5G-NR Direct Mode Communication

NISTIR 8372
Study of 5G New Radio (NR) Support

for Direct Mode Communications
Fernando J. Cintrón
David W. Griffith
Chunmei Liu
Richard Rouil
Yishen Sun
Jian Wang
Peng Liu
Chen Shen
Aziza Ben Mosbah
Samantha Gamboa
This publication is available free of charge from:

https://doi.org/10.6028/NIST.IR.8372
NISTIR 8372
Study of 5G New Radio (NR) Support
for Direct Mode Communications
Fernando J. Cintrón
David W. Griffith
Chunmei Liu
Richard Rouil
Yishen Sun
Jian Wang
Wireless Networks Division
Communications Technology Laboratory
Peng Liu
Chen Shen
Associate, Wireless Networks Division
Department of Physics, Georgetown University, Washington, D.C.
Aziza Ben Mosbah
Samantha Gamboa
Associate, Wireless Networks Division
Prometheus Computing LLC, Sylva, North Carolina

May 2021
U.S. Department of Commerce

Gina M. Raimondo, Secretary
National Institute of Standards and Technology

James K. Olthoff, Performing the Non-Exclusive Functions and Duties of the Under Secretary of Commerce
for Standards and Technology & Director, National Institute of Standards and Technology
Certain commercial entities, equipment, or materials may be identified in this document in order to describe
an experimental procedure or concept adequately. Such identification is not intended to imply
recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to
imply that the entities, materials, or equipment are necessarily the best available for the purpose.
National Institute of Standards and Technology

Interagency or Internal Report 8372
Natl. Inst. Stand. Technol. Interag. Intern. Rep. 8372, 82 pages (May 2021)

______________________________________________________________________________________________________
Abstract
In public safety communications, direct mode communication is essential to keep first re-
sponders connected, especially when there is no network coverage. Direct mode communi-
cation is supported in Third Generation Partnership Project (3GPP) Long Term Evolution
This publication is available free of charge from: https://doi.org/10.6028/NIST.IR.8372
(LTE) via Sidelink (SL). With 3GPP advancing from LTE to Fifth Generation (5G), in this
study we explore new 3GPP 5G New Radio (NR) SL features and evaluate its support for
direct mode communications.
This study focuses on five major topics that are essential to public safety communica-
tions over 5G New Radio (NR) SL: resource management, physical layer, capacity, Quality
of Service (QoS), and LTE/NR coexistence. First, on resource management, we analyze
the supported mapping between applications and SL resources, control of communication
type (unicast, multicast, and broadcast), together with SL physical resource pools. We
then identify several areas for further research, including resource provision for different
types of applications and management of concurrent applications. Second, on physical
layer, we highlight NR’s new features that are expected to improve performance, including
NR’s flexible numerology, feedback channel, advanced channel coding mechanism, and
channel sensing. Third, on capacity, we use maximal achievable data rate as the measure
and analyze in detail the factors that impact NR SL capacity, including various parameters
and overheads. We then study SL capacity under various NR operational configurations
and device capabilities. Using Band n14 for numerical analyses, we also analyze LTE ca-
pacity for comparison purpose, and further quantify NR capacity improvements over LTE
brought by each of the NR new features. Fourth, on QoS, we notice that compared with
the relatively simplified LTE QoS operations, NR SL has much enhanced QoS support by
inheriting various Uu QoS operations and by introducing new features, such as the new
layer Service Data Application Protocol (SDAP) and the new parameter Range. The NR
SL QoS parameters are also backwards compatible with QoS parameters of LTE Proximity
Services (ProSe). We describe these enhancements in detail, as well as the ongoing 3GPP
standardization efforts in supporting public safety communications. Fifth, on LTE/NR co-
existence, we address the transition from LTE to NR before re-farming the full bands. We
discuss in detail Dynamic Spectrum Sharing (DSS), such as the exchange of configurations
between LTE and NR, then point out the limitations in supporting sharing over SL, and
possible solutions in 3GPP standard efforts. Last but not least, we provide an overview
of ongoing 3GPP study items and work items related to the NR SL, including NR ProSe,
direct discovery, direct communication, and UE-based relays, as well as enhancements to
NR SL power efficiency, reliability, and latency.
While our study shows the potential of a promising performance improvement on direct
mode communications for 5G NR SL over LTE SL, the study also identifies its limitations.
Addressing these limitations to better support first responder needs, plus developing mea-
surement tools to quantify performance, will be our next steps.
i
5G New Radio; D2D; Direct Mode; Sidelink; Public Safety.
ii
Key words
______________________________________________________________________________________________________
______________________________________________________________________________________________________
Table of Contents
Definitions vii
Abbreviations vii
1 Executive Summary 1
2 Introduction 4
3 Resource Management 6
3.1 Introduction 6
3.2 V2X Services Provisioning 6
3.2.1 Authorization 7
3.2.2 Service Configuration 7
3.2.3 Resource Provisioning 8
3.3 Sidelink Resource Allocation 8
3.4 Non-V2X applications 10
3.5 Summary 10
4 Physical Layer 11
4.1 Introduction 11
4.2 Frame Structure and Physical Resources 11
4.2.1 Numerologies, Frames, Subframes, and Slots 11
4.2.2 Physical Resources 12
4.3 Sidelink Channels and Reference Signals 13
4.3.1 Physical Sidelink Channels 14
4.3.2 Physical Reference Signals 18
4.4 MIMO and Beamforming 22
4.5 Summary 24
5 Capacity 24
5.1 Introduction 24
5.2 LTE Capacity 24
5.3 NR Capacity 27
5.3.1 Parameter Settings 27
5.3.2 Overheads 29
5.3.3 Blind Transmissions 34
5.3.4 NR Capacity Results 36
5.4 LTE and NR Capacity Comparison 37
5.5 Summary 38
6 Quality of Service 39
6.1 Introduction 39
6.2 Overview of New Radio Quality of Service Model 40
6.3 SDAP Layer for Quality of Service 42
6.4 Quality of Service over New Radio Sidelink 44
iii
______________________________________________________________________________________________________
6.5 Minimum Communication Range 46

6.6 Quality of Service Support for Public Safety 48
6.7 Summary 49
7 LTE/NR Coexistence 49
7.1 Introduction 49
7.2 DSS Overview 49

7.3 Sidelink Considerations 53
7.4 Summary 54
8 Ongoing Standardization 54
8.1 Introduction 54
8.2 Proximity based Services in 5GS work items 55
8.2.1 Study Item phase 55
8.2.2 Work Item phase 58
8.3 NR sidelink enhancement work item 58
8.4 Enhanced V2X services work items 59
8.5 Summary 60
9 Conclusion 61
References 62
iv
______________________________________________________________________________________________________
List of Tables
Table 1 Frame structure parameters and numerologies. 12
Table 2 S-SSB vs. µ and frequency range.
Choices of Nperiod 21
Table 3 64 QAM MCS Table 28
Table 4 256 QAM MCS Table 28

Table 5 NR Channel Bandwidth Table 29
Table 6 Configurations for Band n14 10 MHz NR SL 29
Table 7 Parameter Settings for NR Capacity 29
Table 8 Physical Sidelink Shared Channel (PSSCH) Demodulation Reference Sig-
nal (DM-RS) Time-domain Location 33
Table 9 Resource Allocation of Overhead (OH) Items 35
Table 10 Resource Allocation with Blind Transmissions 36
Table 11 NR SL Capacity in Mbit/s 37
Table 12 NR QoS Mapping Operations 41
Table 13 SL related standardized 5G QoS Identifier (5QI) to QoS characteristics
mapping 42
Table 14 Standardized PC5 5QI (PQI) to QoS characteristics mapping 46
Table 15 LTE SL and NR SL channels 53
List of Figures
Fig. 1 NR SL resource pool configuration. 9
Fig. 2 Relations of physical resources and frame structure. 14
Fig. 3 A slot format without Physical Sidelink Feedback Channel (PSFCH). Nor-
mal cyclic prefix is assumed. 15
Fig. 4 A slot format with PSFCH. Normal cyclic prefix is assumed. 15
Fig. 5 Allocation of PSSCH DM-RS, 2nd-stage Sidelink Control Information (SCI)
and Sidelink Shared Channel (SL-SCH) modulation symbols. 19
Fig. 6 Frequency locations of Physical Sidelink Control Channel (PSCCH) DM-RS. 19
Fig. 7 Transmission of Sidelink Synchronization Signal (S-SS)/Physical Sidelink
Broadcast Channel (PSBCH) blocks with µ = 1. 21
Fig. 8 A slot format of Sidelink Synchronization Signal Block (S-SSB) with normal
cyclic prefix. 22
Fig. 9 LTE’s SL periodic transmission of PSCCH and PSSCH 25
Fig. 10 LTE SL capacity, without considering overhead from Hybrid Automatic Re-
peat Request (HARQ), achieved for each period length, with a 2 ms long
PSCCH. 26
Fig. 11 LTE SL capacity, considering overhead from HARQ, achieved for each pe-
riod length, considering HARQ blind retransmissions. The PSCCH is 2 ms
long and the PSSCH lasting the remaining period length, respectively. 26
Fig. 12 A slot structure configuration with minimal OH. 31
v
______________________________________________________________________________________________________
Fig. 13 Capacity - NR vs LTE. 38

