
qr
r
eeyy
y
ya
where ei is a selection vector of zeros and a “1” in the ith row
• To be determined:
– N2: total number entries of W2 per rank
– Q2: co-phasing granularity between two polarizations
– For each ar
• yr1: column selection for the first polarization
• yr2 : column selection for the another polarization
• mr1: phase shift for the first polarization
• mr2: phase shift for the second polarization
Working Assumption of Codebook Parameters of Rank 1 and Rank 2
• W1 (4 bits)
– N1 16
– Q1, q1: FFS
– For each block matrix Xn
• CR: FFS between 2 and 4 (4 is the default)
• a1,n, …, aCR,n : FFS
• W2 (3 or 4 bits) and three alternatives for column selection (assuming CR=4):
– Alt.1 (4 bits) same as 8Tx W2
– Alt.2 (3 bits) subset of 8Tx W2
– Alt.3 (4 bits) as 8Tx W2 modified to
• For rank 1
2
1 ,
2
1 ,
2
1 ,
2
1
2
1
2
1
2
1
2
1
22Y
Y
Y
Y
Y
Y
Y
YW
jjC 432121
~,~,~,~ , eeeeYY
• For rank 2
2
1,
2
1
22
11
22
11
22YY
YY
YY
YYW
jjC
– Alt.4 (4 bits) as Alt 3 with independent Y1, Y2, Y3, Y4 (Mihai to provide more
detailed description by Friday)
14
Working assumption for rank 3,4 (if new codebook is agreed for rank 3-4):
- W1: 2 bits
- W2: 4 or 3 (FFS) bits for rank 3, 3 bits for rank 4
在 PUSCH feedback mode 方面,主要討論該模式下應反饋的基本內容,包
括 wideband CQI、subband CQI、wideband PMI和 subband PMI等等。其中,各
家公司代表對於 feedback mode 3-2 中 PMI 回報的 subband size 定義有不同看
法。一般來說,較小的 subband size表示較高的 frequency granularity,理論上
能比較有效地讓傳送端評估頻域上的通道變化而提供較好的效能,但也同時意
味著較大的反饋量。部分公司認為該 subband size應與 mode 3-1的 CQI回報相
同,也有公司認為 subband size 可用 RRC訊號來動態調整。最後的共識如下:
Agreements:
• A new aperiodic PUSCH feedback mode is supported in Rel.12 with following feedback:
– A wideband CQI: 4 bits
– Per subband differential CQI with respect to wideband CQI as PUSCH 3-1 : 2 bits
– A wideband PMI based on W1 codebook
• 2 Tx: 0 bit
• 4 Tx:
– 0/0 bits for rank 1-2 respectively if Rel 8 codebook is configured
to the UE
– FFS for rank 3-4
– FFS if Rel 12 dual codebook is configured to the UE
• 8 Tx: 4/4/2/2/2/2/2/0 bits for rank 1-8 respectively
– Per subband PMI(s) based on W2 codebook
• 2Tx: 2/1 bits for rank 1 – 2 based on Rel 8 2Tx codebook
• 4Tx:
– 4/4 bits for rank 1-2 respectively if Rel 8 codebook is configured
to the UE
– FFS for rank 3-4
– FFS for rank 1-4 if Rel 12 dual codebook is configured to the UE
• 8Tx: 0/0/0/0 bits for rank 5 – 8 respectively
– FFS until RAN1#73 for rank 1-4 with codebook sub-sampling
– Subband Size
15
– Working assumption that the CQI and PMI subband sizes are the same
– Revisit if a problem emerges when finalising the feedback report
sizes
– Study further until RAN1#73 whether there is a benefit from
allowing different CQI and PMI subband sizes
– Working assumption that the existing CQI and PMI subband sizes are used
– Study further until RAN1#73 whether there is a benefit from
RRC-configurable subband sizes
– Additional information in the CSI reports is FFS
– For example CSI feedback enhancements targeted at improving MU
performance
• Confirm working assumption of using Rel 10 W=W1W2 codebook structure for 4
antenna feedback for DMRS based TMs
• The new aperiodic PUSCH feedback mode 3-2 can be only configured for DMRS-based
transmission mode 8,9, and 10 when PMI/RI reporting is configured;
• Working assumption that the aperiodic PUSCH feedback mode 3-2 can be configured for
TM4,6 when the Rel-8 codebook is used
• Revisit at RAN1#73 if a problem is found with TM4,6, or if no gain is found.
