Djywade

2013全年3GPP RAN1会议关于D2D（Device-to-Device）技术的提案分析

说明

该文档分析了2013年6次RAN1会议中有关D2D的提案，分析的绝大部分提案是在会议report中noted的。

RAN1#72

Deployment Scenario and Evaluation Methodology

1) R1-130029 HuaweiHiSilicon

l 提出了三种场景（其中只有第二和第三种场景是noted的）

n 网络覆盖下的，高密度用户分布

n 网络覆盖下的，低密度用户分布

n 无网格覆盖的，特别低密度的用户分布

l 给出了场景1和2下的仿真参数

Parameters	Scenario 1	Scenario 2
Range	100 m	500 m
Device density	10,000/km²	100/km²
UE power	Up to 20 dBm (100 mW)	Up to 23 dBm
Evaluation	Discovery (DMC optional)	Discovery and DMC

l 可采用eIMTA中的快衰模型

2) R1-130061 CATT

l 提出的场景：

n 隔离的室内分布（InH）

n 同分布的网络场景

l 提出的分析标准

n D2D Discovery范围、蜂窝用户吞吐量/SINR降低量，不同距离下的总体吞吐量

l 另外提出了D2D撒点的方法

3) R1-130133 ZTE

l D2D discovery的目标发现距离分为三个等级：近、中等、远

l 信道模型

n 对于LOS

n 对于NLOS，可采用基于xia的室外宏蜂窝模型

l Metrics: 开销、效率、覆盖范围、功率消耗

4) R1-130304 Samsung

l Metrics: 发现的UE数量、吞吐量、公平指数

l 场景：城市微小区、城市宏小区、室内热点

l 信道模型：

n Channel models definedfor LTE-advanced (based on Rec. ITU-R M.2135)

n Channel models basedon IST-WINNER II (and WINNER +)

n Channel models basedon Rec. ITU-R P.1411

n Channel models basedon TETRA system

5) R1-130502 Nokia

l 场景：

n D2D与宏蜂窝共信道

n D2D与small cell共信道与macro不共信道

n D2D、small cell、macro三个都不共信道

l 当D2D与常规的蜂窝共用载波时，D2D可重用上行或下行的资源。在TDD/FDD的情况下使用那种资源有待进一步研究

l 考虑室外-室外，室内-室内，室内-室外的场景

l 尽可能地重用TR36.828的信道模型

l 在网络的覆盖下，优先考虑同步的D2D操作

l 重用TR36.814中的业务模型，应优先考虑非full-buffer模型

6) R1-130599 Qualcomm

l 提出了非公共安全和公共安全的使用情况

l 提出的metrics

n 关于discovery的metrics：发现的用户数，资源使用情况，功率消耗

n 关于communication的metrics：吞吐量，时延，资源使用情况

n 公共安全服务的metrics：同步所需时间

7) R1-130676 CMCC

l 提出了三种场景

n 发送端和接收端都在室外

n 发送端和接收端都在室内

n 发送端和接收端都一个在室外一个在室内

l 使用以下的信道模型

n 路损：

If R<=50m, PL=98.45+20*log10(R), R inkm

If R>50m, PL=55.78 +40*log10(R), R in m(Xia model) [2]

n 快衰

可使用TU or ITU 模型

Scenarios	Channel model
Alt 1: D2D transmitters and receivers locate outside the building	[FFS]
Alt 2: transmitter and receiver locate inside the building	InH NLOS/LOS
Alt 3: Transmitter and receiver are separated between indoor and outdoor	[FFS]

RAN1#72b

Scenarios and Evaluation Methodology

1) R1-131166 Huawei,HiSilicon（noted）

l 提出两种discovery方式

n eNB辅助discovery（无需链路级仿真）

n UE间直接discovery（需要链路级仿真）

n 链路级仿真参数表

Parameters	Assumptions
System bandwidth	10 MHz
Carrier frequency	2 GHz
Channel model	Under discussion, but should include indoor-to-indoor, outdoor-to-outdoor, and indoor-to-outdoor
Antenna configuration	2 antennas, in cross-polarization configuration
UE speed	6 km/h

l 按照距离远近提出三种场景，如下

Parameters	Short range	Medium range	Long range
Range	100 m	500 m	2 km
Device density	10,000/km²	100/km²	10/km²
UE power	Up to 20 dBm	Up to 23 dBm	Up to 30 dBm

l 由于场景太多，运营商应选择合适的场景来指导SI work

l 对于long rang场景不进行系统级评估，只进行链路级预算分析

l 把PS场景作为初步研究重点

l 对于discovery和communication需研究以下场景

n Indoor-to-indoor: short-range

n Indoor-to-outdoor: long-range, medium range

n Outdoor-to-outdoor: long range, medium range

l 优先考虑在网络覆盖内的情况，但

n 至少考虑一种不在网络覆盖之内的场景

n 至少考虑一种被混合网络覆盖的场景

l 提出两种metrics

n 被发现的设备与时间的曲线

n 被发现的设备之间的距离分布曲线

n 系统仿真参数

Parameters	Assumptions
System bandwidth	10 MHz
Carrier frequency	2 GHz
Inter-site distance (eNBs)	500 m for dense/medium UE density, 1 km for low density
eNB deployment	The typical 19-cell and 3-sectored hexagon system layout
Minimum distance between UE and Macro	35m
UE antenna gain	0 dBi
UE noise figure	9 dB
UE Antenna pattern:	Omni
Shadowing correlation between UEs	0
UE sensitivity level	-110 dBm
UE-to-UE pathloss and channel configuration	See companion contribution [2]
UE transmit power	20 dBm for consumer, 30 dBm for public safety

l 对于directmobile communications (DMC)

n 需要计算蜂窝系统和DMC的吞吐量，与没有DMC时的蜂窝系统吞吐量做出比较

n 设置在一个小区内DMC连接的数量（如1,2,5）

n 设定给DMC分配的资源比例（25%，50%，75%）

n 计算所有设备之间的干扰

l 如何判定哪些UE参与DMC连接，三种可能

n 同一覆盖网络下的两个随机UE

n 第一个UE附近有热点覆盖

n D2D对之间是固定的距离，被随机抽取

l 需要考虑用哪个频带，三种可能

n 使用上行频带

n 使用下行频带

n 使用专用的频带

l 使用与small cell相同的通信模型（trafficmodel）

2) R1-131686 Samsung（noted）

l 建议D2D场景采用Urbanmacro-cell with dual-strip model

l 初步着重评估简单的场景：urbanmacro场景，UE全部为室外

l 提出D2D部署场景（在网络覆盖范围内）

Deployment scenario for the evaluation	Urban Macro Scenario	Indoor scenario
Layout	Hexagonal grid, 19 cell sites, 3 sectors per site, wrap around (considering macro model in 36.814)	Dual-strip model containing two buildings with 2 floors. Each building contains 20 rooms of 10 m x 10 m each floor (considering HeNB model in 36.814)
Inter-site distance (ISD)	200m, 500 m (considering UMi, UMa in 36.814)	N.A.
Non-D2D link model	outdoor to outdoor	Indoor to outdoor
D2D link model	Outdoor to outdoor	Indoor to outdoor Indoor A to indoor A Indoor A to indoor B

l Discovery的度量和相关的SA要求

Metrics	Related requirements in TR22.803
Discovery Range	[PR.4] ProSe Discovery shall support a minimum of three range classes – for example short, medium and maximum range.
Power consumption and spectral efficiency	[PR.7] The impact of ProSe Services (Discovery and Communications) on radio usage, network usage and battery consumption should be minimized.
Number of discoverable UEs	[PR.43] ProSe Discovery and Communication shall take into account the potentially large numbers of concurrently participating ProSe-enabled UEs.
Impact to legacy services	[PR.125] ProSe Communication and ProSe Discovery shall not adversely affect other E-UTRAN services.
Link reliability	No requirements

l 对于directcommunication的metric

n 在没有D2D和有D2D场景下的吞吐量cdf曲线比较，D2D不同算法间的吞吐量变化比较

n 对于directcommunication最有用的场景为VoIP，所以应评估VoIP的capacity，如一个小区内VoIP连接的数目和中断率

Metrics	Related requirements in TR22.803
D2D throughput	[PR.7] The impact of ProSe Services (Discovery and Communications) on radio usage, network usage and battery consumption should be minimized.
VoIP system capacity	[PR.43] ProSe Discovery and Communication shall take into account the potentially large numbers of concurrently participating ProSe-enabled UEs.
Impact to legacy services	[PR.125] ProSe Communication and ProSe Discovery shall not adversely affect other E-UTRAN services.

l D2D场景图

Figure1: D2D scenario inside network coverage

l D2D评估假设参数

Deployment scenario for the evaluation	Urban Macro Scenario	Indoor scenario
Layout	Hexagonal grid, 19 cell sites, 3 sectors per site, wrap around (macro model in 36.814)	Dual strip model containing two buildings with 2 floors. Each building contains 20 rooms of 10 m x 10 m each floor (HeNB model in 36.814) Dropped randomly in each cell
Inter-site distance (ISD)	200m, 500 m (considering Umi, Uma in 36.814)	N.A.
Non-D2D link model	Outdoor to outdoor	Indoor to outdoor
D2D link model (each channel model should be defined separately)	Outdoor to outdoor	Indoor to outdoor Indoor A to indoor A Indoor A to indoor B
System bandwidth	10MHz
Carrier frequency	2.0GHz
Carrier number	1
Total BS TX power	46dBm
UE TX power	23dBm
BS Antenna Height:	25m
UE antenna Height	1.5m
D2D resource	UL resource
Number of D2D UEs	Peer discovery: 20, 50, 100 UEs Direct communication: 10, 25, 50 pairs of UEs
Number of non-D2D UEs	0, 10 UEs
UE drop	Uniformly random
UE speed	3Km/h
Network synchronization	Baseline is synchronized
UE antenna	1Tx-2Rx

3) R1-131037 Samsung（noted）

(a) Out of network coverage (b) Partial networkcoverage

Figure 1: D2D scenarios

Table1: Deployment scenarios and simulation parameters for out of networkcoverage and partial network coverage

			Out of network coverage	Partial network coverage
Layout			Hexagonal grid, two-tier with 19 cell sites, and wrap around structure	Hexagonal grid with 2 cell sites
Inter-site distance			500m	500m
Number of UEs	Peer discovery		100 UEs per grid	Total 100 UEs (50, 50)*
	Direct communication	Unicast	30 UEs per grid	Total 30 UEs (15, 15)*
		Groupcast	5 groups per grid 6 UEs in each group, (1Tx, 5Rx)	1 group in the layout 30 UEs in the group (15, 15)*
UE dropping	Peer discovery		Uniformly random	Uniformly random
	Direct communication		Uniformly random with pairs or groups within a grid	Uniformly random with pairs or groups across cells
D2D link model			Outdoor to outdoor	Outdoor to outdoor
Synchronization			All UEs are synchronized.	All UEs are synchronized.
System bandwidth			10	10
Carrier frequency (GHz)			2	2
Carrier number			1	1
Antenna pattern			Omni	Omni
UE antenna Height (m)			1.5	1.5
Traffic models for direct communication			Full Buffer, VoIP	Full Buffer, VoIP
UE speed			3km/h	3km/h
Antenna configuration			1Tx/2Rx	1Tx/2Rx

* In (x, y), xis the number of UEs in networkcoverage and y is thenumber of UEs out of network coverage.

l 在每个场景下，除了UEs的数目和UE的撒点对于peer discovery和direct communication是不同的，其他参数都是相同的

l 评估的重点放在所有的户外终端

l 对于ProSe discovery的metric

n 可靠性，需要评估false alarm rate

n Discovery的UE数目

n 需评估UE的电量消耗情况

l 对于ProSe direct communication

n 吞吐量

n 公平性指数

n VoIP capacity

4) R1-131412 Qualcomm Inc（noted）

l 提出两种场景

Parameter	General Scenario	PS Scenario
Network coverage	100% UEs in-coverage	{0% UEs in coverage (out-of-network), 50% eNodeBs disabled (partial coverage)}
Carrier Frequency (GHz)	2	0.7
System Bandwidth (MHz)	10 (FDD), 20 (TDD)	10
System Synchronization	Synchronized	Non-synchronized
UE Drop	Non-clustered	Clustered around events
UE speed	3 km/hr	{3, X*} km/hr indoor {3, 120} km/hr outdoor

*Note that X heredenotes values that are TBD

Table 1: Summary of proposeddeployment scenarios

Parameter	General Scenario	PS Scenario
Network coverage (%)	100% UEs are in-coverage	{0% UEs in-coverage (out-of-network), 50% eNodeBs disabled (partial coverage)}
Carrier Frequency (GHz)	2	0.7
System Bandwidth (MHz)	10 (FDD), 20 (TDD)	10
System Synchronization	Synchronized	Non-synchronized
UE Drop	Non-clustered	Clustered around events
UE speed	3 km/hr	{3,X} km/hr indoor {3, 120} km/hr outdoor
Additional Simulation Parameters	See Table 2	See Table 3
Deployment Layout	Heterogeneous network 3GPP Case 5.3 (Macro + Indoor RRH/Hotzone) [2] o Macro: 3GPP Case 1 [2] o Indoor hotspot: As per Section A.2.1.1.5 in [2] 4 indoor hotspot per macro-cell, uniform in macro-cell area
Channel Model	See [3]
Traffic Model	Full buffer, FTP and VoIP Unidirectional traffic for Unicast and Broadcast Multidirectional traffic for Groupcast

Table 2: Additionalsimulation parameters for general scenario

Parameter	Value
Legacy Ues
Number of legacy UEs per macro cell	100
Fraction of legacy UE indoor	{0.35, 0.80}
Discovery UEs**
Number of discovery devices per sector	{50, 100, 500}
Fraction discovery devices indoor	{0.35, 0.80}
Communication Links*
Number of unicast links per sector	{5, 25}
Fraction of unicast transmitters indoor	{0.35, 0.80}
Number of groupcast UE groups per sector	{5, X}
Fraction of groupcast UE groups with the first UE of the group dropped indoor	{0.35, 0.80}
Number of UEs that belong to a groupcast UE group	X
Number of broadcast UE groups per sector	{1,X }
Fraction of broadcast transmitters indoor	{0.35, 0.80}
Intended fanout of broadcast transmission	{10, X}

* Note:when evaluating a mix of unicast links,groupcast UE groups, and broadcast UE groups, use only the lowest number oflinks/groups proposed in each case.