Fig. 14 QoS architecture 40
Fig. 15 SDAP sublayer structure view 43
Fig. 16 Handling of PC5 QoS Flows based on PC5 QoS Rules 45
Fig. 17 Overview of Minimum Communication Range (MCR) for group communi-
cation. 47
Fig. 18 Xn setup and Node Configuration messages. 50
Fig. 19 Example of protected resources pattern definitions in Uplink (UL). 51
Fig. 20 Xn resource coordination messages. 52
Fig. 21 Example use of the Data Traffic Sharing Resource Indication field. 53
Fig. 22 Protocol stack for direct discovery. 56
vi
______________________________________________________________________________________________________
Definitions
PC5 A reference point where the User Equipment (UE) directly communicates with an-
other UE over the direct channel.
Uu The Radio interface between UTRAN and the User Equipment.
Abbreviations
3GPP Third Generation Partnership Project
4G Fourth Generation
5G Fifth Generation
5GC 5G Core Network
5GS 5G System
5QI 5G QoS Identifier
ACK Acknowledgement
AGC Automatic Gain Control
AMBR Aggregate Maximum Bit Rate
ARFCN Absolute Radio Frequency Channel Number
ARIB Association of Radio Industries and Businesses
ARP Address Resolution Protocol
ARP Allocation and Retention Priority
AS Access Stratum
ATIS Alliance for Telecommunications Industry Solutions
BWP Bandwidth Part
CA Carrier Aggregation
CCSA China Communications Standards Association
CP Cyclic Prefix
vii
______________________________________________________________________________________________________
CQI Channel Quality Indication
CRB Common Resource Block
CRC Cyclic Redundancy Check

CRS Cell Reference Signal
CSI Channel State Information
CSI-RS Channel-State Information Reference Signal
D2D Device-to-Device
DC Direct Current
DL Downlink
DM-RS Demodulation Reference Signal
DRB Data Radio Bearer
DRX Discontinuous Reception
DSS Dynamic Spectrum Sharing
E-UTRA Evolved Universal Terrestrial Radio Access
eMBB enhanced Mobile BroadBand
ETSI European Telecommunications Standards Institute
FDD Frequency Division Multiplexing
FR2 Frequency Range 2
GBR Guaranteed Bit Rate
GFBR Guaranteed Flow Bit Rate
HARQ Hybrid Automatic Repeat Request
IE Information Element
IoT Internet of Things
IPv4 Internet Protocol version 4
viii
______________________________________________________________________________________________________
LCID Logical Channel ID
LCP Logical Channel Prioritization
LDPC Low Density Parity Check

LTE Long Term Evolution
MAC Medium Access Control
MBSFN Multicast-Broadcast Single-Frequency Network
MCPTT Mission Critical Push-to-Talk
MCR Minimum Communication Range
MCS Modulation and Coding Scheme
ME Mobile Equipment
MFBR Maximum Flow Bit Rate
MIMO Multiple-Input and Multiple-Output
mMTC massive Machine Type Communication
mmWave Millimeter Wave
MU-MIMO Multi-User MIMO
NACK Negative Acknowledgement
NAS Non-Access Stratum
NB-IoT NarrowBand-Internet of Things
NG-RAN Next-Generation Radio Access Network
NR New Radio
OFDM Orthogonal Frequency-Division Multiplexing
OH Overhead
PAPR Peak-to-Average-Power Ratio
PCF Policy Control Function
PCI Physical Cell ID
ix
______________________________________________________________________________________________________
PDB Packet Delay Budget
PDCCH Physical Downlink Control Channel
PDCP Packet Data Convergence Protocol

PDSCH Physical Downlink Shared Channel
PDU Protocol Data Unit
PER Packet Error Rate
PFI PC5 QoS Flow Identifier
PHY Physical Layer
PLMN Public Land Mobile Network
PMI Precoder Matrix Indicator
PPPP ProSe Per-Packet Priority
PPPR ProSe Per-Packet Reliability
PQI PC5 5QI
PRACH Physical Random Access Channel
PRB Physical Resource Block
ProSe Proximity Services
PSBCH Physical Sidelink Broadcast Channel
PSCCH Physical Sidelink Control Channel
PSDCH Physical Sidelink Discovery Channel
PSFCH Physical Sidelink Feedback Channel
PSSCH Physical Sidelink Shared Channel
PT-RS Phase-Tracking Reference Signal
PUCCH Physical Uplink Control Channel
PUSCH Physical Uplink Shared Channel
QAM Quadrature Amplitude Modulation
x
______________________________________________________________________________________________________
QCI QoS Class Identifier
QFI QoS Flow Identifier
QoS Quality of Service

QPSK Quadrature Phase Shift Keying
RAN Radio Access Network
RB Resource Block
RE Resource Element
Rel-12 Release 12
Rel-13 Release 13
Rel-14 Release 14
Rel-15 Release 15
Rel-16 Release 16
Rel-17 Release 17
RF Radio Frequency
RI Rank Indication
RLC Radio Link Control
RQA Reflective QoS Attribute
RRC Radio Resource Control
S-SS Sidelink Synchronization Signal
S-SSB Sidelink Synchronization Signal Block
SC-FDMA Single-Carrier FDMA
SCI Sidelink Control Information
SCS Subcarrier Spacing
SDAP Service Data Application Protocol
SDU Service Data Unit
xi
______________________________________________________________________________________________________
SFCI Sidelink Feedback Control Information
SFN System Frame Number
SIB System Information Block