• A new Rel-12 codebook is not applicable to TM4,6
• Assuming a new 4Tx codebook is introduced, it should be supported for all aperiodic
reporting modes that are valid for TMs 8,9,10 when PMI/RI reporting is configured and
periodic feedback modes 2-1 and 1-1
• Assuming a new 4Tx codebook is introduced, RRC configuration per CSI process
determines whether the UE uses the Rel-8 4Tx codebook or the Rel-12 dual
codebook.
會議中也能針對 MU-MIMO的 Feedback Enhancements 做了部分討論,但
由於時間的關係,並沒有特別明確的決議。
3. Study on 3D-channel model for Elevation Beamforming and FD-MIMO
關於 3D-MIMO 的 study item 同樣也是上次會期正式開始探討的新議題。
與傳統MIMO技術相比,3D-MIMO多了 elevation的自由度,因此,該技術重
點為透過主動式天線系統(active antenna systems 或 AAS)及二維的天線配置將
信號能量更精準地聚焦於目標方向,以提升信號品質,並降低對其他用戶所造
成的干擾。
16
此次會議的重點在於模擬平台的環境假設及參數設定,包括 UE 在服務範
圍內的分布,接收端高度的設定,基站天線的高度以及天線的假定模型等等。
大會達成了以下決議:
Agree on the proposals in the two tables plus the following
• Working assumption on number of floors (for UE height):
– Uniformly distributed with an average and variation range
– Average number of floor: 6 for both UMa and UMi
– Variation range: [-2 to 2]
– Additional values or adjustments can be FFS as needed
• For a given UE, the UE height calculation has a two-step process:
– Follow the above-mentioned distribute to find the number of floors x,
– The actual floor the UE is on follows the uniform distribution of [1,x],
• Heterogeneous Networks
– Channel models developed for Urban Micro cell with high UE density and Urban Macro cell
with high UE density scenarios shall support heterogeneous deployment scenarios.
– It is assumed that for heterogeneous deployment scenarios the macro BS
height is at 25m and the lower-power node is at 10m height.
- Assumptions in Table 2 are for calibration purposes only in this SI
- Assumptions in Table 2 to be revisited for evaluating relative performance of proposed
solutions in future Sis
- *Alternative value = 200m (FFS)
Table 1:
Urban Micro cell
with high UE
density
Urban Macro cell
with high UE
density
Layout Hexagonal grid, 19 micro
sites,3 sectors per site
Hexagonal grid, 19 macro
sites,3 sectors per site
UE mobility
(movement 3kmph 3kmph
17
In horizontal plane)
BS antenna height 10m 25m
Total BS Tx Power 41/44 dBm for 10/20MHz 46/49 dBm for 10/20MHz
Carrier frequency 2 GHz 2 GHz
Min. UE-eNB 2D distance 10m [other values FFS] 35m
UE height model
general equation hUE=3(nfl – 1) + 1.5m hUE=3(nfl – 1) + 1.5m
nfl for outdoor UEs 1 1
nfl for indoor UEs Replaced by WA Replaced by WA
Indoor UE fraction 80% 80%
Pathloss
Indoor UE 2D distance from
external building wall din for
pathloss determination
uniform(0,25m) uniform(0,25m)
Table 2:
Urban Micro cell
with high UE
density
Urban Macro cell
with high UE
density
UE distribution (in x-y
plane) Outdoor UEs uniform in cell uniform in cell
Indoor UEs uniform in cell uniform in cell
Building model Dimensions (in x-y plane) not needed not needed
Number of buildings per
macro cell NA NA
Distribution of buildings in a
macro cell NA NA
ISD 200m 500m*
關於更多模擬方法的細節將持續在此次會議後的 email discussion中進行討
論。
4. Network-Assisted Interference Cancellation (NAIC)
LTE Rel-12 NAIC 技術本次會議為第一次討論。RAN 才剛剛通過這個 SI
並準備在今年剩下的幾次會議內完成這項研究議程。在考慮情境的部分,這次
會議主要是討論 NAIC 最需要被評估的場景,以作為給 RAN4的參考,使他們
能夠建立 link level simulation的假設,並去評估出各種不同的Advanced Receiver
18
的表現與複雜度。本項目相關議題如下:
Agree on the deployment scenarios, including any prioritization
Agree on inter-cell interference conditions resulted from considered inter-cell
coordination including the case of no coordination
Agree on any additional intra-cell interference resulted from MU-MIMO operation
Agree on the targeted victim data/control channel(s) of interest and associated
transmission mode, corresponding interference data/control channel(s) and associated
transmission mode, and any additional interference reference signals
Agree on evaluation assumptions
Note: Aim for agreement on as much details as possible, especially if MU
and/or CoMP schemes are considered since both affect inter- and/or intra-cell
interference observed at UEs. Sufficient level of detail, especially the assumed
behavior of UE feedback and scheduler, should be agreed to kick off interference
modeling discussion for link-level evaluation in RAN4.