** Note:when evaluating high number of discovery UEs, a 7-cell macro layout may beused.

Table 3: Additionalsimulation parameters for public safety specific scenario

Parameter	Value
Event Clusters
Number of clusters per macro-cell	1
Cluster radius	125m
Fraction of cluster centres that lie within an indoor hotspot	{0.35, 0.80}
Discovery UEs**
Number of discovery devices per cluster	{50, 100, 500}
Communication Links*
Number of unicast links per cluster	{5, 25}
Number of groupcast UE groups per cluster	{5, X}
Number of UEs that belong to a groupcast UE group	X
Number of broadcast UE group per cluster	{1,X}
Intended fanout of broadcast transmission	{10, X}

*Note:when evaluating a mix of unicast links, groupcastUE groups, and broadcast UE groups, use only the lowest number of links/groupsproposed in each case.

** Note:when evaluating high number of discovery UEs, a 7-cell macro layout may beused.

5) R1-131430 NTT DOCOMO（noted）

Table I. Categorization of D2D Deployment Scenarios

Public / Non-public safety	NW coverage	In-band / Out-band	Observation	No.
General scenarios (applicable to both)	In NW coverage	In-band	• Support of inter-operator operability would be more challenging • Impact on legacy cellular system should be considered	#1
General scenarios (applicable to both)	In NW coverage	Out-band	• Might be more suited for unified solution between public and non-public safety	#2
Public safety specific scenarios	Out of NW coverage / partial NW coverage	Out-band	• Technical differences between out of NW coverage and partial NW coverage should be clarified	#3

Figure 1.D2D deployment scenarios

l UE进行D2D操作时应该保持同步，这样可以避免不断地搜索带来的电量损耗

l General case的需求

n Discovery和communication

ü 支持很多UE

ü 能耗效率

ü 资源效率

ü 对传统UE最小影响

n Discovery

ü 支持最少三种范围

ü Location serviceenhancement

l Public safety case的需求

n Discovery和communication

ü 支持室内/外

ü 没有对传统UE影响

n Communication

ü 支持direct communication

ü 支持组播和广播

ü 支持中继

l 性能指标

n Discovery和communication

ü 资源效率

ü UE功耗效率

ü 对WAN的影响

ü Discovery的可靠性

n Discover

ü Discovery的UE数目

ü Discovery范围

ü 距离/范围估计精度

n Communication

ü 频谱效率

l 不能采用全双工方式

l 对于在NW覆盖下的in-band D2D，复用蜂窝上行资源对于资源利用来说是优先选择

6) R1-131621 Ericsson,ST-Ericsson（noted）

l 链路级仿真采用EPA、ETU mode

l 不要把discover作为communication的先决条件

l 评估groupcommunication

l 仿真两种场景：Suburban（A）和UrbanHotspot（B）

Table 1: Summary of scenarios definitions

A: Suburban scenario	B: Urban hotspot scenario
Macro NW with 2 antennas/sector	Macro NW with 2 antennas/sector
UEs with 1tx/2rx	UEs with 1tx/2rx
Cellular network: deployed according to ITU Suburban Macro (1299m ISD, hexagonal grid, SMa model). 7 cells generated with wrap-around.	Cellular network: deployed according to ITU Urban Macro (500m ISD, hexagonal grid, Uma model). 7 cells generated with wrap-around.
D2D channel model from [4]	D2D channel model from [4]
Uniform UE density (50% indoors, 50% vehicle)	Hotspot UE density (indoor/outdoor and floor assignment follow model in Section 4.1))
UE speed: indoor UEs 3km/h, vehicle UEs 90km/h	UE speed: 3km/h
In-band D2D Operation	In-band D2D Operation
UE Emission model from [5]	UE Emission model from [5]
Options:
A1:full NW coverage A2: partial NW coverage (a subset of the macros is enabled) A3: no NW coverage	A1:full NW coverage A2: partial NW coverage (a subset of the macros is enabled) A3: no NW coverage
Carrier frequency: 700 MHz (NSPS), 2 GHz	Carrier frequency: 700 MHz (NSPS), 2GHz

l 为商业评估划出一个子集

l 对于suburban采用均匀撒点

n 50%车内UE，50%室内UE

l 对于urbanhotspot场景采用热点分布撒点

7) R1-131715 Qualcomm, USDepartment of Commerce, General Dynamics Broadband, ZTE, Samsung, RIM, Fujistu,III, ITRI

l D2D部署

8) R1-131747 Qualcomm, USDepartment of Commerce, General Dynamics Broadband UK, RIM, Fujistu, III, NEC

l discovery Metrics

l communication metrics

9) R1-131755 Alcatel-Lucent,Alcatel-Lucent Shanghai Bell（noted）

l 两个场景（general和public safety scenarios）

六边形网络，19或者7小区，3扇区

n Option 1: Urban macro(500m ISD) + {1} RRH/Indoor Hotzone per cell

n Option 2: Urban macro(500m ISD) + {1} Dual stripe per cell

n Option 3: Urban macro(500m ISD) -- all UEs outdoor

n Option 4: Urban macro(500m ISD) + {3} RRH/Indoor Hotzone per cell

n Option 5: Urban macro(1732m ISD) (UE dropping details FFS)

n Option 6: Urban micro(100m ISD)

Table 1: UEdropping and association for unicast, group cast, broadcast, and Relay

	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 (active) cell	25 for options 1,2,4 10 for options 3,5,6 (FFS)
Number of D2D UEs for discovery	FFS
Number of D2D UEs for communication	FFS
UE drop for discovery and communication	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 option 1,2,4 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 2^nd UE is randomly selected from the remaining UEs within entire 19/7 macro sites 2^nd UE will be re-selected with constraint of minimum RSRP between two UEs if RSRP is less than X dBm (FFS) 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 receivier 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) when UEs are 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 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 2^nd UE is selected from the UEs within eNB coverage FFS: 2^nd UE will be re-selected with constraint of minimum RSRP between two UEs if the RSRP is less than X dBm when UE is transmitted at maximum power
Minimum distance between UE and eNB	>=35m
Minimum distance between UEs	>= 3m

10) R1-131413Qualcomm Incorporated（nottreated）

l 该提案提出的metrics和其他公司提出的相似，但是conclusion不明确

11) R1-131762Fujistu, Qualcomm, US DOC（noted）

l Network coverage基于eNB和UE的RF配置，包括带宽、发射功率、发射和接收天线数目（参考TR36.888的方法）

l 对于10MHz的FDD，基站天线是4发4收，手机是1发2收，用以下下行SINR阈值来决定network coverage

UE power (dBm)

Required

DL SINR (dB)

-6.5dB

-7.8dB

12) R1-131754NTT DOCOMO, US Department of Commerce, Vodafone, KDDI, Deutsche Telekom, CMCC,Ericsson, ST-Ericsson, Qualcomm, ZTE, ITRI（noted）

l It is preferable tohave a common technical solution that accommodates both consumer and PSrequirements

Channel Model

1) R1-131768 Qualcomm, USDepartment of Commerce, General Dynamics Broadband UK, ITRI, Fujitsu, Intel,Samsung, ZTE, III（noted）

RAN1#73

Remaining Details of Scenarios and Evaluation Methodology

1) R1-132719 Qualcomm（noted）（该提案为重点）

l 部署场景总结：

n Option 1: Urban macro(500m ISD) + 1 RRH/Indoor Hotzone per cell

n Option 2: Urban macro(500m ISD) + 1 Dual stripe per cell

n Option 3: Urban macro(500m ISD) (all UEs outdoor)

n Option 4: Urban macro(500m ISD) + 3 RRH/Indoor Hotzone per cell

n Option 5: Urban macro(1732m ISD)

n Option 6: Urban micro(100m ISD)

Table A.2.1.1: Details ofDeployment Scenarios

	General Scenarios		Public Safety Scenarios
LTE Layout	Option 1 shall be mandatory Others layouts are optional in order of decreasing priority: Option 2 / Option 3 Option 4 Option 6		Option 5 shall be mandatory Others layouts are optional in order of decreasing priority: Option 3 Option 1
Carrier Frequency (Note: The performance at 2GHz is expected to be different from the performance at 700MHz.)	2GHz		700 MHz
System bandwidth	10MHz Uplink and 10MHz Downlink for FDD, 20 MHz for TDD		FFS
Network operation	100% eNodeBs enabled		{0, x (FFS)}% eNodeBs enabled (Note that x may be 100%) FFS whether disabled eNodeBs are selected randomly or deterministically
UE out of coverage criterion	N/A		Average SINR < {-x} dB over system bandwidth.
Network synchronization	All cases shall be treated with equal priority: - all eNodeBs synchronized - eNodeBs on different carriers not synchronized - 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 transmit power of 23 dBm for non public safety, 23 dBm, 31 dBm for public safety 1 Tx (2 Tx optional for public safety only), 2 Rx antenna, Antenna gain 0 dBi, Noise figure 9 dB
eNodeB RF parameters	As specified in 3GPP Case 1, except for Option 5 which uses parameters as specified in 3GPP Case 3 (Table A.2.1.1.1 of [3])
Non D2D traffic	With probability {X}, a D2D UE has non D2D (downlink & uplink) traffic. WAN traffic source shall be FTP2.
Number of D2D UEs for discovery	FFS
Number of D2D UEs for communication	FFS
UE drop for all UEs, for both discovery and communication evaluations	For layout options 1,2, 4: 2/3 UEs randomly and uniformly dropped within the clusters of small cell(s). Remaining 1/3 UEs randomly and uniformly dropped throughout the macro geographical area. 20% UEs are outdoor, and 80% UEs are indoor. For layout option 3, 5, 6: Uniform drop : all UEs are randomly and uniformly dropped throughout the macro geographical area 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. Additionally for layout option 5: Drop 2 RRH buildings (without RRHs) in each macro geographical area. (See A.2.1.1.5 in [3]) All UEs randomly and uniformly dropped throughout the macro geographical area such that 20% UEs are outdoor, and 80% UEs are indoor.
UE association for unicast D2D communication	Random pairing: First UE is randomly selected from all UEs within entire 19/7 macro sites and 2^nd UE is randomly selected from the remaining UEs within entire 19/7 macro sites. If random pairing does not lead to target number of unicast D2D link 2^nd UE shall 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 groupcast 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. If random pairing does not lead to target number groupcast groups then UEs shall 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 transmitted at maximum power. The number of receiver UEs ‘Y’ is FFS.
UE association for broadcast D2D communication	N/A	Random pairing: First UE is randomly selected as the UE for broadcast from all UEs within entire 19/7 macro sites. All Y number of receiver UEs are randomly selected from the remaining UEs within entire 19/7 macro sites. If random pairing does not lead to target number groupcast groups then UE shall 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. The number of receiver UEs ‘Y’ is FFS.
UE association for Relay D2D communication	N/A		First UE is randomly selected from all UEs without eNodeB coverage and 2^nd UE is selected from the UEs within eNodeB coverage
Minimum distance between UE and eNodeB	>=35m (except for Option 6 where it shall be 5m)
Minimum distance between UEs	>= 3m

l 通信模型（traffic model）

采用full buffer、VoIP、FTP2^{[3GPP TS 36.814 V9.0.0,“Further advancements for E-UTRA physical layer aspects (Release 11)”]}

l Metrics for discovery

n Number of UEsdiscovered as a function of time. This shall be a system level metric.