SL Sidelink
SL-BCH Sidelink Broadcast Channel
SL-SCH Sidelink Shared Channel
SPS Semi-Persistent Scheduling
SPSS Sidelink Primary Synchronization Signal
SRB Signaling Radio Bearer
SRS Sounding Reference Signal
SSB Synchronization Signal Block
SSSS Sidelink Secondary Synchronization Signal
SU-MIMO Single-User MIMO
TB Transport Block
TDD Time Division Multiplexing
TRPs Transmitter and Receiver Points
TS Technical Specification
TSDSI Telecommunications Standards Development Society, India
TTA Telecommunications Technology Association
TTC Telecommunications Technology Committee
U2N UE-to-Network
U2U UE-to-UE
UE User Equipment
UL Uplink
UP User Plane
xii
______________________________________________________________________________________________________
URLLC Ultra-Reliable and Low Latency Communication
USIM Universal Subscriber Identity Module
V2P Vehicular-to-Pedestrian
V2X Vehicle-to-Everything
V2XP V2X Policy
VRB Virtual Resource Block
VRU Vulnerable Road User
xiii
______________________________________________________________________________________________________
1. Executive Summary
Direct mode communication for first responders over Fourth Generation (4G) Long Term
Evolution (LTE) was first introduced by the Third Generation Partnership Project (3GPP) in
Release 12 (Rel-12). It is supported by combining the Device-to-Device (D2D) capability,
called Sidelink (SL), which enables devices to communicate directly without packets go-
ing through the core network, with the off-network Mission Critical Push-to-Talk (MCPTT)
voice service. Since then, the cellular technology has evolved to Fifth Generation (5G) New
Radio (NR) to provide higher capacity, lower latency, and support for a higher number of
devices. While NR Sidelink (SL) was introduced in Release 16 (Rel-16), only enhanced
Vehicle-to-Everything (V2X) services have been defined to use it. There is an increasing
interest by public safety organizations, such as FirstNet, to leverage NR SL to enable direct
mode communications for first responders. In this study, we focus on the specifications
available in Rel-16 to understand the current capabilities and limitations of NR SL in the
context of direct mode communications for public safety. This document covers the topics
of resource management, physical layer specifications, capacity enhancements, Quality of
Service (QoS), and coexistence between LTE and NR. We conclude the study by highlight-
ing on-going standardization efforts in Release 17 (Rel-17). In each section, we describe
the features or capabilities of NR SL and their applicability to public safety communica-
tions.
We start by analyzing the specifications to understand how SL resources are allocated
to each application, how to determine the type of communication to use (e.g., unicast,
groupcast, or broadcast), and how to provision application service requirements. We de-
scribe the mechanisms to authorize and provision resources for NR V2X services based on
Rel-16 as it is currently the only service defined to use the SL. Configuration information
can be pre-configured inside the devices or dynamically provided by the network through
User Equipment (UE) policies. A UE may be authorized to use the NR SL when in out-
of-coverage, and resources are provided based on geographical areas. When in-coverage,
authorization is based on the Public Land Mobile Network (PLMN) ID and resources are
provided by the serving cell. Further description of the V2X Policy (V2XP) shows that a
V2X service is mapped to a list of authorized NR frequencies (defined by the geographical
area), destination IDs to establish communication, Quality of Service (QoS) parameters,
and security configuration. We also provide a description of the SL resource allocation,
including the SL resource pool, which defines the time and frequency resources that would
be used for SL communication. Based on the analysis of the current specifications, we
discuss several enhancements needed to support non-V2X applications (e.g., D2D applica-
tions), including mechanisms to provision resources for different types of applications and
managing concurrent applications with potentially conflicting SL resource allocations.
We further investigate how devices access the resources allocated to SL by focusing
on the NR SL Physical Layer (PHY). We first introduce the frame structure and physi-
cal resources, then detail the SL channels and reference signals. Finally, we present the
Multiple-Input and Multiple-Output (MIMO) and beamforming techniques. The NR SL
1
______________________________________________________________________________________________________
PHY will have an impact on the performance of supported applications. More specifi-
cally, the flexible numerology can reduce the Orthogonal Frequency-Division Multiplexing
(OFDM) symbol duration and improve communication latency. A configurable number of
retransmissions and the Physical Sidelink Feedback Channel (PSFCH), a newly introduced
feedback channel at PHY layer, can improve the reliability and spectrum efficiency. In ad-
dition, advanced channel coding techniques are used in NR SL, such as Low Density Parity
Check (LDPC) for the data channel and polar code for the control channel, to reduce de-
coding complexity and improve link reliability. We also describe how the Physical Sidelink
Control Channel (PSCCH) in NR SL is used to enable channel sensing to avoid collisions
and interference. Finally, we highlight the beamforming capability as one of the key tech-
niques to improve the communication range, especially at higher carrier frequencies (>
6 GHz).
Leveraging our understanding of the physical layer, we provide our initial investigation
on one of the major performance metrics, capacity, in terms of the maximum achievable
data rate over SL. The public safety band, Band n14, with 10 MHz bandwidth is selected
for numerical analysis, and the focus is on unicast. We study, in detail, NR SL capacity
and compare it to the LTE SL capacity. Like NR, NR SL offers flexibility in configura-
tions of the numerology, modulation, and number of retransmissions to adapt to different
deployment scenarios and service requirements. For this reason, the study covers various
operational configurations and device capacities, together with the associated overheads.
Capacity formula with Hybrid Automatic Repeat Request (HARQ) blind transmissions
(i.e., no feedback channel) is also derived. The initial quantitative results show that NR
capacity varies significantly under different configurations, and the highest is achieved un-
der numerology 0, 256 Quadrature Amplitude Modulation (QAM) capability, and with one
blind transmission. Numerology 0 gives higher capacity than numerology 1, and the limit
to 64 QAM due to device capability leads to a 1/4 reduction in capacity. In addition, NR
capacity is approximately inverse proportional to the number of transmissions when blind
transmission is configured. Furthermore, detailed analyses on the impact of individual NR
features reveals that the greatest NR capacity improvement over LTE comes from the NR
feature of configurable HARQ transmissions, followed by the higher Modulation and Cod-
ing Scheme (MCS) (256 QAM and 948/1024 code rate), and then the support of spatial
multiplexing. Nevertheless, when NR is configured with a very high number of transmis-
sions - NR allows up to 32 transmissions, including the initial transmission and retrans-
missions, compared to the fixed four transmissions in LTE – NR could have less capacity.
Note that this configuration allows NR to meet the requirement of ultra-high reliability
when needed, thus less capacity together with longer latency is balanced out. This work on
capacity will be extended to multicast and broadcast communications, consider other ca-
pacity measures, such as number of users, and correlate the results with the communication
range to obtain an adequate picture of the coverage performance.
We also study the NR QoS model in general, and then take a closer look at the SL re-
lated enhancements. QoS is expected to play a significant role when capacity is limited and
traffic has to be prioritized. Compared with the relatively simplified QoS operations over
2
______________________________________________________________________________________________________
LTE SL, NR SL’s QoS handling has been improved considerably and closely resembles
the QoS mechanism for the 5G system communication over NR’s Uu interface. To name
a few, the application layer data is mapped to QoS flows, and the QoS flows are mapped
to the Data Radio Bearers (DRBs) by the Service Data Application Protocol (SDAP) layer,
which is a new layer introduced by NR. Applications’ QoS requirements can be conveyed
through QoS parameters and QoS characteristics of the corresponding QoS flow. A new
QoS parameter, Range, may be used to specify the area where the QoS needs to be main-
tained. In addition, the new NR SL QoS parameters are backward compatible with QoS
parameters of LTE Proximity Services (ProSe). With these enhancements as well as the
ongoing 3GPP standardization efforts on public safety applications, we are confident that
direct mode communications for public safety services can be supported by NR from the
QoS perspective.
While NR SL looks promising, NR deployments have just started and LTE will remain
the dominant technology for some time, possibly on the same band. This is why we also
investigate Dynamic Spectrum Sharing (DSS), a set of features that allows LTE and NR
base stations to coordinate the sharing of the same spectrum. This is accomplished by
exchanging configuration information, such as Downlink (DL) and Uplink (UL) control
regions, and data resource allocation. However, since DSS does not specifically address SL
communication, there are some limitations into the current specifications, such as how to
signal SL configuration and resource pool allocations. We discuss possible ways to address
current limitations depending on the type of communication, though further work is needed
to identify the modifications necessary for 3GPP specifications.
Finally, we provide an overview of the study items and work items related to the NR
SL that are currently under analysis and specification in 3GPP for the next release (Release
17). The new features include the definition of NR ProSe, comprising direct discovery,
direct communication, and UE-based relays. It also provides enhancements to the NR SL
power efficiency, reliability, and latency for all services that will use it, including enhanced
NR V2X services and NR ProSe services. It is expected that some of those items will
respond, at least partially, to limitations identified in this study. However, as shown in this
study, additional performance evaluation is needed to understand how to efficiently make
use of NR SL for direct mode communications.
3
______________________________________________________________________________________________________
2. Introduction
The Third Generation Partnership Project (3GPP) is responsible for developing standards
to support the deployment of cellular communications technologies. Multiple releases over
the past decade have brought about the widespread deployment of Long Term Evolution
(LTE) cellular networks. These “4G” systems have enabled their end users to communicate
via telephony, text, video, and data transfers.
A feature of 4G LTE cellular systems is their support for direct communication from
one User Equipment (UE) to another UE without the transmitted data traversing the base
station. Directly-exchanged data and signaling traffic between UEs uses a dedicated set of
time and frequency resources known as the “Sidelink (SL),” which 3GPP first introduced
in Release 12 (Rel-12) of its mobile communication standards. In LTE, Device-to-Device
(D2D) communication over the SL can take place between UEs that are within a base
station’s coverage area; if the UEs are being used by public safety personnel, they can also
use D2D communications while outside of any base station’s coverage area.
There are two basic modes of operation for D2D communication via SL: in the first one,
UEs send signaling messages to the base station that request grants of time and frequency
resources that they will use to exchange data, and the base station is responsible for granting
those resources so that multiple UEs can use the SL to communicate without their trans-
missions’ interfering with each other. This mode only applies to in-coverage cases since
the UEs need to communicate with the base station. In the second mode of operations,
UEs communicate without grants by using advertised or pre-configured SL configuration
settings and by selecting time and frequency resources randomly for signaling and data
transmissions, possibly introducing UE-to-UE interference due to message collisions. This
mode can apply to in-coverage cases but is the only option for out-of-coverage cases.
The original version of ProSe defined three basic functions: Discovery, Synchroniza-
tion, and Communications. There was no provision for the normal Hybrid Automatic
Repeat Request (HARQ) feedback, and the communications function was not integrated
with the discovery function, nor was it coupled to the synchronization function. 3GPP
continued to add features to ProSe in subsequent releases. Release 13 (Rel-13) defined
UE-to-Network (U2N) relays, and Release 14 (Rel-14) extended D2D to encompass many
more use cases. These enhancements allow full support for in-coverage, out-of-coverage,
and partial coverage scenarios, enabling first responders to communicate all the time. Rel-
14 also adds support for Vehicle-to-Everything (V2X) communications, albeit via modes
of communication separate from D2D communication. Among other functions that use
broadcast signaling, Rel-14 added the channel sensing and Semi-Persistent Scheduling
(SPS) features, which are designed to reduce the probability of collisions when UEs are
operating in out-of-coverage mode. Release 15 (Rel-15) added more features, including
Carrier Aggregation (CA), transmission diversity, and support for QAM with a 64-symbol
constellation, which allows 6 bits of information to be sent using a single symbol.
The set of releases up to and including Rel-15 are based on the LTE-A interface. The
limitations of this interface may prevent future cellular systems from meeting the IMT-
4
______________________________________________________________________________________________________
2020 performance requirements. Thus, during the past several years, 3GPP has begun
work on standards for 5G cellular communications that are based on a new radio interface.
These standards include definitions for a new set of radio access technologies collectively
known as 5G New Radio (NR). NR operates in two distinct frequency bands: FR1, which
is from 410 MHz to 7125 MHz, and FR2, which is a set of millimeter wave bands between
24.250 GHz and 52.600 GHz [1, Clause 5.1]. NR is enhancing three key capabilities:
The first one, called enhanced Mobile BroadBand (eMBB), is designed to provide higher
capacity and very high peak data rates, with expected data rates up to 20 Gbit/s in the
downlink and 10 Gbit/s in the uplink, and also improved cell edge data rates. This capability
aims at supporting applications such as 3D videos and ultra high definition video. The
second capability is Ultra-Reliable and Low Latency Communication (URLLC) designed
for critical applications such as self-driving (i.e., autonomous) cars, and targets radio access
network latency below 1 ms, which is not possible in LTE since the subframe duration is
1 ms. The third capability is massive Machine Type Communication (mMTC) that falls into
the smart city concept where there are many sensors providing data, with up to one million
devices per square kilometer.
Public safety agencies are interested in how NR can give them the performance im-
provements that will allow their personnel to remain connected and maintain high levels of
situational awareness, even in dangerous environments in which the commercial commu-
nications infrastructure is degraded or destroyed. The higher bandwidth and lower latency
promised by NR standards have the potential to enable enhanced Mission Critical Push-to-
Talk (MCPTT) services, as well as new Internet of Things (IoT) related systems for public
safety use, such as unmanned vehicles with haptic interfaces, which will enable remote
inspection of dangerous environments, such as fires involving hazardous materials. Such
systems can also enable virtual or augmented reality applications to support emergency
field operations like telemedicine.
3GPP’s work includes developing standards for Vehicle-to-Everything (V2X) commu-
nications, and the new SL standards that are part of V2X will offer opportunities for new
types of public safety applications that will improve interoperability and result in a bet-
ter user experience. These standards introduce channel sensing and UE feedback, which
promise to reduce the impact of collisions and interference. However, this additional com-
plexity can introduce overhead that can decrease capacity [2]. Thus, network operators and
equipment manufacturers must use care in designing their systems in order to optimize per-
formance. There are numerous dimensions in the performance metric space that industry
must consider, such as capacity and quality of service. The purpose of this document is
to examine these major performance metrics and discuss how NR addresses them differ-
ently than LTE does, and to identify some of the design tradeoffs that manufacturers and
operators will face as they build and deploy NR networks for public safety applications.
This document is organized as follows. In Section 3, we discuss resource management
for the SL. We describe the mechanisms to allocate resources for V2X services, and discuss
the SL’s pool of time-frequency resources. In Section 4, we examine the SL PHY, with an
emphasis on the new features that 3GPP has introduced to support NR, such as multiple
5
______________________________________________________________________________________________________
numerologies and support for antenna beamforming. The NR PHY layer will allow in-
creases in channel capacity, and we perform a detailed analysis in Section 5 to quantify
NR’s capacity gain relative to LTE, and we examine the effect of various system param-
eters on capacity, such as the number of retransmissions. In Section 6, we consider the
new mechanisms that 3GPP has introduced in NR to support Quality of Service (QoS), and
we compare these to the legacy mechanisms in LTE. We then visit the issue of spectrum
coexistence between LTE and NR in Section 7; because spectrum management is limited to
the Uplink (UL) and Downlink (DL), we summarize the main differences between LTE SL
and NR SL and introduce some possible means of supporting coexistence in the SL. Next,
we review ongoing standardization work in 3GPP in Section 8, which at the time of this
writing is working on Rel-17, and we focus on study items and work items that affect the
NR SL. Finally, we summarize the work in this study in Section 9.
3. Resource Management
3.1 Introduction
Applications and services using direct communication between devices need to be mapped
to SL configurations and resources in order to operate properly and obtain the desired per-
formance. In this section, we use the specifications for provisioning V2X services in Rel-16
as our baseline since it is the only type of services that is clearly defined to use SL over
the 5G NR air interface. In particular, we focus on the mechanisms in place to allocate re-
sources to a given service, specify QoS parameters, and control the type of communication
to be used by a service. We then introduce the NR SL resource pool that specifies the time
and frequency resources available for the UEs to communicate over NR SL. Finally, we
discuss limitations and possible extensions to support non-V2X applications.
3.2 V2X Services Provisioning