由於是第一次討論,各家公司對於需要考慮的情境都各持意見,並且對於
技術報告的文字都頗斤斤計較。除了模擬情境外,對於所以考量的互相干擾的
通道,也有許多公司提出了意見。如 Ericsson 所提出的這篇 contribution
(R1-131333) 把所有的可能皆考慮進來:
Nine different channel interference conditions (CHICOs)
due to three different physical channel types.
Interfering Channel
PDSCH ePDCCH PDCCH
Vic
tim
Chan
nel
c
PDSCH CHICO1 CHICO2 CHICO3
ePDCCH CHICO4 CHICO5 CHICO6
PDCCH CHICO7 CHICO8 CHICO9
Proposal 1: Techniques for mitigating interference on PDSCH, ePDCCH and PDCCH
shall be studied assuming interference can come from PDSCH, ePDCCH, PDCCH, or a
mixture thereof.
19
但大部分公司比較支持優先考量 PDSCH 會遭受到的干擾 (CRS 或其他
PDSCH)。最後大家把主要焦點專注於一篇由 Orange, Telecom Italia, Deutsche
Telekom, T-Mobile USA, China Unicom, KDDI, Docomo, Vodafone 等電信業者
所共同提出的 Way Forward (R1-131687)。
對於此篇 WF其他公司也提出了以下幾點評論修正:
- Consider including CoMP/eICIC schemes as baseline
- Channels of interest: Consider taking PDSCH as first priority
- Consider replacing last bullet with prioritisation of CRS-based TMs over DMRS based
TMs
- Consider intra-cell MU-MIMO separatelyConsider larger / smaller ISDs .
經過許多討論,最後對於考慮情境的決議如下:
Agreements:
- NAICS Scenario 1:
o Homogeneous network, macro only, ISD = 500m
o ITU UMa channel model
o Non-ideal backhaul between sites (same assumptions as for SCE SI)
o Coordination assumptions:
Intra-site information exchange is possible
Inter-site information exchange is subject to the backhaul latency
FFS whether complexity of information exchange is also taken into
account
o NB: This scenario is similar to CoMP scenario 1 in TR36.819
Proposal 2: Partially overlapping physical channels is a typical interference condition in
the investigations
aggressor PDCCH
aggressor PDSCH
victim PDCCH
victim PDSCH
20
eNB
Coordination area
- NAICS Scenario 2a:
o SCE Scenario 1, with the modification that the small cell deployment is sparse not
clustered (FFS: 4 or 10 per macro)
o Backhaul assumptions:
Between macro-cell and small cells within its coverage, and small nodes
under the coverage of one macro: Non-ideal
Between macros of different sites: Non-ideal
o Coordination assumptions:
Intra-site information exchange is possible
Inter-site information exchange is subject to the backhaul latency
FFS whether complexity of information exchange is also taken into
account
- NAICS Scenario 2b:
o Same as NAICS Scenario 2a, with the following exceptions:
21
o Backhaul assumptions between macro and small nodes within its coverage, and
between small nodes under the coverage of one macro: “fibre access 4” as per
TR36.932
o Coordination assumptions:
According to the backhaul assumptions, information exchange is possible
in the following cases:
Intra-site
Between a macro and a small node within its coverage
Among small nodes within the coverage of the same macro
According to the backhaul assumptions, the information exchange is
subject to the backhaul latency (+ FFS complexity) in the following cases:
Inter-site between macros
Between a macro and a small node outside its coverage
Among small nodes within the coverage of different macros
- Synchronization error
o To be defined by RAN4
- Exact latency value corresponding to non-ideal backhaul is FFS from the values
considered in the SCE SI
- General for all scenarios:
o Baseline is Rel-11
o Baseline for comparison should be the appropriate Rel-11 technique(s) for each
scenario – to be agreed for each scenario
o CRS interference modelling is included
FFS number of antenna ports and number of MBSFN subframes
CRS interference cancellation at the UE is assumed for all subframes for
up to 2 interfering cells
o Traffic model: FTP model 1
5. Further Enhancements to LTE TDD for DL-UL Interference Management
22
and Traffic Adaptation
在 3GPP RAN#58 會議中,分時多工系統增強干擾消除和流量調整議題
(TDD eIMTA)已同意被 3GPP Rel-12採納,其應用為未來非常重要的研究與發
展標的。此次會議的討論著重於兩大方向:
干擾消除的方法
分時多工系統重新組態(reconfiguration)訊號通知的機制
在干擾消除方法的部分,此次會議做出了以下決議:
7.2.3.1 Interference mitigation schemes
(Focus on understanding the benefits and standard impacts of the proposed IM schemes.)