n CDF of number of UEsdiscovered as a function of time. This shall be a system level metric.

n Probability ofdiscovery as a function of time. Zero time penalty shall be assumed for eachfalse alarm. This shall be a system level metric.

n Probability ofdiscovery vs. pathloss. This shall be both a link & system level metric.

n Probability of false alarm.This shall be both a link & system level metric.

n Amount of resourceused for discovery per cell if in network coverage. This shall be a systemlevel metric.

n FFS metrics related tothroughput loss and/or interference.

n Power consumptionmodeled through ON time or equivalent power consumed. Transmit power should becaptured differently than received power. Detailed model FFS.

l 需要注意：

n Time shall be measuredfrom start of simulation without prior synchronisation.

n Unless explicitlystated same metrics shall be used for in-network, partial network and out ofnetwork coverage scenarios with possibly different emphasis.

n Same metrics shall beused for public safety and non-public safety cases with possibly differentemphasis.

l Metrics forcommunication

n For full buffertraffic model: mean, 5%, and CDF of user throughput. This shall be a systemlevel metric.

n For FTP2 traffic model:mean, 5%, and CDF of perceived user throughput. This shall be a system levelmetric.

n For VOIP traffic model:VOIP system capacity. The VOIP delay requirement shall be {X} ms. This shall bea system level metric.

n Performance versuspathloss or distance. Performance shall be in terms of either user throughput,perceived user throughput, or probability of satisfied VOIP user depending ontraffic model. This shall be both a link and system level metric. For linklevel performance use only full buffer.

n Physical layer latencyfor call setup for out of coverage only. This should only model L1 relatedaspects; higher layer aspects should be considered in RAN2. This shall be botha link and system metric.

n Change in cellthroughput/cell spectral efficiency for full buffer traffic model. This shallbe a system level metric.

n CDFs of perceivedper-user throughput for FTP2 with and without D2D. This shall be a system levelmetric.

n Power consumptionshould be modeled; detailed model is FFS.

l 需要注意：

n Same metrics shall beused for in-network, partial network and out of network coverage scenarios.

n Same metrics shall beused for public safety and non-public safety cases with possibly differentemphasis.

n Same metrics shall beused for unicast, groupcast and broadcast with each receiver countedseparately.

2) R1-132458 Ericsson,ST-Ericsson（noted）（该提案提出的UE功耗模型为重点）

l Public safetyscenarios的带宽是10MHz

l 不在基站覆盖下的UE接收信号强度<-120dBm，PSS/SSS Es/Iot<-6dB

l 以上准则需在所有的D2D场景中使用，使UEs能够被蜂窝系统服务

l 总的UE数目是指宏小区范围内的UE数目

l 每个想进行D2D通信的UE对被随机的选择，如果两个UE恰好在D2D范围内，它们可能会进行D2D通信

l 如果对高速移动的UE进行建模，RAN1应该讨论合适的移动模型

l 功耗模型：

n 功耗模型适用于discovery和communication，但是相应的参数值不同

n General和public safety场景使用相同的功耗，模型和参数设置如下：

Table 1: Example power consumption model parameters

Parameter	Value	Comment
	250 mW	E.g., cell search
	2 mW	Inactive state of DL RX, i.e., only maintaining clock, memory, etc.
	300 mW	For UE transmitting at <0 dBm
	335 mW	For UE transmitting at 0 dBm
	380 mW	For UE transmitting at 5 dBm
	480 mW	For UE transmitting at 10 dBm
	1280 mW	For UE transmitting at 20 dBm
	1750 mW	For UE transmitting at 23 dBm
	TBD	For UE transmitting at >23 dBm (public safety UEs only, with the maximum output power of 31 dBm)
	250 mW	Receiving in UL, i.e., D2D activity
	2.5 mW	Inactive state of UL RX
	19 mW	GPS power consumption included only for solutions requiring GPS-based synchronization

3) R1-132065 Alcatel-Lucent,Alcatel-Lucent Shanghai Bell（noted）

l public safety场景使用5MHz的带宽

l out-of coverage场景下所有基站不能参与，部分覆盖场景有50%的基站不能参与

l 决定UE为out-of coverage的标准为-6dB

l 对于UE的移动性，需要考虑0km/h的情况

l 简单起见，研究D2D时不需要对emission mask建模

l 激活UE的总数目应该和LTE-A的一样，对于同构网络，每个小区有10个用户，对于异构网络，每个小区有25个用户

l discovery的UE数目应为激活UE数目的3倍

l 对于open discovery，D2D communication的UE数目为激活UE数目的40%。对于network assisted，D2D communication的数目由调度策略决定

l 建议10%的D2D用户同时并行参与LTE network通信

l RSRP门限值不应被确定，应由具体的调度策略在仿真的时候确定

l 组播接收UE的数目为3

l 广播接收UE在具体仿真时确定

Table 1: Simulation parameters for D2D discovery and communication

		General Scenarios	Public Safety Scenarios
LTE Layout		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)
		Option 1 (mandatory) 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*		2GHz	700 MHz
System BW		10MHz (FDD), 20 MHz (TDD) **	FFS
Network Operation		100% eNBs enabled	0, x (FFS)}% eNodeB enabled( x may be 100%) FFS disabled eNBs are selected randomly or deterministically
UE out of coverage criterion			Average SINR < {-x (-6dB working assumption – can be revisited at RAN1#73)} dB over system bandwidth.
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 23 dBm • 1 Tx 2 Rx antenna, • Antenna gain 0 dBi, • Noise figure 9 dB	• Max Tx power 23 dBm, 31 dBm • 1 Tx (2 Tx optional), 2 Rx antenna • Antenna gain 0 dBi • Noise figure 9 dB
eNB RF parameter		3GPP case 1	3GPP case 1 (case 3 for option 5)
Network Synchronization		With equal priority: • all eNodeBs synchronized • eNodeBs on different carriers not synchronized • eNodeBs on a given carrier not synchronized
Traffic model		Full buffer or FTP-2 in 36.814
Emission mask		FFS under what circumstances (if any) this is needed; if needed, as per 36.101 s.6.5.2.3
	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
	Minimum distance between UE and eNB	>=35m (except for Option 6 is 5 m)
	Minimum distance between UEs	>= 3m
	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
	Non D2D traffic	a) With probability {X}, a D2D UE has non D2D (downlink & uplink) traffic. b) WAN traffic is FTP2

Table 2: Simulation parametersfor device pairing for different deployment scenarios

	General Scenarios	Public Safety Scenarios
UE association for unicast D2D communication	Random pairing: First UE is randomly selected from all UEs within entire 19/7 macro sites and 2^nd UE is randomly selected from the remaining UEs within entire 19/7 macro sites 2^nd 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 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

l D2D对系统性能的影响可以通过系统级仿真总吞吐量简单地计算出来

l 对D2D discovery和communication使用的平均电量进行功率消耗的建模

l Discovery Metrics

Table3: Performance matrix for D2D discovery

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 (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)

*Samemetrics used for in-network, partial network and out of network with possibledifferent emphasis

*Samemetrics used for public safety and non-public safety cases with possibledifferent emphasis.

l Communication Metrics

Table 4: Performance matrix for D2D communication

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

*Samemetrics used for in-network, partial network and out of network

*Samemetrics used for public safety and non-public safety cases with possibledifferent emphasis.

*Samemetrics used for unicast, groupcast and broadcast with each receiver countedseparately

**Performance means throughput, perceived throughput, prob. of satisfied VOIPuser depending on traffic model

4) R1-132500 QualcommIncorporated, US Department Of Commerce（noted）

l 使用R1-131747中提出的D2D信道模型，并参考其中的表1

l 关于部署场景的变更：

n 确保仿真所有的场景使用wrap-around

n 为了评估D2D性能，small cell的节点（如pico和femto等）不应被仿真

l UE的撒点进行一下修改

n 对于option1，2,5（室内外混合撒点情况），35%是室内，剩下的是室外

l 使用下表的参数

Parameter	Value
Total number of active UEs per cell area*	Layout	UE drop	Value****
	Option 1	Indoor-outdoor mix**	{15} + (#D2D UEs)
	Option 5	Indoor-outdoor mix**	{15} + (#D2D UEs)
	Option 5	Uniform (outdoor)	{10} + (#D2D UEs)
	For other layout options: FFS
Number of D2D UEs for discovery per cell area*	Layout	UE drop	Value
	Option 1	Indoor-outdoor mix**	{50, 150}
	Option 5	Indoor-outdoor mix**	{50, 75}
	Option 5	Uniform (outdoor)	{50, 100, 400}
	For other layout options: FFS
Number of receiver UEs for broadcast and groupcast (Y)	Groupcast: {5, 10} Broadcast: {10,75} for uniform (outdoor) drop; {10} for indoor-outdoor mix drop.
Number of D2D UEs for communication per cell area*	Unicast: {5, 25} * 2 Groupcast: {2, 5} * Y Broadcast: {2, 5} * (1+Y) where Y is the ‘number of receiver UEs’ parameter listed above.
Minimum RSRP constraint for UE association	-97 dBm

l 在对部分覆盖建模时，只使用中心eNB

l 对于全覆盖和部分覆盖的场景，使用10MHz FDD或20MHz的TDD，对于不覆盖场景，使用10MHz专用频带

l 对于UE移动性，使用表5中的参数，包括一个改动，在PS场景，限制室外UE最大运动速度为60km/h

l 同频带内信号发射功率会均匀的分配给所有的子载波

l D2D communication使用VoIP的time budget：150ms（无中继）

l 功率消耗模型与不同的发射/接收功率有关，也与能量开销有关

5) R1-132560 GeneralDynamics Broadband UK（noted）（该提案有仿真结果）

l 每个扇区的UE数目应该比10大

l 链路重选的FFS constrain，如果两个UE间的路损大于120dB，然后应该重新选择第二个UE，这适用于UE的发射功率为23dBm

l RRH building的场景，室外用户的比例至少为50%

Table 1 - Evaluation Assumptions for simulation

Property	PS Specific Scenario Value
Layout	Option 5 - Urban macro (1732m ISD) 19 site, 3 sectors per site
Carrier Frequency	700MHz
Network operation	0 % eNodeB enabled ¹
UE RF parameters	1 Tx, 2 Rx antenna, Antenna gain 0 dBi, Noise figure 9dB

Table 2 - Evaluation Assumptions for UE dropping andassociation

Property	PS Specific Scenario Value
Layout	Option 5 - Urban macro (1732m ISD)
Total number of active UEs per cell area	{10, 27}²
Number of D2D UEs for communication	FFS ³
UE drop ⁴	Case 1: 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. Case 2: Uniform drop - all UEs are randomly and uniformly dropped throughout the macro geographical area Case 3: 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 macro sites and 2^nd UE is randomly selected from the remaining UEs within entire 19 macro sites 2^nd UE will be re-selected if the path loss between the two UEs is greater than 120dB ⁵
UE association for group cast D2D communication	Random pairing: First UE is randomly selected as the UE for group cast from all UEs within entire 19 macro sites All Y number of receiving UEs are randomly selected from the remaining UEs within entire 19 macro sites · FFS-Number of receiver UEs “Y” ⁶ · UEs will be re-selected if the path loss between the two UEs is greater than 120dB ⁵

Table 3 - Mean number of links and Mean % of UEs inrange

Scenario	Mean number of links	Mean % of UEs in range
O2O case 2, 10 UEs	11	2.2
O2O case 2, 27 UEs	27	2.2
O2O case 3, 10 UEs	13	2.5
O2O case 3, 27 UEs	34	2.4

Figure 1 – Outdoor tooutdoor Simulation case 2. UEs/sector =10

Figure 2 – Outdoor to outdoor Simulation case 2. UEs/sector = 27

Figure 3 – Outdoor to outdoor Simulation case 3. UEs / sector = 10

Figure 4 – Outdoor to outdoor Simulation case 3. UEs/sector = 27

6) R1-132695 LGElectronics, LG Uplus, Qualcomm, General Dynamics Broadband UK（noted）

n 对于public safety specific场景

n 系统带宽

ü 全覆盖（in-coverage）和部分覆盖（partial coverage）场景：10MHz（FDD），20MHz（TDD）

ü 不覆盖（out-of-coverage）场景：10MHz专用频谱

n 对于部分网络覆盖的网络的措施

ü 中心宏基站enable，剩余的disabled

n 不覆盖场景标准

ü 平均SINR-6dB

n WAN traffic UE

ü 如果不覆盖场景下UE产生WAN traffic，并且UE找不到一个或更多的中继UE，那么WAN traffic will be dropped

n UE association

ü UE中继：第一个UE从不覆盖的所有UE中随机选择，覆盖下的第二个UE的选择方法也应为设计方案的一部分

7) R1-132718 QualcommInc., General Dynamics Broadband UK, US Dept.of Commerce, Fujitsu, III（notnoted）

l UE的撒点和association

* Note that a cell refers to asector of the geographical macro-cell (hexagon).
** Active UEs are UE with WAN traffic