The V2X configuration parameters can be pre-configured in the Mobile Equipment (ME)
(i.e., UE), configured in the Universal Subscriber Identity Module (USIM), provided by
UE policies or a V2X application server, or a combination of those [3]. Priority is given
to dynamic information with the highest precedent going to the configuration parameters
provided by UE policies [4]. The V2X policies are provided by the Policy Control Function
(PCF) as part of the UE policy delivery procedure upon request by a UE connected to
the network. The policies are provided in a V2XP UE policy part inside the message
sent by the PCF. While V2XP defines policies for V2X communication over both the Uu
interface (i.e., with the network) and PC5 interface (i.e., the new D2D interface also known
as SL), we only focus on policies related to the PC5 interface. A UE requests updated V2X
policies when the validity timer associated with previous policies expires or when there is
no information in the UE (e.g., going into a new geographical area or a different network).
6
______________________________________________________________________________________________________
3.2.1 Authorization
Authorization to use V2X services is based on whether a UE is connected to a network or
not. Inside the V2XP Info message, there is a single parameter to authorize out-of-coverage
operations, which is defined by the UE not being served by LTE Evolved Universal Ter-
restrial Radio Access (E-UTRA) and not being served by 5G NR. This parameter specifies
if a UE can perform V2X communication over E-UTRA-PC5 and/or NR-PC5. However,

as we will discuss in Section 3.2.3, the resources that can be used will depend on the geo-
graphical area where the UE is located. When the UE is in network coverage, authorization
to perform V2X is specified per PLMN ID, with specific radio resources provided by the
serving cell when the UE connects to the network.
3.2.2 Service Configuration