Working assumption:
• At least for UL, the following scheme is supported for dynamic TDD UL-DL
reconfigurations:
– Depending on the type of a subframe and/or type of interference seen by a
subframe, the power control parameters and/or mechansim could be different
between a flexible subframe and a fixed subframe
• Details of subframe-type dependent power control is FFS
• Companies are encouraged to bring detailed proposals and performance evaluations in
the next meeting.
Agreement:
• Backhaul signaling capturing eNB-to-eNB interference can be beneficial for TDD
eIMTA
• Working assumption that New backhaul signaling capturing eNB-to-eNB interference is
to be introduced
– To be confirmed if gains are shown by evaluations in following meeting(s)
– FFS on the detailed contents of the information on eNB-to-eNB interference
• Any new backhaul signaling capturing eNB-to-eNB interference shall be assumed not to:
– impose mandatory behaviour in the receiving eNB
– impose new requirements on the accuracy of eNB measurements (unless shown to
be beneficial)
– impose new architecture for LTE
Note that this does not preclude consideration in RAN1 of any feedback from RAN4.
另外,關於重新組態訊號通知,此次會議最大的爭議在於是否要使用
23
PBCH/MIB 來進行直接式或者間接式的通知。詳細的決議如下:
7.2.3.2 Signalling mechanisms for TDD UL-DL reconfiguration
Agreement:
• No new TDD UL-DL configurations are introduced in the BCT (in WI on TDD eIMTA)
• Alternative 1 below is agreed.
Note: “PHY signaling” includes possibility of
UE specific or UE common signaling
Using either existing or newly defined DCI formats
Alternative 1:
• A signaling mechanism which explicitly or implicitly indicates TDD UL-DL
reconfiguration by either
• PHY signaling (not including PBCH/MIB signaling), or
• MAC signaling
• PBCH/MIB signaling issue could be revisited if reliability issue of the above method
becomes severe
Support: Renesas Mobile Europe, RIM, Intel, Samsung, Nokia, NSN, Qualcomm, Sharp , NTT
DoCoMo, Potevio, NEC, Panasonic, LGE, Ericsson, ST-Ericsson, MediaTek, ITRI,
Interdigital,
Alternative 2:
• A signaling mechanism which explicitly or implicitly indicates TDD UL-DL
reconfiguration by either
• PHY signaling ( including PBCH/MIB signaling), or
• MAC signaling
Support: CATT, CMCC, CATR, New Postcom, Coolpad, Alcatel Lucent, ASB, ZTE,
Huawei, Hi Silicon
6. Device to Device Communications
根據上個會期的共識,本會期需要針對模擬情境進行討論,本會期主要工
作在於討論 General Scenario以及 Public Safety Scenario的模擬情境,並討論這
兩種情境的模擬設定。本會期針對模擬的佈建情境,載波頻率,頻寬,評估基
準等參數進行討論。根據目前的討論,有以下六種佈建情境:
24
Option 1: Urban macro (500m ISD) + {1} RRH/Indoor Hotzone per cell
Option 2: Urban macro (500m ISD) + {1} Dual stripe per cell
Option 3: Urban macro (500m ISD) -- all UEs outdoor
Option 4: Urban macro (500m ISD) + {3} RRH/Indoor Hotzone per cell
Option 5: Urban macro (1732m ISD) (UE dropping details FFS)
Option 6: Urban micro (100m ISD)
根據會場討論,目前根據兩種情形進行投票表決,分成 General Scenario
跟 PS Specific scenario,並根據投票結果排出模擬的優先順序,投票結果如下:
o General Scenario
1: QC, USDoC, ALU, ASB, Fujitsu, GDB, CATT, NEC
2: Samsung, Nokia, NSN, RIM, Renesas,
3: ZTE, ITRI, EADS, E///, ST-E
4: LGE, Telecom Italia, Intel,
5:
6: HW, HiSi
o PS Specific scenario
1: QC, USDoC, GDB,
2:
3: ALU, ASB, Samsung, CATT, E///, ST-E, ZTE, VF, ITRI,
4