***AUE can be part of only at most one unicast link

****AUE can be part of only at most one groupcast group

l 指标和场景

8) R1-132722 QualcommInc.（noted）

l UE撒点和association

*Notethat a cell refers to a sector of the geographical macro-cell (hexagon)

**ActiveUEs are UE with WAN traffic

***WorkingAssumption: The unicast traffic will flow from the first selected UE to thesecond selected UE

9) R1-132754 ALU, ASB,Ericsson, Hi-Silicon, Huawei, ST-Ericsson（noted）

l 对于广播和组播，discovery并不是必要地步骤

l 对于组播，并不是说组内的所有UE都必须临近

l 为了仿真，组播是一个UE对两个或更多组内UE的单向通信

10) R1-132759Qualcomm, Fujitsu, US Department of Commerce, General Dynamics Broadband（noted）

l 发射接收功率

n Sleep功率=0.01 unit每子帧

n 接收功率=1unit每子帧

n 发射功率

ü 31dBm时，20unit每子帧

ü 0dBm时，1unit每子帧

n GPS功率=0.08unit每子帧

n 对于D2D discovery和D2D communication和WAN signaling使用相同的值

11) R1-132769Ericsson, ST-Ericsson, ALU, ASB, Hi-Silicon, Huawei（noted）

l 与之前分析过的提案R1-132754相同，只是增加了如下一点：

l 广播是first UE向所有其他UE的单向通信

12) R1-132804Ericsson, ST-Ericsson, USDoC（noted）

l 仿真组播和广播

n 每个小区撒3 groupcasting (broadcasting) UEs

n 根据R1-132722通过的最小RSRP准则，每个组撒9个接收UE

13) R1-132810Qualcomm Inc., Huawei, HiSilicon, General Dynamics Broadband UK, III（noted）

对于unicast

l 每次仿真只能仿真一种D2D communication方式

l Unicast

n 根据TR规定的撒点准则撒first UE

n 根据TR规定的撒点准则撒second UE

ü Second UE直到满足RSRP准则才重新撒

对于groupcast

l 每个cell（宏小区的一个扇区）有3个组

l 每个组有10个UE

n 1个发射UE，9个接收UE

l 根据TR定义的撒点准则撒发射UE

l 根据TR定义的撒点准则撒接收UE

n 接收UE直到满足RSRP准则才重新撒

对于broadcastoption 1

l 每个小区有3组broadcast sessions

l 每broadcast sessions有10个UE

n 1个发射UE，9个接收UE

n 其撒点方式同groupcast

对于broadcastoption 2

l 每个cell有3个broadcast transmitters

l 每个小区有24个UEs

n 3个发射UEs，剩下的是接收UEs

14) R1-132750LG Electronics, LG Uplus, Qualcomm（noted）

l 为了保证不被覆盖UE有足够的数量，建议使用ITU信道模型

Table 1 Percentage ofout-of-coverage UE (uniform UE drop, ISD=1732m)

Case	Percentage of out of coverage UE
Case 1) ITU channel, with wraparound	39.1%
Case 2) ITU channel, without wraparound	57.0%
Case 3) 3GPP channel, with wraparound	0.9%
Case 4) 3GPP channel, with wraparound	10.4%

l 对于部分覆盖场景，建议使用19小区3扇区，并且使用wraparound，即下图所示

Figure 1 Proposed 3-site clustered eNB enabling pattern

Channel Model

1) R1-131895 CATT（noted）

l 该提案总结了信道模型

Table 1:Summary of channel model for I2O link

Case		Pathloss and Penetration Loss [dB] (d in m, f_c in GHz)	Shadow fading	Applicability range, default values	Fast Fading
Indoor-indoor		InH as baseline: PL_LOS = 16.9log₁₀(d) + 32.8 + 20log₁₀(f_c) PL_NLOS = 43.3log₁₀(d) + 11.5 + 20log₁₀(f_c) Reference for optional penetration loss: k*d+ 18.3 n ^{((n+2)/(n+1)-0.46)} n is the number of floors, the value of k is FFS;	Log-normal distribution with σ=3dB(LOS)/4dB(NLOS)	3m<d<100 m 10m<d<150 m	InH (LOS/NLOS)
Outdoor-outdoor		ITU-R P.1411-6 with p=50: PL_LOS = 20 log₁₀(d) + 32.45 + 20 log₁₀(f_c) PL_NLOS = 40log₁₀(d) + 24.5 + 45log₁₀(f_c) + L_urban Upper limit of LOS probability:	LOS: Log-Rayleigh distribution with s= 7dB NLOS: Log-normal distribution with s= 7dB	L_urban = 6.8 dB	ITU UMi (LOS/NLOS)
Indoor-outdoor	I2O pathloss	PL = PL_b+ PL_bw + PL_in and PL_b= PL_b (d_out+ d_in) PL_tw = 14 + 15(1-cos(θ))² PL_in= 0.5d_in	Log-normal distribution withs= 7dB	3m<d_out+d_in<1 000 m, 0m<d_in<25 m,	ITU-UMi (LOS/NLOS)
	PL_b for h_IN≥ 7.5 m	For LOS: PL_b=max(PL_free-space, PL_b1) PL_b1 = 22.7log₁₀(d) + 27.0 + 20log₁₀(f_c) PL_b1 = 40log₁₀(d) + 7.56 – 17.3log₁₀(h_IN) –17.3log₁₀(h_OUT) + 2.7log₁₀(f_c) For NLOS: f_c: 0.45-1.5 PL_b1 = (44.9 - 6.55log₁₀(h_OUT))log₁₀(d) + 5.83log₁₀(h_OUT) + 16.33 + 26.16log10(f_c) – 0.8 h_IN f_c: 1.5-2.0 PL_b1 = (44.9 - 6.55log₁₀(h_OUT))log₁₀(d) + 5.83log₁₀(h_OUT) + 14.78 + 34.97log10(f_c) – 0.8 h_IN f_c: 2.0-6.0 PL_b1 = (44.9 - 6.55log₁₀(h_OUT))log₁₀(d) + 5.83log₁₀(h_OUT) + 18.38 + 23log10(f_c) – 0.8 h_IN		3m< d < d’_BP¹⁾ d’_BP<d<5000m 10m<d<2000 m h_IN =(n-1)N+1.5 m, h_OUT* = 1.5 m n >1 is the number of floors; N is the height of one floor;
	PL_b for h_IN= 1.5 m	For LOS: PL_b = 20 log₁₀(d) + 32.45 + 20 log₁₀(f_c) For NLOS: PL_b = 40log₁₀(d) + 24.5 + 45log₁₀(f_c) + L_urban L_urban = 6.8 dB		h_IN=1.5m, h_OUT=1.5 m

2) R1-131922 IntelCorporation（noted）

l 对于链路级分析，在研究D2D discovery和communication时，建议使用EPA-5Hz和ETU-30Hz模型

l 使用下表中的信道模型参数来进行D2D系统级仿真

Table 1.Recommended UE-UE channel characteristics for D2D studies

O2O UE-UE link	General description	Pathloss, shadow fading based on ITU-R P.1411-6 [4] for urban area case. Large scale parameters (except for shadowing standard deviation) and their cross-correlation values as for ITU-R UMi LOS and NLOS in Table B.1.2.2.1-4 of 3GPP TR 36.814 [6]. Small scale channel modeling (except those listed below) based on SCM model for ITU-R UMi LOS and NLOS [6].
	Distance dependent path-loss	Distance d is given in m, frequency f in MHz
	Shadowing standard deviation	LOS: , log-normal distribution NLOS: , log-normal distribution
	Shadowing correlation	As proposed in [3].
	LOS Probability	Modified model exponentially decaying at large distances
	Fast fading channel model	Modified ITU-R UMi channel model: · Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values). AOD/AOA spread (log10(degrees)): o LOS: μ = 1.75, σ = 0.19; o NLOS: μ = 1.84, σ = 0.15. · Doppler Spread o Low mobility scenario, UE speed is 3km/h
O2I UE-UE link	General description	Pathloss, shadow fading, based on WINNER+ O2Ia channel model [5]. Large scale parameters (except for shadowing standard deviation) and their cross-correlation values as for ITU-R UMi O2I in Table B.1.2.2.1-4 of 3GPP TR 36.814 [6]. Small scale channel modeling (except those listed below) based on SCM model for ITU-R UMi O2I in Table B.1.2.2.1-4 of 3GPP TR 36.814 [6].
	Distance dependent path-loss	L_tw - penetration loss; - effective BS and UE antenna heights accordingly - BS and UE antenna heights accordingly Distances are given in m, frequency f in GHz.
	Shadowing standard deviation	LOS: , log-normal distribution NLOS: , log-normal distribution
	Shadowing correlation	As proposed in [3].
	LOS Probability	Modified model exponentially decaying at large distances
	Penetration Loss	Frequency f in GHz
	Fast fading channel model	Modified ITU-R UMi O2I channel model: · Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values). AOD/AOA spread (log10(degrees)): o O2I: μ = 1.76, σ = 0.16. · Doppler Spread o Low mobility scenario, UE speed is 3km/h
I2I UE-UE link (same building)	General description	Pathloss, large scale parameters and their cross-correlation values as in [6] for ITU-R InH. Small scale channel modeling based on SCM model for ITU-R InH [6].
	Distance dependent path-loss	Distances are given in m, frequency f in GHz.
	Shadowing standard deviation	LOS: , log-normal distribution NLOS: , log-normal distribution
	Shadowing correlation	As proposed in [3].
	LOS Probability	Modified model exponentially decaying at large distances
	Fast fading channel model	Modified ITU-R UMi InH channel model: · Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values). AOD/AOA spread (log10(degrees)): o LOS: μ = 1.62, σ = 0.22. o NLOS: μ = 1.77, σ = 0.16. · Doppler Spread Low mobility scenario, UE speed is 3km/h
I2I UE-UE link (different buildings)	General description	Pathloss, shadow fading, based on WINNER+ O2Ia channel model [5]. Pathloss equation has the following modifications: - Penetration loss is added to reflect the signal propagation through the second wall - Indoor distance for the second indoor propagation part is generated Large scale parameters (except for shadowing standard deviation) and their cross-correlation values as in [6] for ITU-R UMi O2I. Small scale channel modeling (except those listed below) based on SCM model for ITU-R UMi O2I [6].
	Distance dependent path-loss	- total indoor distance L_tw - penetration loss; - effective BS and UE antenna heights accordingly - BS and UE antenna heights accordingly Distances are given in m, frequency f in GHz
	Shadowing standard deviation	LOS: , log-normal distribution NLOS: , log-normal distribution
	Shadowing correlation	As proposed in [3].
	LOS Probability	Modified model exponentially decaying at large distances
	Penetration Loss	Frequency f in GHz
	Fast fading channel model	Modified ITU-R UMi O2I channel model: · Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values). AOD/AOA spread (log10(degrees)): o I2I: μ = 1.76, σ = 0.16. · Doppler Spread Low mobility scenario, UE speed is 3km/h

l 对于Macro-UE和RRH-UE链接，建议使用RAN1 WG中的信道模型，在表1和2中列出

Table 1.Proposal on large scale propagation characteristics for proximity studies inagreed D2D scenarios

UE-UE

O2O

Pathloss:

According to Section 4.3 of ITU-R P.1411-6 [4]

Shadow fading:

According to Section 4.3 of ITU-R P.1411-6 [4] with the following modifications: for LOS case lognormal distribution is used

Probability of LOS:

According to Table 2 from the companion contribution [2]

O2I

Pathloss, Shadow fading:

According to scenario O2Ia from Table 4-1 of WINNER+ [5]

Probability of LOS:

According to Table 2 from the companion contribution [2]

I2I
(same building)

Pathloss, Shadow fading:

According to ITU-R InH (Table B.1.2.1-1 of 3GPP TR 36.814 [6])

Probability of LOS:

According to Table 2 from the companion contribution [2]. The formula is a modified equation for ITU-R InH scenario from Table B.1.2.1-2 of 3GPP TR 36.814 [6]

I2I
(different buildings)

Pathloss, Shadow fading:

According to scenario O2Ia from Table 4-1 of WINNER+ [5] with the following modifications (Table 5 of accompanied contribution [2]):

- Penetration loss is added to reflect the signal propagation through the second wall

- Indoor distance for the second indoor propagation part is generated

Probability of LOS:

According to Table 2 from the companion contribution [2]

Macro eNB-UE

O2O

Pathloss, Probability of LOS, Shadow fading:

According to 3GPP Macro + Indoor RRH/Hotzone scenario. See Table A.2.1.1.5-1 of 3GPP TR 36.814 [6])

O2I

Pathloss, Penetration Loss, Probability of LOS, Shadow fading:

According to 3GPP Macro + Indoor RRH/Hotzone scenario. See Table A.2.1.1.5-1 of 3GPP TR 36.814 [6])

RRH-UE

I2O

Pathloss, Shadow fading:

According to scenario O2Ia from Table 4-1 of WINNER+ [5]

Probability of LOS:

According to Table 2 from our companion contribution [2]

I2I

(same building)

Pathloss, Probability of LOS, Shadow fading:

According to ITU-R InH (Table B.1.2.1-1 of 3GPP TR 36.814 [6])

I2I

(different buildings)

Pathloss, Shadow fading:

According to scenario O2Ia from Table 4-1 of WINNER+ [5] with the following modifications (Table 5 of contribution [2]):

- Penetration loss is added to reflect the signal propagation through the second wall

- Indoor distance for the second indoor propagation part is generated

Table 2. Proposal on small scale channel models forsystem level proximity studies in agreed D2D scenarios.