The list of V2X services that can be used, along with their configuration is also included in
the V2XP. Each service is uniquely identified by a binary code [5]. There are separate lists
for E-UTRA and NR V2X communication, thus allowing the same service to potentially
be using either technology.
With regards to NR, a service is mapped to the following parameters:
• NR Frequency to use in given geographical area: configuration allows to specify a list

of frequencies that can be used within a list of geographical areas. Each frequency is
coded according to the NR-Absolute Radio Frequency Channel Number (ARFCN)
value defined in [6], and limited to bands n38 and n47 in Rel-16.
• Destination L2 ID for broadcast: specifies the L2 ID to be used by lower layers for

transmitting packets in broadcast communications.
• Destination L2 ID for groupcast: specifies the L2 ID to be used by lower layers for

transmitting packets in groupcast communications.
• Destination L2 ID for unicast initial signaling: specifies the L2 ID to be used by lower

layers for transmitting packets during initial connection. During unicast connection
establishment, the UEs will exchange the L2 IDs to use for data communication.
• PC5 QoS parameters: includes information such as guaranteed and maximum flow
bit rate, and range (see Section 6.4 for more details).
• Unicast security policy: indicates cyphering and integrity protection for both signal-
ing and user plane.
• Default communication mode: specifies the default mode of communication (e.g.,

unicast, broadcast, or groupcast) for this service.
7
______________________________________________________________________________________________________
Not all parameters are configured for all services. For example, if a service only sup-
ports broadcast mode, it will not be included in the mapping rules for groupcast or unicast,
nor will it have an entry in the unicast security policies. Similarly, mapping rules can be
used to map multiple services to the same set of configuration parameters.
Based on that information, when a V2X service starts, its service code will be checked
against the list of services that were (pre-)configured. The UE will use the list of NR
frequencies configured for that service to see if it can operate in any of those frequencies
in its current location. If the UE is out of coverage, this will be done by checking the
pre-configured resources. Otherwise, the UE will check if any of the authorized PLMNs
is operating in the right frequency and make a selection. If a frequency is available, the
service can start and information about destination L2 ID, QoS, and security can be passed
to the lower layers to activate SL communication.
3.2.3 Resource Provisioning

Similar to what was done with LTE SL, there are different mechanisms for providing NR
SL resources to the UEs based on the situation and operator preference. For an out-of-
coverage UE, the resources must be pre-configured in the device. To accommodate the
fact that the same set of resources may not be available everywhere, the resources are
location dependent. The UE policies associate a geographical area defined by a set of
latitude/longitude coordinates to a set of radio resources that can be used. For in-coverage
cases, an operator can use the System Information Block (SIB) 12 to broadcast common
resource pool information to all UEs. This message also includes special resource pools
when the UE experiences radio link failure or handover. Alternatively, an operator can
decide to use dedicated resource pools for each UE and transmit the resource configuration
via Radio Resource Control (RRC) signaling (i.e., in an RRCReconfiguration message).
We note that even though the Information Elements (IEs) defined in TS 38.331 [7]
contain a list of configurations for multiple frequency carriers, Rel-16 indicates that only
one element can be configured in the list. That is, in Rel-16, multiple applications will
share the same frequency band and SL Bandwidth Part (BWP), but they could use different
resource pools.
3.3 Sidelink Resource Allocation

The configuration of the SL BWP and resource pools is established by the RRC layer [7]
and provided to lower layers when activated. In Rel-16, UEs can only have one active SL
BWP at a time in a given frequency band. Just like a regular BWP, a SL BWP is defined by
its frequency, bandwidth, Subcarrier Spacing (SCS), and Cyclic Prefix (CP). In addition,
it also defines parameters common to all the resource pools that are contained within it,
namely the number of symbols and starting symbol used for SL in all slots (except those
with Synchronization Signal Block (SSB)), power control for Physical Sidelink Broadcast
Channel (PSBCH), and optionally the location of the Direct Current (DC) subcarrier.
8
______________________________________________________________________________________________________
The SL BWP has different lists of resource pools for transmissions and receptions, to
allow for a UE to transmit in a pool and receive in another one. For transmissions, there
is one pool for UE selected mode, one for scheduled mode (e.g., when the gNodeB helps
with resource selection), and one for exceptional situations. Resource pools are expected
to be used for only transmission or reception, except when the feedback mechanisms are
activated, in which case a UE would transmit Acknowledgement (ACK) messages in a

reception pool and receive ACK messages in a transmission pool.
As shown in Fig. 1, a resource pool located inside an SL BWP is defined by a set
of sl-Rb-Number contiguous Resource Blocks (RBs) in the frequency domain starting at
RB sl-StartRBsubchannel. The resource pool is further divided into subchannels of size
sl − SubchannelSize, which can take one of several values (i.e., 10, 12, 15, 20, 25, 50, 75,
and 100). The notion of a subchannel is essential for the sensing mechanisms designed by
3GPP. Depending on the value of sl-RB-Number and sl-SubchannelSize, some RBs inside
the resource pool may not be used by the UEs. As such, some consistency is needed in
order to avoid wasting resources.
In the time domain, a resource pool has some available slots configured by various
parameters. To determine which slots belong to the pool, a series of criteria are applied:
1) Slots where SSB will be transmitted cannot be used. The number and locations of
those slots are based on configuration, as detailed in Section 4.3. 2) Slots that are not
allocated for UL (e.g., in the case of Time Division Multiplexing (TDD)) or do not have all
the symbols available (as per SL BWP configuration) are also excluded from the resource
pool. 3) Some slots are reserved such that the number of remaining slots is a multiple
of the sl-TimeResource-r16 bitmap length (also defined as Lbitmap ), that can range from
10 bits to 160 bits. The reserved slots are spread throughout 10240 ∗ 2µ slots, where µ is
the numerology. 4) The bitmap sl-TimeResource-r16 is applied to the remaining slots to
compute the final set of slots that belong to the pool, which are labeled t 0SL i , with i between
0 , and repeats every 10240 ∗ 2µ slots.
0 and Tmax
Fig. 1. NR SL resource pool configuration.
Transmissions within the resource pools are based on one of two modes: in Mode 1, the
9
______________________________________________________________________________________________________
resources are allocated by the gNodeB and it supports dynamic and periodic allocations,
while in Mode 2, UEs autonomously select resources based on sensing. A description of
those two modes can be found in [8, Section 6.3.2] while Section 4 provides details on the
SL channels transmitted within the resource pools.
3.4 Non-V2X applications

From an Access Stratum (AS) point of view, the SL can be used by any type of applica-
tions. Unlike LTE SL, which defined different modes for D2D and V2X, there are only 2
modes of communication depending on whether the resource selection is assisted by the
network or independently done by the UEs. As long as the performance provided by NR
SL meets public safety requirements, it is likely that changes required in the AS will be
minimal. However, though mechanisms for provisioning non-V2X applications will be
very similar to those for V2X services, new specifications will be required. Specifically, a
new addressing scheme to identify applications will be needed.
Another issue with supporting non-V2X applications is the handling of concurrent ap-
plications. It is more likely that non-V2X applications and V2X services will need to
operate on different bands. For example, Direct Mode for MCPTT may be configured to
use a regular cellular band (e.g., n2, n4, n5, n12, n13) or a public safety band (e.g., n14 in
the US), while V2X services are configured with any operating bands defined for V2X (i.e.,
n38 or n47 in Rel-16 [6]). This could be the case if a fire truck acts as a U2N Relay and
is equipped with V2X services for collision avoidance and other safety services. Similarly,
a public safety handheld device could also have Direct Mode and Vehicular-to-Pedestrian
(V2P) services running. Supporting those use cases will require further evaluation in order
to handle device limitations, such as operating on multiple bands simultaneously (or in an
alternating fashion) or prioritizing traffic.
3.5 Summary
In Rel-16, 3GPP has defined mechanisms to enable and configure V2X services over SL.
Resources can be allocated based on the device location and the network it is connected
to. Similar mechanisms are required for non-V2X applications and operators will need
to ensure consistency in the UE policies so the devices can communicate with each other.
Concurrent applications between V2X services and non-V2X applications will need further
study to ensure proper operations, especially due to the current restrictions in the number
of SL BWPs and resource pools that can be configured at a given time. Finally, while NR
provides a lot of flexibility to configure the resource pools, analysis of the SL performance
considering public safety traffic and scenarios is needed to guide the allocation policies.
10
______________________________________________________________________________________________________
4. Physical Layer
4.1 Introduction
Like any wireless communication system, the PHY is an important layer in NR. Various
new PHY features have been introduced in NR to support different emerging applications,
which demand high throughput, high reliability, and low latency, among others. In this sec-
tion, we introduce some critical SL PHY features, channels, reference signals, and MIMO
and beamforming, all of which having an impact on user coverage. In particular, Section 4.2
covers configurable numerology, flexible time-domain frame structure, and frequency do-
main resource structure. Section 4.3 introduces the PHY channels and PHY reference
signals defined for NR SL. Section 4.4 presents the MIMO and beamforming development
in NR standardization. Finally, Section 4.5 summarizes the key PHY features introduced
in NR.
4.2 Frame Structure and Physical Resources