5: HW, HiSi, EADS, Fujitsu, NSN, Nokia, LGE, Nomor, NEC, Renesas
6
下表為目前 RAN1會議同意的模擬情境,該模擬情境還說明了在一般的情
境,載波中心頻率為 2GHz,在 PS Specific Scenario,中心頻率是 700MHz。對
於 General Scenario 來說,TDD 使用的 20MHz,FDD 使用的是 10MHz,對於
是否需要一個額外的頻率去表述 PS 的頻率,目前還在討論中。對於超出網路
覆蓋範圍的 UE,目前先討論以-6dB 為界去設定。目前主要是先假設在網路同
步的情境下去,但在非相同頻率的情況下,會錯開彼此間的頻率,亦即異頻下
網路是非同步狀態。用戶速度在 General Scenario 考慮為 3km/hr 的低速狀態,
在 PS Specific Scenario 考慮 120km/hr跟 3km/hr的混和速度狀態。
Property General Scenario PS Specific Scenario
25
Layout Option 1 (mandatory)
Others optional in order of
decreasing priority:
Option 2 / Option 3
Option 4
Option 6
Option 5 (mandatory)
Others optional in order of
decreasing priority:
Option 3
Option 1
Carrier
Frequency
(note that the
performance at 2GHz is
expected to be different
from the performance at
700MHz.)
2GHz 700 MHz
System
bandwidth
10MHz (FDD), 20 MHz
(TDD)
FFS whether an additional
bandwidth reflecting the
PS-specific scenario may
be added
FFS
Network
operation
100% eNBs enabled {0, x (FFS)}% eNodeB enabled
FFS whether disabled eNBs are
selected randomly or
deterministically
(Note that x may be 100%)
UE out of
coverage
criterion
Average SINR < {-x (-6dB working
assumption – can be revisited at
RAN1#73)} dB over system
bandwidth.
Network
synchronization
Mandatory:
- all eNodeBs synchronized
- eNodeBs on different carriers not synchronized
Optional: eNodeBs on a given carrier not synchronized
UE mobility {3,X} km/hr 120 km/hr for {x} fraction of
outdoor UEs
{3,X} km/hr for other UEs
UE RF
parameters
Max Tx power of 23 dBm for non PS -- 23 dBm, 31 dBm
for PS
1 Tx (2 Tx optional for PS only), 2 Rx antenna, Antenna gain 0
26
dBi, Noise figure 9 dB
UE inband
emission mask
FFS under what circumstances (if any) this is needed; if
needed, as per 36.101 s.6.5.2.3
eNodeB RF
parameters
As specified in 3GPP Case 1, except for Option 5 which uses
parameters as specified in 3GPP Case 3
Traffic models Full buffer, VoIP, & FTP2 from 36.814
Non D2D traffic With probability {X}, a D2D UE has non D2D (downlink &
uplink) traffic.
WAN traffic is FTP2
根據目前的討論,對於 unsynchronised network,Ericsson, ST-E, Telecom
Italia, Docomo, Vodafone, Deutsche Telekom, ALU, ASB, KDDI, NEC 等公司認為
是一定要考慮的情境,僅有 Qualcomm覺得這並非一定要考慮的情形。目前這
兩種情境人需要被討論。
根據 UE 的擺置,目前同的內容詳述於 R1-131789,請參考下表中,黃色
的部分為尚須討論並研究的內容。
General Scenarios Public Safety Scenarios
LTE Layout Option 1 (mandatory)
Others optional in order of
decreasing priority:
Option 2 / Option 3
Option 4
Option 6
Option 5 (mandatory)
Others optional in order of
decreasing priority:
Option 3
Option 1
Total number of
active UEs per cell
area
FFS
Starting point:
25 for options 1,2,4
10 for options 3,5,6
Number of D2D
UEs for discovery
FFS
Number of D2D
UEs for
communication
FFS
27
UE drop for all UEs, for
both discovery and
communication
evaluations
For Layout option 1,2, 4:- 2/3 UEs randomly and uniformly dropped within the clusters
of small cell(s), 1/3 UEs randomly and uniformly dropped throughout the macro
geographical area.
a) 20% UEs are outdoor and 80% UEs are indoor.
For Layout option 5, UEs randomly and uniformly dropped throughout the macro
geographical area; 20% UEs are outdoor and 80% UEs are indoor. Drop 2 RRH
buildings (without RRHs) in each macro geographical area.