UE-UE
O2O	According to SCM model parameters for ITU-R UMi LOS and NLOS from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6] with the following modifications: - Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values)
O2I	According to SCM model parameters for ITU-R UMi O2I from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6] with the following modifications: - Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values)
I2I (same building)	According to SCM model parameters for ITU-R InH from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6] with the following modifications: - Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values)
I2I (different buildings)	According to SCM model parameters for ITU-R UMi O2I from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6] with the following modifications: - Aligned AOD and AOA statistical distributions (mean and std. deviation values for AOD are set equal to the AOA values)
Macro eNB-UE
O2O	According to SCM model parameters for ITU-R UMa from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6] for LOS and NLOS accordingly
O2I	According to SCM model parameters for ITU-R UMa NLOS case from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6] for both LOS and NLOS cases
RRH-UE
I2O	According to SCM model parameters for ITU-R UMi O2I from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6]
I2I (same building)	According to SCM model parameters for ITU-R InH from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6]
I2I (different buildings)	According to SCM model parameters for ITU-R UMi O2I case from Table B.1.2.2.1-4 of 3GPP TR 36.814 [6]

3) R1-132622 AniteTelecoms Ltd., Elektrobit Corporation（noted）

l 对于D2D O2O部分，建议使用ITU-R IMT-Advanced 3D模型

4) R1-132501 QualcommIncorporated, US Department Of Commerce, General Dynamics Broadband UK（noted）

l 对于O2O大尺度衰落，使用ITU-1411-6模型

l 对于I2I，使用3GPP InH模型

l 使用winner+O2I模型

l 对于所以D2D场景，使用ITU-R UMi LOS

l 使用简单的空间损耗模型

l 使用36.184中的小尺度衰落，并有微小的修改

5) R1-132030 Ericsson,ST-Ericsson（noted）

l 对于不同的D2D communication场景，不同的信道模型总结如下：

Considered scenario		Pathloss model	Shadowing	Fast fading
Both devices indoors	Same floor	, dB/m	lognormal	InH (LOS)
	Same building, different floor	, dB , dB	lognormal	InH (NLOS)
	Different buildings	Berg’s recursive model (first and last hop by expression above) dB , dB/m , dB and dB	lognormal	InH (NLOS)
One device outdoors and one device indoors		Berg’s recursive model (Outdoor to indoor hop by expression above for exterior walls j = 1-4) dB , dB/m dB and dB	lognormal	InH(NLOS)
Both devices outdoors		Berg’s recursive model (Section 4)	lognormal	UMi

6) R1-132341 NEC Group（noted）

l 建议使用以下信道模型

Table 1: Proposed models for D2D channel

	Indoor to Indoor	Outdoor to Outdoor	Indoor to Outdoor/ Outdoor to Indoor
Pathloss	Dual-strip model [4] or the proposed model in the following equation	ITU-R P.1411-6 [5]	Dual-strip model [4]
Shadowing	Log-normally distributed with standard deviation 4dB (NLOS)	Log-normally distributed with standard deviation 10dB	Log-normally distributed with standard deviation 8dB
Shadow fading correlations	FFS	FFS	FFS
Fast fading	TGn channel models [6]	3GPP Interim channel model [7]	FFS

l 在RAN1中应该深入研究通信双方都移动的新型多普勒效应

l D2D链路级仿真的多普勒功率谱使用以下公式：

l D2D系统级仿真时空间信道模型（spatial channel model）使用以下公式：

7) R1-132247 LGElectronics（noted）

l 对于I2I，建议使用InM模型（TR36.814）

l 对于O2O，建议使用ITU-R P1411-6 model (p=50)

l 对于I2O，建议使用Winner II -A2模型

l 对于700MHz载波，-9dB offset，使用TR 36.814和TR 36.828提出的eNB-to-UE信道模型

8) R1-132316 Nokia, NokiaSiemens Networks（noted）

l 当选择信道模型时，应该考虑spatial consistency。路损和相近链路的相关性是D2D信道模型的重要特点

l 对于O2I路损，建议使用TR36.828室外UE-HeNB

l 对于I2I路损，建议使用TR36.828室内UE-HeNB

l 对于O2O路损，建议使用ITU-R P.1411-6 (Section 4.3) 中提出的low height model， p = 50 and L_urban=-8dB

l 建议增加对数正态阴影，O2O使用7dB，O2I和I2I使用4dB

l 如果RAN1同一对相关性建模，建议使用简单模型，与蜂窝场景下相同，在链路末端取和，，SF_UEi是从相同2D map取得

l RAN1应该继续研究多普勒效应，D2D discovery和communication性能的初步评估可能不需要对快衰落建模

9) R1-132502 Qualcomm, USDepartment of Commerce, General Dynamics Broadband UK, ITRI, Fujitsu, Intel,Samsung, ZTE, III（noted）

l 对于三种情况的建模

10) R1-132760Qualcomm, Huawei, HiSilicon, ITRI, ZTE, CATT, Samsung, US Department ofCommerce, General Dynamics Broadband, III（noted）

l 对于三种情况的建模

11) R1-132621Anite Telecoms Ltd., Elektrobit Corporation（noted）

l 讨论了P.1411-6对于D2D仿真的适用性，建议使用IMT-Advance而不是P.1411-6

12) R1-132761Qualcomm, ANITE, Intel, Fujitsu, US Department of Commerce, General DynamicsBroadband, III（noted）

l D2D多普勒建模：修改ITU-R UMi/InH模型使其适用于双向移动

13) R1-132745NEC, Ericsson, ST-Ericsson, Huawei, HiSilicon（noted）

l 讨论双向移动对链路级和系统级性能的影响

l 讨论在发射机和接收机的散射对链路级和系统级性能的影响

l 如果需要，修改链路级仿真的多普勒频谱，至少在UE移动超过3km/h时考虑上双向移动

l 如果需要，因为天线高度较低，修改现在的3GPP SCM或者定义新的SCM

14) R1-132762Qualcomm, Intel, US Department of Commerce, General Dynamics Broadband, III（noted）

l From R1-132390, Intel

l dcorr = 10m

l dTh1 = 1m

l dTh2 = 13.5m

15) R1-132803Qualcomm, Anite, RIM, Intel, III, General Dynamics Broadband（noted）

l 该提案是R1-132761的完善，增加了以下两条：

n Direction of Travel(velocity vector) independent and random

n Doppler is determinedby path AOA/OA

D2D Discovery

1) R1-132371 NTT DOCOMO（noted）

l D2D发射和接收使用相同的频带

l D2D UE不能使用全双工方式

l 研究在/不在网络覆盖场景的同步问题

2) R1-132028 Ericsson,ST-Ericsson（noted）

l D2D复用上行资源（FDD 上行频谱，TDD 上行子帧）

l 考虑引进新的数据和控制信道/信号

l 分配上行子帧资源子集的一个子集给D2D

l 用户只能在分配的资源中传输D2D信号

l D2D复用上行蜂窝TA

l 为了满足各种场景的需求，需支持异步discovery

l 用户能解码异步信标

l 至少在public safety场景下，connection-less通信应作为D2D communication的一个候选方案

3) R1-132413 Huawei,HiSilicon（noted）

l RAN1应研究现有的上行信号是否适合基站定向发现（eNB-directed discovery）

l RAN1应研究对discovery信号如何最恰当的选择时域/频域资源

l RAN1应研究如何最优化复用discovery资源，网络采用什么机制可以最好的评估和控制复用

4) R1-132503 QualcommIncorporated（noted）

l Discovery使用所有的PUSCH

l Discovery信号使用PUSCH的调制和编码

l 对于Open ProSe and Restricted ProSediscovery使用相同的物理层传输方案

l Discovery资源的选择可以由UE或者网络完成

l Discovery子帧应使用下行链路timing

l Discovery资源的跳频需要降低远近效应和半双工问题的影响

l 使discovery子帧与少数上行链路子帧交织

5) R1-132726 QualcommInc., General Dynamics Broadband UK, US Dept.of Commerce, Fujitsu, Intel, III（noted）

l 对于FDD部署时，discovery需要使用上行链路频谱

l Discovery的设计应包括RRC_IDLE UEs

l Direct discovery设计应使预留资源以半静态方式周期性的分配

6) R1-132772 QualcommInc., Huawei, HiSilicon, LG Electronics, LG Uplus, Intel, Samsung, NEC, RIM,ETRI（noted）（三种discovery方式，重要）

l Type 1: a discoveryprocedure where resources for discovery signal transmission are allocated on anon UE specific basis

Note: Resources can befor all UEs or group of UEs

l Type 2: a discoveryprocedure where resources for discovery signal transmission are allocated on aper UE specific basis

n Type 2A: Resources areallocated for each specific transmission instance of discovery signals

n Type 2B: Resources arepersistently allocated for discovery signal transmission

RAN1#74

D2D Communication

1) R1-133908 QualcommInc., US Department of Commerce, Ericsson, ST-Ericsson, General DynamicsBroadband UK, ZTE（noted）

l RAN1 #74会议重点研究D2D communication

l D2D discovery也会研究，包括Public Safety和General Scenario

l 对于D2D communication，RAN1重点研究广播D2D communication，并将其作为Rel-12的最高优先级

l 在Rel-12中，RAN1不研究组播和单播D2D communication

2) R1-13030 CATT（noted）

l D2D communication应该在TDD模式下运行

l 在网络覆盖下时，D2D communication应运行在上行资源上

3) R1-13031 CATT（noted）

l 在研究D2D communication时，应研究cluster head assistant manner和contention based manner

4) R1-133911 Huawei,HiSilicon, Fujitsu, KDDI, Intel, Panasonic（noted）

l 对于in-network、partial network和out-of network，提出一个统一的框架

l Group master有以下功能：timing、资源分配、组内管理

l 对于in-network，群主是基站

l 对于out-of network，群主是UE

n 讨论是否所有PS UE或者只有一些高容量的PS UE可以作为群主

n 注意：增强D2D通信协议可能与基站与用户间的通信协议有区别

n 讨论如何选择群主

l 对于partial network，讨论如何选择群主

l 注意：可以有多个群主

5) R1-132898 Huawei,HiSilicon（noted）

l 修改评价方法，以便组播UE占用小片区域

6) R1-133146 ZTE（noted）

l 分析Public safety广播的详细情景，以方便进一步研究潜在的解决方案

l Public safety专用频谱使用TDD双工模式

l Public safety广播的资源分配应该预先配置或者半静态配置

l 在子帧水平，public safety广播和单播的资源是TDM

7) R1-133988 QualcommIncorporated, Nokia, NSN（noted）

l W = 3 [dB]、X = 6 [dB]、Y = 3 [dB]、Z = 3 [dB]