4.2.1 Numerologies, Frames, Subframes, and Slots
In terms of frame structure, NR SL is similar to LTE in that 1) the duration of each frame
and subframe are 10 ms and 1 ms, respectively [9, Section 4.3.1], and 2) the number of
consecutive Resource Elements (REs)/subcarriers per RB, Nsc RB , is 12 [9, Section 4.4.4].
NR SL is advantageous over LTE SL in the reconfigurability in both frequency and time

domains. One contributing factor is the numerology µ, which defines SCS, number of slots
in a subframe, and cyclic prefix options. In NR SL, the value of µ ranges from 0 to 3 [9,
Section 8.2.2], whereas in NR UL and DL, 4 is also supported [9, Section 4.2]. In addition,
the supported µ values vary at different frequency bands.
In the frequency domain, SCS is defined by µ as [9, Section 4.4.4]
∆ f = 2µ · 15 [kHz]. (1)
Two types of cyclic prefix are defined: normal and extended cyclic prefix. Normal
cyclic prefix is supported for all µ, whereas only µ = 2 supports extended cyclic prefix [9,
Table 4.2-1]. When normal cyclic prefix is applied, each slot contains 14 OFDM sym-
bols [9, Table 4.3.2-1]. The number of OFDM symbols is 12 for extended cyclic prefix [9,
Table 4.3.2-2].
subframe,µ subframe,µ
Within a subframe, slots are numbered from 0 to Nslot − 1, where Nslot is
the number of slots per subframe for µ, in the form of
subframe,µ
Nslot = 2µ , (2)
and the duration of a slot is
1
τslot,µ = [ms]. (3)
2µ
11
______________________________________________________________________________________________________
To sum it up, parameters and numerologies for frame structure are defined by µ as in
slot is the number of symbols per slot, N frame,µ is the number of slots per
Table 1, where Nsymb slot
subframe,µ
frame for µ, and Nsymb is the number of symbols per subframe for µ.
Table 1. Frame structure parameters and numerologies.

frame,µ subframe,µ Cyclic slot subframe,µ

µ SCS[kHz] Nslot Nslot τslot,µ [ms] Nsymb Nsymb
Prefix
0 15 10 1 1 Normal 14 14
1 30 20 2 0.5 Normal 14 28
Normal 14 56
2 60 40 4 0.25
Extended 12 48
3 120 80 8 0.125 Normal 14 112
4.2.2 Physical Resources

Here we introduce several terms commonly used in the NR physical resource description.
Resource Grid: The resources for NR SL can be represented by a resource grid. Each
of the UL, DL, and SL channels are configured with its set of resource grids, and for SL,
size,µ RB subcarriers in the frequency domain and N subframe,µ OFDM
it is composed of Ngrid, SL Nsc symb
size,µ
symbols in the time domain, where Ngrid, SL is the size of the resource grid for µ that
indicates the number of RBs in the frequency domain, and is related to carrier bandwidth.
µ start,µ µ
The resource grid starts from nCRB = Ngrid , where nCRB is the index of Common Resource
start,µ
Block (CRB) and Ngrid is configured at the RRC layer [7, Section 6.3.5].
Resource Element: The unit element in a resource grid is RE, which occupies one
subcarrier in the frequency domain and one OFDM symbol in the time domain. REs can be
indexed as (k, l) p,µ , where k is the index in the frequency domain, l is the symbol position
in the time domain, and p is the antenna port. RE with k = 0 coincides with the subcarrier
at Point A, which, for NR SL, is the frequency-location expressed as in ARFCN [9, Section
4.4.4.2], [9, Section 8.2.7].
Resource Block and Common Resource Block: In the frequency domain, the general
concept of an RB contains 12 consecutive subcarriers, as mentioned above. On the other
µ µ
hand, CRB is the indexed RB with its index nCRB incremented from 0. nCRB is related to
the frequency index of RE as follows [9, Section 4.4.4.4]:

µ k
nCRB = RB
. (4)
Nsc
Bandwidth Part and Physical Resource Block: Unlike LTE, 5G NR introduces BWP,
which provides more flexibility in frequency resource configuration. For a given numerol-
ogy µi , a BWP i is a subset of contiguous CRBs. Only one BWP can be configured for NR
12
______________________________________________________________________________________________________
SL [10, Section 16], whereas for each of the NR UL and DL channels, up to four BWPs
can be configured and only one can be active [9, Section 4.4.5].
µ start,µ size,µ
The BWP starts at nCRB = NBWP,i , and it contains NBWP,i RBs. It is bound by the
resource grid as [9, Section 4.4.5]
start,µ start,µ start,µ size,µ

Ngrid, SL ≤ NBWP,i < Ngrid, SL + Ngrid, SL , (5)
and
start,µ start,µ size,µ start,µ size,µ
Ngrid, SL < NBWP,i + NBWP,i ≤ Ngrid, SL + Ngrid, SL . (6)
RBs within a BWP are called Physical Resource Blocks (PRBs) and are numbered from
start,µ
0 to NBWP,i − 1. PRB indexes are related to CRB indexes as [9, Section 4.4.4.4]
µ µ start,µ
nCRB = nPRB + NBWP,i . (7)
In NR, since each device may operate in a different BWP and using a different numerol-
ogy, the common resource block is necessary for a device to locate the frequency resources
within the carrier bandwidth. Fig. 2 summarizes the relations of physical resources and
frame structure as described above, including the resource grid, CRB, RE, BWP, PRB,
slot, and subframe. Their sizes and boundaries of indexes are also specified, in both the
time and frequency domains.
4.3 Sidelink Channels and Reference Signals

The PHY of NR SL carries the physical channels of
• Physical Sidelink Shared Channel (PSSCH),
• Physical Sidelink Control Channel (PSCCH),
• Physical Sidelink Broadcast Channel (PSBCH), and
• Physical Sidelink Feedback Channel (PSFCH),

reference signals of
• Demodulation Reference Signal (DM-RS) for Physical Sidelink Shared Channel (PSSCH),
PSCCH, and PSBCH,
• Phase-Tracking Reference Signal (PT-RS), and
• Channel-State Information Reference Signal (CSI-RS),

and Sidelink Synchronization Signal (S-SS) of
• Sidelink Primary Synchronization Signal (SPSS), and
• Sidelink Secondary Synchronization Signal (SSSS).
13
______________________________________________________________________________________________________
Fig. 2. Relations of physical resources and frame structure.
In addition, a series of consecutive OFDM symbols of Sidelink Primary Synchroniza-

tion Signal (SPSS), Sidelink Secondary Synchronization Signal (SSSS) and PSBCH form
a Sidelink Synchronization Signal (S-SS)/PSBCH block, or S-SSB.
These physical channels, physical signals and the S-SS/PSBCH block are discussed in
this section.
As discussed in Section 4.2, a slot contains 14 OFDM symbols with a normal cyclic
prefix and 12 symbols with an extended cyclic prefix. The physical channels of PSCCH,
PSSCH and/or PSFCH, as well as the reference signals, fit into the same slot. Fig. 3 il-
lustrates a slot format that contains PSCCH and PSSCH, with the PSFCH disabled. When
the PSFCH is requested, three slots that were supposed to be allocated to the PSSCH are
reassigned to the PSFCH, as well as its Automatic Gain Control (AGC) and guard symbols.
An example of such a slot format is shown in Fig. 4. Examples in both Fig. 3 and Fig. 4
assume that a normal cyclic prefix is applied. [8, Section 6.2.1]
4.3.1 Physical Sidelink Channels