For Layout option 3, 5, 6 –
a) Uniform drop - all UEs are randomly and uniformly dropped throughout the
macro geographical area
b) Hotspot drop – Randomly select an area within each macro geographical area.
Randomly and uniformly drop 2/3 UEs within 40 m of the selected area.
Randomly and uniformly drop the remaining 1/3 UEs to the entire macro
geographical area of the given macro cell
UE association for
unicast D2D
communication
Random pairing: First UE is randomly selected from all UEs within entire 19/7 macro
sites and 2nd
UE is randomly selected from the remaining UEs within entire 19/7 macro
sites
2nd
UE will be re-selected with constraint of minimum RSRP between two UEs if RSRP
is less than X dBm (FFS; in the meantime, companies may choose the value, including
- ) when UE is transmitted at maximum power
UE association for group
cast D2D communication
N/A Random pairing: First UE is randomly
selected as the UE for group cast from all
UEs within entire 19/7 macro sites
All Y number of receiving UEs are
randomly selected from the remaining UEs
within entire 19/7 macro sites
FFS-Number of receiver UEs “Y”
UEs will be re-selected with
constraint of minimum RSRP between two
UEs if the RSRP is less than X dBm (FFS;
in the meantime, companies may choose
the value, including - ) when UEs are
28
transmitted at maximum power
UE association for
broadcast D2D
communication
N/A Random pairing: First UE is randomly
selected as the UE for group cast from all
UEs within entire 19/7 macro sites
All Y number of receivier UEs are
randomly selected from the remaining UEs
within entire 19/7 macro sites
FFS: Number of receiver UEs “Y”
UE will be re-selected with constraint
of minimum RSRP between two UEs if the
RSRP is less than X dBm (FFS; in the
meantime, companies may choose the
value, including - ) when UE is
transmitted at maximum power
UE association for Relay
D2D communication
N/A First UE is randomly selected from all UEs
without eNB coverage and 2nd
UE is
selected from the UEs within eNB coverage
Minimum distance
between UE and eNB
>=35m (except for Option 6…)
Minimum distance
between UEs
>= 3m
以下兩個表格所列,說明 D2D的 Discovery與 Communications 效能指標,
討論 Discovery與 Communications 兩種情境。
Discovery效能指標
Aspect Metrics*
Performance
target
Open discovery:
- Number of UEs discovered as a function of time
(system)
- CDF of number of UEs discovered as a function of
time (system)
Closed discovery (i.e. knowing the UEs to be discovered):
- Probability of discovery as a function of time
29
(assume zero time penalty for each false alarm)
(Time measured from start of simulation without prior
synchronisation)
Range &
reliability
Prob. of discovery vs pathloss (link & system)
Prob. of false alarm (link & system)
Impact on
WAN
Amount of resource used (system) (per cell if in network
coverage)
FFS metrics related to throughput loss and/or interference
Power Power consumption modeled through ON time or equivalent
power consumed (transmit power should be captured
differently than received power --detailed model FFS)
*Same metrics used for in-network, partial network and out of network with possible different
emphasis
*Same metrics used for public safety and non-public safety cases with possible different emphasis.
Communications 效能指標
Aspect Metrics*
D2D Throughput
/spectral efficiency
User throughput (mean, 5%, CDF) for full buffer
(system) Perceived user throughput (mean, 5%, CDF)
for FTP (system)
VOIP system capacity (system) (VOIP delay
requirement {X}ms)
Range and
Reliability
Performance** vs pathloss or distance (link and
system)
For link level, use only full buffer
Call setup latency Phy. layer latency for call setup for out of coverage
only (link and system)
(This should only model L1 related aspects; higher
layer aspects should be considered in RAN2)
Impact on WAN Change in cell throughput/cell spectral efficiency
(system)
Cdfs of perceived per-user throughput for FTP2 with
and without D2D
Power
consumption
Power consumption should be modelled; detailed
model is FFS
*Same metrics used for in-network, partial network and out of network
*Same metrics used for public safety and non-public safety cases with possible different emphasis.