D2D Discovery

1) R1-132900 Huawei,HiSilicon（noted）

l OFDM和SC-FDMA之间没有太大的差别

l 由于每个PRB的信道估计性能，频率分集不合适在低SNR

l 频率分集可以改善在中、高信噪比时的结果

2) R1-133034 CATT（noted）

l Discovery的设计应考虑时间同步的问题

l 对于时间同步，进一步研究single-point-synchronization机制和UE-specific synchronization机制

l 对于不同的discovery设计，进一步研究频率同步错误是否不是问题

3) R1-133600 QualcommIncorporated（noted）

l Open ProSe和Restricted ProSe使用相同的物理层和MAC层设计，Restricted ProSe的discovery信息应该被适当的编码

l Discovery的Network reserves周期性发生在上行链路子帧上

l 在out- of- network/partial-network情况下，需预先配置

l Discovery期间，子帧使用所有的PUSCH

l PUSCH应为频率复用，discovery资源包括一对RB

l Discovery信号应再利用PUSCH调制和编码

l 资源选择应基于discovery资源的接收能量

l Discovery子帧应使用下行链路timing

l 采用Discovery资源跳频来降低远近效应和半双工问题的影响

l Discovery子帧和小部分上行链路子帧交织，这个信息应该基于Proposal 2a机制被添加到SIB

4) R1-133972 NTT DOCOMO,Ericsson, ST-Ericsson, Qualcomm, LG Electronics（noted）

l 对于小区间discovery，同步和异步部署都应该被研究

5) R1-133980 QualcommInc., Fujitsu, General Dynamics Broadband UK, BlackBerry（noted）

l 对于discovery，周期性上行链路资源以半静态方式分配

l 一个分配周期内的discovery资源被分成时间-频谱资源，至少以FDM或者TDM分

l UE发射和接收discovery信号使用半双工方式

l Discovery传输可以使用X bits（128 bits）的信息或者物理信号（PRACH、SRS、PSS/SSS）

l 继续研究信号传输室SC-FDM还是OFDM

RAN1#74b

D2D Communication

Synchronization and Timing

1) R1-134624 QualcommIncorporated（noted）

l 低占空比的PSS传输作为链路层信号

l 不同UEs同步信号的时间正交，以便被更好的检测到

l 支持解决冲突的同步帧

l 使用非线性算法，多用户时间同步选择算法

2) R1-134306 ZTE（noted）

l 在部分覆盖和无覆盖情况下，考虑分布式本地同步，在覆盖情况下，考虑小区间同步

l 今后研究集中全局同步

l 如果没有集中全局同步，建议使用单独同步信号以提供同步基准和携带discovery消息或communication控制信息的提示

l 使用基站head配置UE-SpecificTA

Physical Channel Design

1) R1-134344 Huawei,HiSilicon（noted）

l 相比于LTE蜂窝链路，尽量减少D2D链路物理层设计的改变

l D2D communication使用OFDM

l 至少预留一个符号用作保护时间

l D2D链路的PRB结构与LTE的相同

l D2D链路可以复用TDM或者FDM的PDCCH

2) R1-134625 Huawei,HiSilicon（noted）

l 结论如下：

Design Issues	Proposed Solution
Method for achieving synchronization	See [2]
Method for scheduling and resource allocation	See [3]
Closed loop physical layer feedback	None
Physical channel design	Reuse uplink/PUSCH signal structure
MCS selection	Fixed but configurable
H-ARQ	Blind retransmissions – possibly chase combining only
Power setting	Max power transmissions
Handling of interference coordination	See [2]
Method of multiplexing D2D and Uu	Should be studied further

Method of Resource Allocation

1) R1-134113 CATT（noted）

l 如果D2D链路使用FDM，需要研究克服远近效应的方案

D2D Discovery

Discovery Signal Design

1) R1-134629 Qualcomm Inc（noted）

l 允许RRC_IDLE UEs参与discovery消息，不用提前timing

2) R1-134502 NTT DOCOMO,INC（noted）

l 如果不能合适的设置频率偏移量，则会显著降低系统性能

l 在接收端频域补偿可以减小由频率偏移量带来的性能下降

3) R1-134075 Huawei,HiSilicon（noted）

l 无论SINR target是多少，所需比特数对占用的PRBs数目有巨大影响

4) R1-134628 QualcommIncorporated（noted）

l 对于open和restricted discovery，基于use message的信号应复用PUSCH

5) R1-134355 Huawei,HiSilicon（noted）

l 基于OFDM与SC-FDMA的discovery信号没有明显的性能区别

l 在低SNR水平上，使用2个PRBs来传输discovery信号似乎更加合适

l 无论SINR target是多少，所需比特数对占用的PRBs数目有巨大影响

l 链路级仿真参数表

Table1. Link level simulation parameters assumptions

Bandwidth	10MHz
Carrier frequency	2G Hz
Channel model	AWGN ETU: The delay profiles refer to 36.101 Table B.2.1-2
MIMO configuration	1x1for AWGN 1x2 with low correlation for ETU
Modulation	QPSK
Channel coding	CC with 16-bit CRC
UE receiver	MMSE-IRC
Channel estimation	Ideal or realistic
Noise estimation	realistic
Received timing delay (us)	0
Frequency offset (Hz)	0
UE speed	3km/h

6) R1-134074 Huawei,HiSilicon（noted）

l 评估时使用NLOS模型

l 基站控制着discovery资源的数目，可以在一个discovery周期内改变discovery子帧的数目

l 研究跳频对discovery性能的影响

l Power hopping考虑作为发送discovery信号的备选方案

l 将距离接近的UE分为一组，在组内进行资源分配，这样可以在一定discovery范围内快速的discovery大量的UEs。UE分组和资源划分应被视为候选discovery方案

7) R1-134627 QualcommIncorporated（noted）

l Open和restricted discovery情况下，Discovery信号需要传输message

l Discovery信号应使用SC-FDMA，并复用PUSCH参考信号，调制和编码

l Discovery需使用下行链路timing，对于同步部署，UE可以使用最早的下行链路timing

l 对于异步部署，UE可以使用所联系的宏基站的下行链路timing来进行小区内discovery

l 下面附上重要结果

Results for GeneralScenario Synchronous Deployment

LayoutOption 1:

Figure A1.1 Average Number of UEs discovered versusnumber of discovery periods

Figure A1.1 plots the number of UEsdiscovered using decoding only and using decoding & partial matching. Onaverage 240 UEs can be discovered after 40 discovery periods using decodingonly, and on average 310 UEs can be discovered after 40 peer discovery periods,using decoding & partial matching.


Figure A1.2 CDF of Number of UEs discovered using decoding only	Figure A1.3 CDF of Number of UEs discovered using decoding & partial matching

Figure A1.2 plots CDF of number of UEsdiscovered as a function of time using decoding only, the step behaviour isbecause there are 80% indoor UEs. The median point of number of UEs discoveredafter 40 discovery periods is 142. Figure A1.3 plots CDF of number of UEsdiscovered as a function of time for decoding & partial matching. We againobserve the step behaviour because of the 80% indoor UEs. The median point ofnumber of UEs discovered after 40 discovery periods is 167.


Figure A1.4 Probability of discovery versus pathloss using decoding only	Figure A1.5 Probability of discovery versus pathloss using decoding & partial matching

Figure A1.4 plots probability of discoveryversus pathloss using decoding only, at 143.5dB pathloss there is 10%probability of discovery after 40 discovery periods. Figure A1.5 plots theprobability of discovery versus pathloss using decoding & partial matching;note that at 148.5dB pathloss there is 10% probability of discovery after 40discovery periods.

Note that there is non-monotonic behaviourof probability of discovery versus pathloss curve in the indoor-outdoordeployment. This is the combined effect of different types of UEs, i.e., theUEs outdoor, UEs inside buildings and UEs virtually indoor. (See [1] fordetailed explanation of the non-monotonic behaviour.)

LayoutOption 3:

Figure A1.9

Figure A1.9 plots the number of UEsdiscovered using decoding only and using decoding & partial matching. Onaverage 1052 UEs are discovered after 40 discovery periods in the case ofdecoding only, and on average 1332 UEs are discovered after 40 discoveryperiods in the case of decoding & partial matching.


Figure A1.10 CDF of Number of UEs discovered using decoding only	Figure A1.11 CDF of Number of UEs discovered using decoding & partial matching

Figure A1.10 plots CDF of number of UEsdiscovered as a function of time using decoding only. The median point ofnumber of UEs discovered after 40 discovery periods is 1052. Figure A1.11 plotsCDF of number of UEs discovered as a function of time using decoding &partial matching. The median point of number of UEs discovered after 40discovery periods is 1332.


Figure A1.12 Probability of discovery versus pathloss using decoding only	Figure A1.13 Probability of discovery versus pathloss using decoding & partial matching

Figure A1.12 plots probability of discoveryversus pathloss using decoding only, at 135.5dB pathloss after 40 discoveryperiods there is 10% probability of discovery. Figure A1.13 plots theprobability of discovery versus pathloss using decoding & partial matching,at 139.5dB pathloss there is 10% probability of discovery after 40 discoveryperiods.
Results for General ScenarioAsynchronous Deployment

LayoutOption 1:

Figure A2.1 Average Number of UEs discovered versusnumber of discovery periods

Figure A2.1 plots the number of UEsdiscovered using decoding only and using decoding & partial matching. Onaverage 191 UEs can be discovered after 40 discovery periods using decodingonly, and on average 231 UEs can be discovered after 40 peer discovery periods,using decoding & partial matching.


Figure A2.2 CDF of Number of UEs discovered using decoding only	Figure A2.3 CDF of Number of UEs discovered using decoding & partial matching

Figure A2.2 plots CDF of number of UEsdiscovered as a function of time using decoding only, the step behaviour isbecause there are 80% indoor UEs. The median point of number of UEs discoveredafter 40 discovery periods is 129. Figure A2.3 plots CDF of number of UEs discovered as a function of timeusing decoding & partial matching, the step behaviour is because there are80% indoor UEs. The median point of number of UEs discovered after 40 discoveryperiods is 155.


Figure A2.4 Probability of discovery versus pathloss using decoding only	Figure A2.5 Probability of discovery versus pathloss using decoding & partial matching

Figure A2.4 plots the probability ofdiscovery versus pathloss using decoding only, at 145.5dB pathloss there isstill 10% probability of discovery. Figure A2.5 plots probability of discoveryversus pathloss using decoding & partial matching, at 150.8dB pathlossthere is still 10% probability of discovery.

Note that the non-monotonic behaviour ofprobability of discovery versus pathloss curve is because the probability ofdiscovery is averaged across indoor, virtual indoor and outdoor UEs, while theydiffer in the interference experienced due to in-band emissions. Similar to theresult presented in Appendix A.1 (Figures A1.6 through A1.8), the probabilityof discovery versus pathloss observed individually for the indoor, virtualindoor, and outdoor UEs is monotonic, and is not presented here for brevity.

LayoutOption 3:

Figure A2.6 Average Number of UEs discovered versusnumber of discovery periods

Figure A2.6 plots the number of UEsdiscovered using decoding only and using decoding & partial matching. Onaverage 1037 UEs can be discovered after 40 discovery periods using decodingonly, and on average 1309 UEs can be discovered after 40 peer discoveryperiods, using decoding & partial matching.


Figure A2.7 CDF of Number of UEs discovered using decoding only	Figure A2.8 CDF of Number of UEs discovered using decoding & partial matching

Figure A2.7 plots CDF of number of UEsdiscovered as a function of time using decoding only. The median point ofnumber of UEs discovered after 40 discovery periods is 1042. Figure A2.8 plotsCDF of number of UEs discovered as a function of time using decoding &partial matching. The median point of number of UEs discovered after 40discovery periods is 1315.


Figure A2.9 Probability of discovery versus pathloss using decoding only	Figure A2.10 Probability of discovery versus pathloss using decoding & partial matching

Figure A2.9 plots probability of discoveryversus pathloss using decoding only, at 139.7dB pathloss there is still 10%probability of discovery. Figure A2.10 plots probability of discovery versuspathloss using decoding & partial matching, at 143.4dB pathloss there isstill 10% probability of discovery.