As listed in [11, Table 4.3-1] and [11, Table 4.3-2], SL transport channels are mapped to
physical channels as:
14
______________________________________________________________________________________________________
Fig. 3. A slot format without PSFCH. Normal cyclic prefix is assumed.
Fig. 4. A slot format with PSFCH. Normal cyclic prefix is assumed.
• SL-SCH =⇒ PSSCH,
• SL-BCH =⇒ PSBCH,
and SL control information and SL feedback control information are mapped to physical
channels as:
• 1st-stage SCI =⇒ PSCCH,
• 2nd-stage SCI =⇒ PSSCH,
• SFCI =⇒ PSFCH.
PSSCH:
PSSCH carries Sidelink Shared Channel (SL-SCH) Transport Block (TB) from the
transport layer, as well as the 2nd-stage Sidelink Control Information (SCI). At the trans-
port layer, the SL-SCH data and 2nd-stage SCI are attached with Cyclic Redundancy Check
(CRC), channel coded, rate matched and then multiplexed. Channel coding for SL-SCH is
LDPC, whereas the 2nd-stage SCI is Polar coded.
15
______________________________________________________________________________________________________
At the PHY, the block of bits are first scrambled and then modulated. The modulation
for the 2nd-stage SCI bits is performed using Quadrature Phase Shift Keying (QPSK),
(q) (q)
resulting in Mbit, SCI2 /2 modulation symbols, where Mbit, SCI2 is the number of coded bits
for the 2nd-stage SCI, and 2 is the QPSK modulation order. Modulation schemes for SL-
SCH bits include QPSK, 16 QAM, 64 QAM, and 256 QAM, and their modulation orders
Qm are 2, 4, 6, and 8, respectively. Using one of the modulation schemes, the number of
(q) (q)
modulation symbols is Mbit, data /Qm , where Mbit, data is the number of coded bits for the
SL-SCH. [9, Section 8.3.1.2]
(q)
The modulation symbols d (q) (0), . . . , d (q) (Msymb − 1) can be mapped up to v = 2 lay-
ers [9, Section 8.3.1.3]:
x(i) = [x(0) (i) ... x(v−1) (i)]T ,

layer
i = 0, 1, ..., Msymb − 1, (8)
v ∈ {1, 2},
(q)
where Msymb is the combined number of modulation symbols for the 2nd-stage SCI and
SL-SCH. When v = 1 [9, Table 7.3.1.3-1],
x(0) (i) = d (q) (i),

layer (q) (9)
Msymb = Msymb ,
and when v = 2 [9, Table 7.3.1.3-1],
x(0) (i) = d (q) (2i),

x(1) (i) = d (q) (2i + 1), (10)
layer (q)
Msymb = Msymb /2.
The modulation symbols are first precoded, then scaled by the amplitude scaling factor
for power control, and finally mapped to the REs indexed with (k0 , l) p,u , in the correspond-
ing Virtual Resource Blocks (VRBs), following the rules [9, Table 8.3.1.5]:
1. Starting from the lowest k0 and the first PSSCH OFDM symbol that contains Demod-
ulation Reference Signal (DM-RS) for PSSCH, the precoded symbols for the 2nd-
stage SCI are mapped to the REs that are not assigned to DM-RS, Phase-Tracking
Reference Signal (PT-RS), or PSCCH. Mapping of the REs first fills the frequency
domain by increasing k0 , and then moves up in time domain by increasing l;
2. Starting from the lowest k0 and l, the precoded symbols that are not for the 2nd-stage
SCI fills the REs for the PSSCH that are not assigned for the 2nd-stage SCI, DM-RS,
PT-RS, or PSCCH, in the increasing order of first k0 and then l.
16
______________________________________________________________________________________________________
Taking the example in Figs. 3 and 4, mapping of the precoded symbols for the 2nd-stage
SCI starts in the first available RE at the OFDM symbol l = 3 in Fig. 3, and l = 1 in Fig. 4.
On the other hand, in both figures, mapping of the precoded symbols for the SL-SCH starts
at the first available RE for the PSSCH from l = 1. Mapping of 2nd-stage SCI symbols, as
in Fig. 3, is also illustrated in Fig. 5.
The VRBs above are then mapped to the corresponding PRBs.

PSCCH:
The PSCCH carries the 1st-stage SCI, which schedules the time and frequency re-
sources for the corresponding PSSCH, as well as the associated 2nd-stage SCI, DM-RS,
and configurations such as MCS, PSFCH, etc. The format of the 1st-stage SCI is SCI for-
mat 1-A. The UE procedure for PSCCH transmissions is introduced in [10, Section 16.4].
As illustrated in Figs. 3 and 4, PSCCH begins at the second OFDM symbol within a slot,
and it may occupy 2 or 3 consecutive symbols in the time domain. In the frequency do-
main, the number of PRBs allocated for it can be 10, 12, 15, 20, or 25 but cannot exceed
the number of PRBs assigned to a subchannel. Allocation of time and frequency resources
are configured at the RRC [7, Section 6.3.5].
Similar to the processing of the 2nd-stage SCI at the transport layer, the 1st-stage
SCI bits are attached with CRC bits, channel coded with Polar coding, and then rate
matched [11, Section 8.3]. The rate matched bits are then scrambled and modulated with
QPSK at the PHY. The modulation symbols are multiplied with the amplitude scaling fac-
tor for power control, and then mapped to the REs for the PSCCH but not for the associated
DM-RS symbols. This mapping increases in frequency first and then in time [9, Section
8.3.2].
PSFCH:
PSFCH is a new PHY channel introduced by NR SL, and is used to carry Sidelink Feed-
back Control Information (SFCI) (i.e., 1-bit ACK or Negative Acknowledgement (NACK)).
Instead of directly sending the 1-bit information bit, a pseudo random sequence is trans-
mitted. A base pseudo random sequence (a low-Peak-to-Average-Power Ratio (PAPR)
sequence) is constructed based on the length of the sequence to be generated, the group
number, as well as the sequence number within a group [9, Section 5.2.2]. There are a total
of 30 groups. The group ID number u and the sequence ID number v within a group can be
computed based on nID , which is given by the higher-layer parameter, sl-PSFCH-HopID.
If nID is configured by the higher layer, we have group ID
u = nID mod 30, (11)
and sequence ID v = 0; otherwise, we have u = 0 and v = 0.

To multiplex multiple PSFCH transmissions in the same frequency-time resource, on
top of the base sequence, a cyclic shift is applied. The cyclic shift is also generated based
on a pseudo random sequence [9, Section 5.2.1] and is determined by the HARQ-ACK
information and the resource allocation schemes. To be specific, the phase rotation that
results from the cyclic shift is a function of the slot number and the symbol index within
17
______________________________________________________________________________________________________
the slot, along with m0 and mcs values determined by the UE [10, Section 16.3], and other
parameters, which can be computed as [9, Section 6.3.2.2.2]:
2π
+ ncs (ns, f 0 l + l 0 )) mod NSC
µ RB ),
αl = RB ((m0 + mcs + mint
NSC (12)
where
ncs (ns, f 0 , l) = ∑7m=0 2m c(8Nsymb

µ µ
slot n + 8l + m).
s, f (13)
0
Here, mint = 0, l is the OFDM symbol number in the PSFCH transmission, l is the index
of the OFDM symbol in the slot that corresponds to the first OFDM symbol of the PSFCH
in the slot, nus, f is the slot number in the radio frame, mcs is the information bit to be
transmitted, and m0 depends on the cyclic shift pair index in the cyclic shift pairs configured
for the resource pool. Also, c(·) is the pseudo random sequence defined by a length-31
Gold sequence and initialized by cinit . If nID is configured by a higher layer, cinit = nID ;
otherwise cinit = 0. After the sequence is generated, it shall be power scaled to conform to
the transmit power specification and mapped to one OFDM symbol. Note that the resource
elements used for the PSFCH in the first OFDM symbol, including the reference signals,
shall be duplicated in the immediately preceding OFDM symbol.
4.3.2 Physical Reference Signals