*Same metrics used for unicast, groupcast and broadcast with each receiver counted separately
30
** Performance means throughput, perceived throughput, prob. of satisfied VOIP user depending
on traffic model
7. Low-cost MTC
本會期針對 Low-cost MTC為了提供額外 20dB對每個通道做出了討論,針
對每個通道,從放開效能需求,到從新設計有了許多的討論。
PSS/SSS
同意事項如下:
• The coverage enhancement requirement for PSS/SSS can be achieved by longer acquisition
time
– Initial synchronization requires up to 2 seconds for FDD per a center carrier
frequency
– Re-synchronization can be performed quicker
• PSD boosting can be considered a complementary solution
• New PSS/SSS design may need to be considered if the longer time and freq. acquisition
and associated power consumption increase are not considered acceptable
• In addition to coverage enh. analysis, TP should include followings for all channel/signals
– Impact on specification
– Other impacts
• Power consumption
• Cell spectral efficiency
• Analysis/evaluation of cost reduction
PBCH
同意事項如下:
• The coverage requirement for PBCH may be met with a combination of repetition of the
current PBCH in each subframe and PSD boosting (e.g., 4dB) within 40ms (for FDD
systems)
– Repetition alone cannot meet the coverage requirement for the current PBCH
• The coverage requirement for PBCH may alternatively be met with a new PBCH design
(for TDD and FDD systems)
– A new design can consider: a longer period, reduced MIB content, intermittent
transmission. Repetitions and/or PSD boosting will be helpful for new design to
meet the requirement
31
– Also other system information that needs to be broadcasted to enhanced coverage
MTC UEs beside MIB contents can be considered in the new PBCH design
– Other low rate coding schemes or spreading can be considered for new design
• Further study could determine if there is a benefit of
– Using implementation-based solution such as new decoding techniques
PRACH
同意事項如下:
• The coverage requirement of PRACH can be achieved by preamble repetition and/or new
preamble format
– PRACH resources for coverage limited MTC UEs need to be defined and allocated
• Relaxing PRACH miss probability will make it easier to meet the coverage requirement,
and can be used in addition to repetition and/or new preamble format for PRACH coverage
improvement
• System impacts should be clarified including collision probability increase, PRACH
latency increase
• As a complement to the other techniques, PSD boosting over a narrower bandwidth may be
considered
Note:
・ Include RAN1 #72 following observation of PRACH in TP
PRACH can be used to inform eNB the amount of coverage enhancement a low cost
MTC UE needs
Observations:
• Coverage enhancement must be provided for scenarios where no small cells are deployed
• An operator may decide to deploy additional macro or LPN sites to achieve a coverage
benefit
• For deployments that already contain small cells, there may be a benefit to allow decoupled
UL and DL for delay tolerant MTC UEs
同意事項如下:
• A section is added to 9.4 to capture the above observations on the technique of using small
cells
32
• Email discussion can also discuss details of the proposal, and additional text that could be
added to the PRACH section
PDSCH
同意事項如下:
• The coverage requirements for PDSCH can be met by time domain repetition
– Cross-subframe channel estimation requires less number of repetition times than
single-subfame channel estimation
• RS power boosting and/or increased RS density may further improve the channel
estimation performance
• PSD boosting for PDSCH can help to improve the coverage
PUSCH
同意事項如下:
• PUSCH coverage enhancement requirement for MTC UE can be achieved by repetition
– Channel estimation over multiple subframes could be helpful to reduce the number
of repetitions
– Nevertheless, compared to PDSCH case, relatively larger number of repetitions
would be required for the PUSCH coverage
– Selection of TBS needs to consider the spectral efficiency and channel coding gain
• Complementary schemes can be considered to reduce the number of repetitions, for
example
– Increased DMRS density
– PSD boosting
– Frequency hopping during repetition
– Shorter length CRC
– Code spreading
• Companies may bring potential PUSCH performance issues at the next meeting
(E)PDCCH
同意事項如下:
• Repetition of (E)PDCCH across multiple subframes is required to achieve the coverage
enhancement target
33