8) R1-134919 Qualcomm,General Dynamics Broadband UK, Samsung, Intel, LG Electronics, ETRI, Kyocera（noted）

l 对于同步和异步部署，所有UE应该使用固定的偏移量，设置T2为恒定值，具体数值有待研究

l Discovery信号可以被配置为normal CP也可以extended CP

Method of Resource Allocation

1) R1-134884 CATT,Ericsson, ETRI, Intel, Huawei, HiSilicon, Samsung, ZTE（noted）

l Discovery资源配置包括一定数量的子帧和一个discovery周期

l 在D2D发射端UE处进行discovery资源选择

2) R1-134918 Qualcomm,General Dynamics Broadband UK, Fujistu, LG Electronics, ZTE, Samsung,ETRI, DoCoMo, Kyocera（noted）

l 对于同步部署，discovery资源分配可以是整个共同部署

l 对于同步部署，不同的基站分配不同资源

RAN1#75

D2D Communication

Synchronization Design

1) R1-135114 IntelCorporation（noted）

l 定时估计(timingestimation)对多径效应很敏感，尤其是在低SNR的区域中；当SNR为-2 dB 左右时，只采用PSS同步序列能达到可接受的时间估计性能。

l 在缺少频率偏移补偿的条件下，大的频率偏移会导致初始时间估计性能严重降低；引入频率偏移会使初始时间估计变得更复杂

l 基于一次性的频率估计剩余频率偏移量的估计误差(residual frequency offset error)需要经过后来重复的D2DSS信号的校正。

l 对于单跳的传播场景

n 在均匀或热点用户分布的场景下，在小区覆盖的范围内，I-SS（独立同步源）的数量的平均值与小区数量相近。

n 对于室内-室外混合的场景：

u 室内I-SS数量的平均值与建筑的数量大致相等

u 几乎所有的虚拟室内用户(virtualindoor UE)都用作一个I-SS

u 少部分的室外UE用作I-SS

n 对于两跳的场景，I-SS会有所减少

l 对于两跳的传播场景

n 能够检测到的最大的I-SS数量不超过4

n 根据不同的场景，会有10%到20%的用户不在I-SS的同步范围之内

n 检测到的timing的数量根据不同的场景会有不同的变化

u 在均匀或热点用户分布的场景下，大部分UE(大于90%)检测到少于10个timing

u 对于室内-室外混合的场景，大部分室内的UE(大于90%)检测到少于4个timing；而室外的检测到的timing的数量的平均数为10

l 减小D2DSS传输数量、拓展局部的同步范围以及减少终端timing数量的机制还需进一步研究

2) R1-135219 Samsung（noted）

l 系统级的同步可以用来解决资源分配和干扰的问题

l 同步源UE之间的同步可以通过它们之间的直接的互操作或是间接的通过另一个同“志愿的”同步源UE来实现

l 如果D2D UE的操作是基于同步源UE的参考信号而不是“志愿的”同步源UE，则系统的稳定性能能得到保证

l 在无网络覆盖范围内，位于同一位置的（co-located，这里指的是在互相的同步区域内）同步源UE需要进行同步以支持D2D的广播通信

l 应该允许D2D UE成为一个“志愿的”同步源以支持无隐藏节点（hidden node）的系统级的同步

l D2D UE应该同步于一个同步源UE而不是一个“志愿的”同步源UE

l 当同步源UE检测相邻的同步源UE的时间，其接收时间应该被分离出来而忽略它自己的参考时序

l 同步源UE应尽力保持它自己的传输的参考时序，同时也要按相邻的D2D的参数信号去调整它自己的参考时序，以达到同步源UE之间的同步

3) R1-135316 Qualcomm Inc（noted）

l 基于同步簇头（synchronizationcluster head）的两跳同步会导致多个不同步的时序在同一区域出现

l 通过增加同步的跳数，能够减少不同步的时序的数量

Proposal:

l 大于两跳的同步应该引入到WF中去

图1 两跳同步协议

图2 四跳同步协议

图3 提高跳数少的时间同步的优先级（三跳同步协议）

图4 无网络覆盖的室内/外的分布

Average Number of SCHs
Protocol	2-hop	3-hop	4-hop
Avg no	306	134.7	62.5

表1SCH(Synchronization Cluster Head)的数量

Average Number of Acquired Tx Timings in Vicinity
Protocol		2-hop	3-hop	4-hop
Within 1 hop	All UEs	3.4	2.2	1.5
	Indoor UEs	1.3	0.7	0.4
	Outdoor UEs	11.9	8.6	5.7
	Virtually Indoor UEs	1.7	0.8	0.7
Within 2 hops	All UEs	23.7	15.3	9.3
	Indoor UEs	13.8	9.4	5.9
	Outdoor UEs	61.4	38.1	21.8
	Virtually Indoor UEs	18.9	12.8	8.1

表2

图5 不同的单次timing传输的数量分布

图6无网络覆盖的室的热点分布

Average Number of SCHs
Protocol	2-hop		3-hop		4-hop
Avg no	55.8		5.9		1.9
Average Number of Acquired Tx Timings in Vicinity
Protocol		2-hop		3-hop		4-hop
Within 1 hop		11.84		3.01		0.82
Within 2 hops		46.94		4.95		0.89

图7不同的单次timing传输的数量分布

4) R1-135113 IntelCorporation（noted）

l D2D进行传输前先进行同时信号的搜索。

l 在小区覆盖范围内，eNB用作同步源

l 在一个不同步的网络中，在小区边缘的UE可发送D2DSS和PD2DSCH和传播D2D时序信息。

l 在一个同步网络中，上面一条proposal同样适用。

l D2DSS可采用PSS/SSS序列，除非是覆盖范围外的同步情况；周期性的D2DSS和PD2DSCH传输FFS。

5) R1-135014 Huawei,HiSilicon（noted）

l 在同一时刻时面簇里面只有一个UE发送同步信号

l 在小区覆盖的范围内，clusterhead应处于RRC connected状态

l 在小区覆盖的范围内，clusterhead发送的同步信号的功率由服务基站决定；在小区覆盖的范围内，同步信号的功率可以是UE的最大发送功率或一个预定的值

l 在不同步的网络中，小区间的D2D需要一个公共的timing

l 分层次的同步协议与蜂窝网络的同步流程相近

l 在分层次的同步过程中，同时限制同步的跳（hop）数，可能会产生同步的边界，即多个同步源会造成互相间的干扰

l 分面式的同步协议需要一个新的同步流程，以及会有一个比较长的收敛时间

6) R1-135368 ZTE（noted）

l 如果一个D2D UE有广播数据需要传输，则它可以自动地在预定义的资源中通过分布式的方法去传输同步信号

l 如果一个clusterhead能够传输同步信号，非cluster head的用户不能传输或者是去中继同步信号

l D2D用户需要去同时检测多个不同步的CH下发的同步信号

l 单独的同步信号能够映射到PSS/SSS序列或SRS序列中去，但是否重用Rel-8中的PSS/SSS需要进一步的研究

l 可通过复用数据的方式来支持非单独的同步信号。同步信号的密度和详细的资源映射还需要进一步的研究

l 同步信号能携带简单的D2D广播控制信息

7) R1-135317 Qualcomm Inc（noted）

l 多种不同方案性能的排序如下：TDM2 (best) > TDM1 > SFN (worst)

l 同步的功率消耗由Latency2决定，如果Latency2达到了30秒，则1秒的同步周期接近最优的性能

l 应该支持D2DSS和PD2DSCH的TDM方式的资源分配方式

(a) SFN (b) TDM1

(c) TDM2

图1 三种不同的多用户同步方式

图2部分网络覆盖的同步

图3 AWGN信道下D2DSS的SNR分布

图4 PED-A信道下的D2D的平均SNR分布

图6不同时延下的功率消耗

8) R1-135532 Huawei,HiSilicon（noted）

l D2DSS应能到达粗同步和细同步的作用，且由PD2DSS和SD2DSS组成

l 传输PD2DSS需要72个子载波

l PD2DSS采用ZC序列，且采用与传统的ZC序列不同

l SD2DSS倾向于采用72载波以及多于一个符号的格式，对于超过72个载波的需要进一步的研究

l SD2DSS可通过突发或周期的方式发送

9) R1-135175 Alcatel-Lucent,Alcatel-Lucent Shanghai Bell（noted）

l 同步信号分配到一个每一个子帧固定的无线资源位置，这样的话同步信号同时还能作为一个discovery信号

图1 3个D2D UE之间组通信（groupcommunication）的半双工时序例子

图2 D2D同步信号信道的固定的资源分布

D2D Data communication Physical Channel Design Details

1) R1-135092 CATT（noted）

l 使用表1中的物理信道类型

l 考虑到用户间的发送/接收的半双工操作，UE到CH、CH到UE、UE到UE的信道应在时间上隔离开来

l D2D的帧结构能被CH或eNB配置或预配置

n 连接部分：D2D UE能通过争夺的方式占用资源

n 连接空闲部分：D2D UE能通过被CH/eNB调度的方法占用资源

Direction	Channel type	Scheduling based manner	Contention based manner
CH to UE	D2DSS	Y	Y
	PD2DSCH	Y	Y
	Cluster head control channel	Y	N
	Cluster head data channel	Y	Y
UE to UE	D2D data channel	Y	Y
UE to CH	REQ channel	Y	N

表1 D2D物理信道的类型

图1 基于调度方式的帧结构

图2 基于争夺方式的帧结构

图3 部分覆盖的示意图

图4 D2D通信的帧结构

2) R1-135176 Alcatel-LucentShanghai Bell, Alcatel-Lucent（noted）

l D2D数据信道中使用一个固定的RVs序列，采用自动的盲和重传

l 除了PUSCH DM RS外，不使用额外参考信号

l 至少使用正常的CP。拓展的CP需要进一步的讨论

l 在每个子帧里面，无线资源按每组N_b个PRB均匀的分配

3) R1-135220 Samsung（noted）

l 在D2D广播通信不中支持分隔的物理控制信道设计

l 在一个D2D通信的资源集合里面使用TDM的复用方式时，需要控制信息去提供频域上的资源分配信息

l 使用一个固定的MCS等级足以支持广播传输，不需要L1/L2反馈以及控制自信以指示MCS等级/TBS size

4) R1-135318 Qualcomm Inc（noted）

l 通过TTI绑定、时域和频域上的分集以提高VOIP的性能

l TTI绑定能提供一个大于2dB的重传增益

l 通过频域上的交织能获得一个4dB的增益

l 通过时域上的交织能获得一个4dB的增益

Parameters	Assumptions
Carrier Frequency, System Bandwidth	700 MHz, 10 MHz
VoIP Packet Size	44 Bytes (including CRC)
Channel Model	ITU UMi NLOS (reduced variability)
Number of RX antennas	2
Coding/Modulation	Turbo/QPSK
HARQ Transmissions	Chase Combining
Channel Estimation Algorithm	Quadratic Interpolation
Transmission Bandwidth	2 PRBs

表1 仿真参数表

图1 提供链路预算的方案

图2 时域上的信道变化

图3 频域上的信道变化

图4 不同重传数的BLER结果

图5 不同TTI绑定数的BLER结果

图6不同频域交织的BLER结果

图7不同时域交织的BLER结果

图8 BLER的全部结果

5) R1-135319 Qualcomm Inc（noted）

l 图1的控制框架

l 使用DM-RS作为新的数据指示(newdata indicator)

图1 WAN和D2D的控制框架

Field	Proposal for D2D	Note
Resource information	Not needed -- pre-configured or implicit	Receiver monitors preconfigured /implicitly determined resources
MCS/RV	Not needed – pre-configured	E.g. for Chase Combining it is not needed
NDI	Needed – propose to use DMRS
DMRS offset	Not needed – pre-configured or implicit
Precoding information	Not needed	Current focus on single TX antenna
CSI request	Not needed	Precluded by RAN1 agreement
SRS request	Not needed	Precluded by RAN1 agreement
Uplink Index/DAI	Not needed
PUSCH power control	Not needed	Unclear how (and why) to do power control without physical layer feedback
0/1A differentiation	Not needed
Padding	Not needed
Identity	Not needed -- implicit	Could also be carried as part of data

表1 UL调度中DCI格式0/4中的控制元素

6) R1-135321 Qualcomm Inc（noted）

l D2D通信子帧应与WAN的上行子帧相似，然后子帧里的最后一个符号可能用作间隔

l 使用正常的CP，对于长距离的D2D通信可考虑使用拓展的CP

l DMRS作为 D2D通信的信号，不由小区决定

l 使用的MCS等级是一个配置的选项

l 可以使用符合流程的现行的速率

l D2D通信中采用同样的交织技术

l D2D通信在一个子帧里面不应有间隙间的跳动。资源的映射应与PUSCH的映射相同

l D2D通信的机制应与PUSCH的掩码机制相同

7) R1-135369 ZTE（noted）

l 在中心调度的情况下，D2D直接的控制信道可以在CH传输的PUCCH/PUSCH中携带，或者通过PUSCH中的MAC或RRC信令

l 在法中心调度或分面式调度的情况下，D2D直接的控制信道能在上行的频谱中传输，通过PUSCH或PUCCH。不过这需要新的DCI设计和资源映射。

图1 D2D的直接控制信道，与PUSCH的CQI传输相似

图2 D2D的直接控制信道，与PUSCH的RI/ACK传输相似

图3 D2D的直接控制信道，SC-FDMA符号的粒度。

图4 D2D的直接控制信道使用PUCCH格式1/1a/1b的结构

图4 D2D的直接控制信道使用PUCCH格式2/2a/2b/3的结构

8) R1-135385 Huawei,HiSilicon（noted）

l 在D2D广播通信的情况下，应有一个共享的ID可选以减小干扰

l 在D2D广播通信的情况下应支持HARQ，以及HARQ应联系于时间资源从而使盲合并的HARQ可用

l 在D2D广播通信的情况中使用A/N反馈以支持链路自适用

图1 一个HARQ到时间资源的映射

9) R1-135386 Huawei,HiSilicon（noted）

l 除了MCS和RA，RV和NDI也应在D2D通信的控制信令中使用。功控相关的信息、MIMO相关的信息作为可选。其他的信息如HARQ相关信息FFS

l 应继续研究如何在D2D链路中传输控制信令。在单载波的条件下，我们倾向于使用控制信令与数据复用的方式

l 现有系统的数据信道结构尽最大可能的重用

l 不能在PUCCH（Cellular）和D2D中同时传输。只能通过TDM的方式传输

l 不需要使用CSI-RS。如果需要SRS则需进一步的研究。其他参考信号也待进一步的研究。RS的开销应被考虑

图2 控制信令与数据复用

Alternatives	PUCCH-like control channel	Control information carried by PUSCH
Reuse legacy mechanisms	PUCCH	UCI on PUSCH
Single carrier character	No	Yes
Cubic Metric (CM)	High	Low
Can be used for current subframe	Yes	No
Capacity	Limited	Larger
Simultaneous transmission capacity requirement	Yes	No
Power control may be needed	Yes	No
Relationship with data	Separate	Multiplexed with data

表1选项总结

图3数据信道结构

10) R1-135613Intel Corporation（noted）

l 在MCL和频谱效率的观点下，窄带分配具有较好的优势

l 多TTI传输可以到达一个135dB的MCL

l 根据信道模型和传输条件4到8次的TTI重传用以支持VOIP业务

l 通过全编码增益以最大化MCL

l 跳频能够获得4-5dB的MCL增益

l 在系统级的观点来看，TTI重传的数量与处理频带内传输问题的灵活性是一个折中的问题。相当于频率复用因子和带内发送之间的关系。

l 在一个同步的系统里面，可能通过在多个连续的TTI内保持同样的发送机集合以避免在宽广的动态的范围内的AGC集合内的频率更新

l 在单个UE的观点下，在蜂窝和D2D传输间使用TDM的方式

l 避免蜂窝和D2D传输的冲突应进一步研究

l D2D链路中的控制信令中的细节应进一步的研究

图1 VOIP业务的MCL分析

11) R1-135805Ericsson（noted）

l 对于D2D广播通信，D2D同步信号序列能作为同步源的身份识别功能。比如在网络覆盖场景下的控制网络节点或无网络覆盖场景下的同步源

l 调度分配通过加扰具有了同步源识别的功能。SA包含了广播传输识别的信息，以及SA的CRC通过预定义的广播RNTI来进行加扰

l 数据包通过加扰具有了广播传输识别的功能

l 数据包的CRC是否需要进行加扰还需进一步的研究

l D2D信标信道中的DMRS应与同步源的识别联系在一起

l D2D信标信道通过加扰具有了广播传输识别的功能

l D2D信标信道的CRC是否需要进行加扰还需进一步的研究

l PD2DSCH通过加扰具有了广播传输识别的功能，负责PD2DSCH的传输。更进一步的，PD2DSCH的CRC通过预定义的广播RNTI来进行加扰

12) R1-135806Ericsson（noted）

l GP在D2D子帧的开始和结束是非常有用的

l 将第一个和最后一个OFDM符号作为保护周期会产生巨大的开销

l 可以减少一半的保护周期的SC-OFDM符号，同时不改变LTE的发身机/接收机块

l 尽管没有引入保护周期，接收机能够最小化SC-OFDM符号间的干扰，可通过调速交织器使得没有系统的比特映射到第一个或最后一个OFDM符号中去

l D2D广播数据的传输与蜂窝传输采用FDM的方式

l 发现以及其他D2D/蜂窝的信道通过TDM的方式分隔在不同的子帧中

l 在网络中使用上行的TA作为数据通信的传输时序，以允许D2D物理数据信道与传统的LTE通过FDM的方式复用

l 携带调度分配信息的物理直接数据信道采用拓展的CP

l 携带数据的物理直接数据信道采用拓展的CP

l 根据表1采用不同的D2D物理信道的CP长度

图1 UE内的冲突

图2 UE间的干扰

图3 第一个和最后一个符号用作保护周期的D2D子帧

Channel/payload type:

Discovery channel

Physical direct data channel carrying a scheduling assignment [4]

Physical direct data channel carrying broadcast data

PD2DSCH

Multiplexing:

FDM in discovery subframes;

TDM with other D2D/cellular channels;

FDM in scheduling assignment subframes;
TDM with other D2D/cellular channels;

FDM with legacy UL channels/signals;

TDM with other D2D/cellular channels [6];

CP length:

FFS: normal/extended CP

FFS normal/extended CP

Normal CP

FFS normal/extended CP

表1 在不同物理信道下的CP和复用方案

D2D Discovery

Discovery Signal and Message Design Details

1) R1-135323 Qualcomm Inc（noted）

l Open ProSe 和Restricted ProSe discovery使用相同的物理层设计。Restricted ProSe的discovery信息应该被适当编码

l Discovery信号的信息部分被用于传输discovery information，序列部分应被用于信道估计。序列部分应为DMRS

l Discovery信息和序列在时域和频域应一同传输

l Discovery子帧应和PUSCH传输使用的上行WAN子帧相似，然而，子帧的最后一个符号可以被用作间隔

l Discovery传输使用普通循环前缀，对于大的discovery范围，可使用扩展的循环前缀

l DMRS信号传输作为discovery信号的一部分不能依赖于小区标识

l Discovery应使用QPSK调制

l Discovery应使用Turbo编码

l Discovery应使用像PUSCH一样的信道交织

l Discovery传输不能出现跨时隙跳频。资源元素映射应与PUSCH资源映射相同，并且没有跨时隙跳频

l Discovery扰码机制应与PUSCH扰码机制相同，例外的是，加扰序列不能依赖UE或者小区标识

l Discovery信息传输应使用长度为24的循环冗余校验码（CRC）

Figure 4

Figure 5

Resource Allocation Method

1) R1-135045 Huawei,HiSilicon（noted）

l Observation 1:在用户密度低的场景下，Type 2比Type 1更加高效

l Observation 2:在用户密度高的场景下，对于一些UE来说，Type 1太慢，Type 2A提供了快速的discovery alternative

l Observation 3:在用户密度高的场景下，Type 2A不会像Type 1一样有范围限制

l Proposal 1:至少支持Type 2A和Type 1这两种解决方式

Figure 1.Number of discovered UEs vs. time

2) R1-135046 Huawei,HiSilicon（noted）

l 使用越多的DRs(Discovery Resources)会越提高discovery性能，但会带来更大的discovery开销和时延

l D2D UE和控制节点应该持续监测和报告分配的discovery资源池的使用情况，以此来确定discovery资源池的大小

l 设计Type 1的资源分配时，应考虑短discovery循环

Figure1. Discovery performance with different combining reception schemes

3) R1-135089 CATT（noted）

l Discovery资源单元应为一对PRB对（PRB-pair），而非两队

l 一个discovery周期中，对于特定的消息，多个discovery资源分配可被考虑为大尺度发现

l 在一个消息有多discovery资源的情况下，资源分配机制的设计需要考虑到接收机中多discovery资源的信息结合

l 当discovery资源有限时，需采用基于UE分组的discovery周期/资源配置

l 随机资源选择可成为今后D2D discovery研究的基线

Figure 1: Discovery performancefor different sizes of discovery resource (Option 1 layout)

Figure 2: Average number ofdiscovered UEs for different discovery resources per UE (Option 1 layout)

Figure 3: Discovery probabilityvs. pathloss for different options

Figure 4: Transmissionprobability based discovery period/resource configuration

Figure 5: UE Groupsbased discovery period/resource configuration

Figure6: Performance comparison of different resource allocation schemes (16subframes)

Figure7: Performance comparison of different resource allocation schemes (64subframes)

Figure 8: Average discovered UEsfor different resource selection schemes (Option 1)

Figure 9: Average discovered UEsfor different resource selection schemes (Option 5, all outdoor UEs)

Figure 10: CDF of discovered UEs for RAM1/2/3 in the first discoveryperiod (Option 1, 16 subframes)

Figure 11: Discovery probability vs. pathloss for RAM1 and RAM4(Option 1, 64 subframes)

Table 1: System level simulation assumptions

Parameter	Assumption
Layout	Hexagonal grid, 3 sectors per site with 19 macro-site with wraparound Option 1: (Urban macro (500 m ISD) + 1 RRH/Indoor Hotzone per cell) Option 5: Urban macro (1732 m ISD) with uniform UE distribution
Channel model	According to TR 36.843 v0.1.0
Carrier frequency	2 GHz
System bandwidth	10 MHz
Network synchronization	All eNodeBs are synchronized
UE antenna configuration	1 TX 2 RX
Transmit power	23dBm, Antenna gain 0 dBi, Noise figure 9 dB
Number of D2D UEs per sector	150 UEs
UE drop for D2D UEs, for discovery	As described in TR 36.843 v0.1.0
Discovery Bandwidth	44RBs
Discovery subframes number in one period	16/32/64
Discovery signal format	1 PRB PUSCH with two slots
Resource allocation	Random allocation within each period as baseline
In-band emission	[W,X,Y,Z] = [3,6,3,3]dB
Multiple access type	SC-FDMA
Modulation type	QPSK
Message size	104bits
UE mobile speed	3km/h

4) R1-135325 Qualcomm Inc（noted）

l 对于RRC_IDLE and RRC_CONNECTED UEs，Type 1的资源分配比Type 2A及Type 2B更加高效

l Type 2需要eNodeB之间的协调，以避免资源碰撞，同时需要伪随机跳频来克服半双工的约束

l 加上eNodeBs的资源协调和管理，Type 1的性能与Type 2相似

l Type 1应被视作D2D discovery的基线

Figure 1: Radio resources consumed for resourceallocation (all UEs start in RRC_IDLE)

Figure 2: Radio resources consumed for resourceallocation (all UEs start in RRC_CONNECTED)

Figure 3 Resource reuse pattern for Type 2discovery

Figure 4 Type 1 vs. Type 2 discoveryresource comparison

5) R1-135372 ZTE（noted）

l 对于时域资源分配，LTE规范化中signaling method reusing subframeallocation schemes有更高的优先级；对于频域资源分配，从系统角度来看，PUSCH和discovery信号同时传输

l 对于小区间discovery，在同步部署下，需研究多小区公共资源分配方案，对于异步部署，UE需能锁定同步信号，并解码来自相邻小区的discovery资源分配指示

l 对于Type 1，周期内随机资源选择可作为今后工作设想

Figure1 An example of resource layout of discovery signals

Figure2 An example of time resource allocation for discovery signal.

Figure3 Resource allocation example for synchronous deployments

Figure4 Resource allocation example for asynchronous deployments

Figure5 CDFs of discovery distance.

Figure6 CDFs of number of discovered UEs per discovering UE.

Table A.1:System level simulation assumptions

Parameter	Assumption
Layout	Option 5: Urban macro (1732 m ISD) with uniform UE distribution Mixed Senrio
Channel model	According to the agreements in chairman notes of #73meeting and email discussion [#73-10/10a]
Carrier frequency	2G MHz
System bandwidth	10 MHz
Network synchronization	All eNodeBs are synchronized
UE antenna configuration	1 TX 2 RX
Transmit power	23dBm, Antenna gain 0 dBi, Noise figure 9 dB
Number of D2D UEs per sector	150 UEs
UE drop for D2D UEs, for discovery	As described in TR 36.843 v0.1.0
Discovery Bandwidth	44RBs
Discovery subframes number in one period	8
Discovery period	320ms
Discovery signal format	1 PRB PUSCH with two slots
Resource allocation	Random allocation within each period as baseline
In-band emission	[W,X,Y,Z] = [3,6,3,3]dB
Multiple access type	SC-FDMA
Modulation type	QPSK
UE mobile speed	3km/h

6) R1-135373 ZTE（noted）

l 考虑到Type 2在特殊场景下的优越性，不应该被排除，需要对其进一步研究

l 对于Type 2，UE-specific discovery资源分配可基于Type 1discovery资源池分配，也可单独分配

7) R1-135522 NTT DOCOMO（noted）

l 需同时支持Type 1和Type 2discovery

l 为Type 2B半静态分配资源域（resource region）

--资源域的大小应根据Type 2B UE的数目调整

--分配给Type 1的资源域应该是FDM/TDM

l 应研究一种预定跳频模式，使Type 2B UEs可以在最短的时间内发现对方，同时实现频率分集增益

l 为了弥补定时误差和满足不同场景下不同的discovery范围，需研究如何配置CP的长度

--在网络的覆盖下，eNB需能够配置和控制CP长度

--CP长度配置的资源应该为TDM

Figure 1 Resource hopping for type 2B discovery

2013全年3GPP RAN1会议关于D2D（Device-to-Device）技术的提案分析

说明

RAN1#72

Deployment Scenario and Evaluation Methodology

RAN1#72b

Scenarios and Evaluation Methodology

Channel Model

RAN1#73

Remaining Details of Scenarios and Evaluation Methodology

Channel Model

D2D Discovery

RAN1#74

D2D Communication

D2D Discovery

RAN1#74b

D2D Communication

Synchronization and Timing

Physical Channel Design

Method of Resource Allocation

D2D Discovery

Discovery Signal Design

Results for GeneralScenario Synchronous Deployment

Method of Resource Allocation

RAN1#75

D2D Communication

Synchronization Design

D2D Data communication Physical Channel Design Details

D2D Discovery

Discovery Signal and Message Design Details

Resource Allocation Method

你可能感兴趣的:(技术文献)