– Specification impact may exist, such as starting subframe of first transmission, the
maximum number of repetitions, timing relationship between PDSCH and
(E)PDCCH
• PSD boosting can help to reduce the required number of repetition
• The new design of downlink control channel, for example compact DCI, higher
aggregation level, can help to reduce the required number of repetition
• Note that control overhead should be considered for further analysis in the potential WI
phase, concerning the overall system efficiency optimization
PUCCH
同意事項如下:
• The coverage requirements for PUCCH format 1a can be met with time domain repetition
• Further evaluation could show whether the following contents could be reduced or
eliminated
• SR
• CSI
• HARQ-ACK
8. Small Cell Discovery
針對 Small Cell Discovery,本次會期主要針對如何辨識 Small Cell進行討
論,目前有三種方案,第一種方案是根據同步通道來設計,第二種方案是將傳
統的同步通道跟參考信號作修改,第三種方案是設計新的 Discovery 通道去支
援 Small Cell Discovery。目前有多個需求如下:
• Requirements
– Requirement 1: To minimize UE battery consumption for discovery
– Requirement 2: To solve PCI confusion/collision
– Requirement 3: To detect sufficient number of cells
• FFS: actual target number of cells
– Requirement 4: To support detection between small cells
– Requirement 5: To support DTX small cell operation
– [FFS] Requirement 6: To have a commonality with D2D discovery
本會期同意內容如下:
34
- Metrics for evaluation:
o UE battery consumption for discovery
o Number of supportable individually identifiable small cells
Baseline is current number of supported PCIDs
Identify whether the current number is sufficient
o Number of detectable cells in the chosen scenarios
Target FFS for each scenario (or for a given SINR)
Target false alarm probability FFS
Detectability as defined in 36.133 for initial evaluation
o Probability of detecting a cell as a function of distance
o Detection time (e.g. taking into account ability to support small cell DTX operation
/ energy consumption)
o Ability to estimate the signal strength of a small cell
o Overhead
o Impact on legacy UEs
- Begin by evaluating performance of legacy mechanism (i.e. PSS/SSS/CRS)
- If inadequacies are identified with the legacy mechanism, evaluate:
o first, approaches based on modified SS/RS
o second, approaches based on new discovery signal
- Evaluation methodology:
o Up to companies to decide between e.g.:
Alt.1:
Step-0:system level simulation to model the interference profile for
link level simulation
Step-1: link level simulation to derive the performance curve (i.e.,
SINR – detection probability) based on the interference profile
derived by the Step-0 simulation
[FFS] Step-2: system level simulation based on LLS to SLS
mapping
Alt.2: System level evaluation including link-level signal generation and
detection
- Scenario:
o Scenario 2a with dense deployment of small cells
Baseline: 1 cluster per cell, 10 cells per cluster; other values can also be
evaluated.
- Synchronisation cases (in order of decreasing priority):
o 1: Synchronized transmission of discovery signal both within and between clusters
in the same or different macro cells
35
o 2: Synchronized transmission of discovery signal within clusters; unsynchronized
between clusters
o 3: Unsynchronized
o FFS: Level of synchronization (including timing offset between cells)
Resolve FFSs at RAN1#73.
9. WLAN/3GPP Radio Interworking (Rel-12) 議題
36
七、心得與建議
本團隊在這次會議中關注多個 Rel-12的新興技術議題。以下為此次會議後
本團隊對於這幾個技術項目的心得:
Further Downlink MIMO Enhancement
DL MIMO enhancements 針對 4-TX設計了新的 codebook以因應不
同天線架構,並新增 PUSCH feedback mode 3-2以改善效能,相關
的模擬平台應開始著手以助於評估各項 MIMO 技術能帶來的系統
增益。
Study on Elevation Beamforming and FD-MIMO
該 Study Item需要定義新的 3D通道模型,這也是目前會議在該項
目中的討論重點之一。應持續開發 spatial channel model 的模擬平
台,並與將來新的 3D 通道模型進一步整合。
Network-Assisted Interference Cancellation (NAIC)
繼續掌控此項研究議程的進行方向並繼續做出貢獻,幫忙模擬各種
情境,評估此項技術的可行性
可參考 RAN4的初步接收器評估
開始著手了解各種不同的接收器,並且將其效應於系統層模型化,
可於下幾次會期在會議提供簡單而不失準確性的模型
盡量以可以直接反應在 SINR上的模型為主
Further Enhancements to LTE TDD for DL-UL Interference Management
and Traffic Adaptation
此次會議著重於討論分時多工系統增強干擾消除和流量調整議題
(TDD eIMTA)的信號通知方式,目前結論為偏好能夠支持快速變換
重新組態的信號機制,例如透過 PHY或 MAC來通知 Rel-12 UEs,
以達到較好的干擾消除和流量調整效能。此外未來也將會繼續討論
信號為直接通知(explicit)或間接通知(implicit)給 Rel-12 UEs。
目前為初步的討論信號傳送機制,未來將有許多衍生議題需要我們
參與共同探討,例如分時多工系統增強干擾消除和流量調整議題
(TDD eIMTA)的 HARQ-ACK和 DCI detection timeline。
37
本次會期在 Device to Device有許多的討論,由於上個 Plenary會期決議
將 Device to Device的解決方案提前到五月會期開始,因此本會期針對
模擬情境、效能指標作了討論,因此會場針對 Device to Device的討論
是升溫的。