DM-RS for PSSCH:
In an OFDM symbol for the PSSCH and its associated DM-RS, the frequency loca-
tions for DM-RS has an index of k = 0, 2, 4, 8, . . ., where k is the index for an RE relative to
subcarrier 0 in CRB 0, and k should fall within the CRBs allocated for the PSSCH transmis-
sion. As discussed above, in the first symbol that contains the PSSCH and its associated
DM-RS (in ascending order), the subcarriers for PSSCH are first filled with modulation
symbols for the 2nd-stage SCI, and then filled by modulation symbols for the SL-SCH.
Fig. 5 illustrates their inter-dependencies in the frequency domain.
Time locations of the OFDM symbols that contain the PSSCH DM-RS are determined
by 1) the total number of OFDM symbols for the PSSCH, its associated PSCCH and AGC,
2) the number of PSCCH symbols, and 3) the number of PSSCH DM-RS symbols, using
the reference from [9, Table 8.4.1.1.2-1]. Figs. 3 and 4 provide two examples of time
location allocations, with a different number of PSCCH, PSSCH and DM-RS symbols.
Another example is detailed in Section 5.3.
Prior to being mapped to physical resources, the PSSCH DM-RS sequence is generated
at the PHY and scaled by amplitude scaling factor for power control.
DM-RS for PSCCH:
Similar to the DM-RS for thePSSCH, the sequence for PSCCH DM-RS is also gener-
ated at the PHY and scaled for power control before being mapped to physical resources.
18
______________________________________________________________________________________________________
Fig. 5. Allocation of PSSCH DM-RS, 2nd-stage SCI and SL-SCH modulation symbols.
The frequency locations of PSCCH DM-RS REs are defined as:
k = nNscRB
+ 4k0 + 1,
k0 = 0, 1, 2, (14)
n = 0, 1, . . . ,
where k is the subcarrier index in reference to subcarrier 0 in CRB 0, and falls within the
CRBs allocated for the PSCCH transmission. Therefore, we can derive k from Eq. 14 as
k = 4m+1, 4(m+1)+1, 4(m+2)+1, . . ., as illustrated in Fig. 6, where 4m+1 corresponds
to the first RE allocated for the PSCCH DM-RS.
Fig. 6. Frequency locations of PSCCH DM-RS.
PT-RS for PSSCH:
19
______________________________________________________________________________________________________
The PT-RS is a newly introduced Reference Signal in 5G NR. The purpose of PT-
RS is to track phase changes and compensate phase noise during transmissions, which is
more prominent for higher carrier frequencies. Therefore, the PT-RS is only supported for
Frequency Range 2 (FR2). If the PT-RS is configured to be present, it should be transmitted
in the resource blocks used for the PSSCH, and the time density and frequency of the PT-
RS can be configured by the upper layer and is configured per resource pool [12, Section
8.2.3]. The sequence generation and the details of how the reference signal is mapped to
physical resources can be found in [9, Section 8.4.1].
Channel-State Information Reference Signal (CSI-RS):
The CSI-RS is used for channel sounding. The receiving device measures the received
CSI-RS, then reports back the Channel State Information (CSI) to the transmitter via the
PSSCH. The CSI-RS is configurable in both the time and frequency domain, and can be
used to provide fine channel state information. Therefore, the CSI-RS can support mobil-
ity and beamforming. For SL, the CSI-RS is configured with a maximum of two antenna
ports, and is transmitted in every resource block within the bandwidth CSI-RS being con-
figured [9, Section 8.4.1.5].
S-SS/PSBCH Block:
A slot that transmits the S-SS/PSBCH block is also known as a Sidelink Synchroniza-
tion Signal Block (S-SSB). It consists of PSBCH, SPSS and SSSS symbols. The period of
the S-SSB transmission is 16 frames [10, Section 16.1], and within each period, the number
S-SSB is configured at the RRC [7, Section 6.3.5]. The range of choices
of S-SSB blocks Nperiod
S-SSB varies with µ and the frequency range. Their relationships are listed in Table 2.
for Nperiod
S-SSB , and time interval, N S-SSB , within a period are also configured at
The time offset, Noffset interval
the RRC. Both are measured in terms of a slot. The relationship between slot index iS-SSBslot ,
S-SSB and N S-SSB are illustrated in Fig. 7 (with µ = 1)
S-SS/PSBCH block index iS-SSB , Noffset interval
and is expressed as [10, Section 16.1], [7, Section 6.3.5]:
iS-SSB
slot
S-SSB
= Noffset S-SSB
+ Ninterval · iS-SSB , (15)
where index 0 of iS-SSB

slot coincides with a first slot of a frame with subframe number
iSFN mod 16 = 0, (16)
and
S-SSB
0 ≤ Noffset ≤ 1279,
S-SSB
0 ≤ Ninterval ≤ 639, (17)
S-SSB
0 ≤ iS-SSB ≤ Nperiod − 1.
The SPSS and SSSS are generated using the M-sequence and Gold sequence, respec-
tively. PHY SL identities are used while generating the sequences.
20
______________________________________________________________________________________________________
Fig. 7. Transmission of S-SS/PSBCH blocks with µ = 1.

S-SSB vs. µ and frequency range. [7, Section 6.3.5]
Table 2. Choices of Nperiod
Frequency S-SSB
µ SCS[kHz] Nperiod
Range
0 15 {1}
FR1 1 30 {1, 2}
2 60 {1, 2, 4}
2 60 {1, 2, 4, 8, 16, 32}
FR2
3 120 {1, 2, 4, 8, 16, 32, 64}
Similar to 1st- and 2nd-stage SCIs, at the transport layer, the Sidelink Broadcast Chan-
nel (SL-BCH) payload is generated, attached with CRC bits, channel coded with Polar
coding, and then rate matched [11, Section 8.1]. At the PHY, the block of bits are scram-
bled and modulated with QPSK before being mapped to the physical resources in an S-SSB
slot [9, Sections 8.3.3, 8.4.3].
The DM-RS for the PSBCH is assigned to all PSBCH symbols. The PSBCH DM-RS,
along with the SPSS and SSSS, are first scaled for power control, and then assigned to the
resources following [9, Table 8.4.3.1-1]. A slot format of an S-SSB is illustrated in Fig. 8,
where a normal cyclic prefix is assumed:
1. For a normal cyclic prefix, as shown in Fig. 8, the number of S-SS/PSBCH OFDM
symbols is 13, whereas it is 11 for an extended cyclic prefix, and the last OFDM
symbol in the slot is left unused as a guard symbol for both cases [9, Section 8.4.3.1].
21
______________________________________________________________________________________________________
Fig. 8. A slot format of S-SSB with normal cyclic prefix.
2. In the frequency domain, 132 subcarriers are assigned for the OFDM symbols, with
k = 0, 1, . . . , 131. Each PSBCH symbol, as well as its associated DM-RS, occu-
pies all the 132 subcarriers. For the SPSS and SSSS symbols, the subcarriers k =
1, 2, . . . , 127, 128 are occupied, and the rest are set to zero [9, Table 8.4.3.1-1].
3. In the time domain, SPSSs are assigned to the 2nd and 3rd OFDM symbols, with
indexes l = 1 and l = 2, respectively, whereas the SSSS follows the SPSS and are
assigned to the OFDM symbols with indexes l = 3 and l = 4. PSBCH symbols
are assigned to the rest of the assigned S-SS/PSBCH time resources, and the guard
symbol follows the last PSBCH symbol.
4. In each PSBCH symbol, the associated DM-RS REs are assigned to the subcarriers
with indexes 4n, where n = 0, 1, . . . , 32 [9, Table 8.4.3.1-1].
In order to determine the location of the S-SSB in the frequency domain, the RX UE is
configured with an absolute frequency via the parameter sl-AbsoluteFrequencySSB in the
RRC configuration [10, Section 16.1], which provides the position of the subcarrier k = 66
in the S-SSB.
4.4 MIMO and Beamforming

Multiple-Input and Multiple-Output (MIMO) technology can facilitate beamforming, spa-
tial multiplexing, or spatial diversity by using multiple transmit or receiver antennas or an
antenna array with a large number of antenna elements, and it supports different transceiver
architectures such as analog, digital, or hybrid. In NR, MIMO standardization focuses on
the following areas: MIMO transmission schemes, CSI measurement and report, reference
signal design, and beam management [13].
Two schemes are defined for DL transmission: (i) closed-loop transmission (scheme 1),
where the DM-RS is transmitted using the same precoding matrix as data transmission; (ii)
22
______________________________________________________________________________________________________
open-loop and semi-open loop transmission (scheme 2), where data and DM-RS may use
different precoding matrices. Note that since NR minimizes the always-on Cell Reference
Signal (CRS) usage, DM-RS will be used for channel estimation and data demodulation
for DL signals. For open-loop transmission, a transmitter can use a fixed precoding matrix
without the CSI knowledge. Therefore, in a closed-loop scenario, the receiver needs to send
back a Precoder Matrix Indicator (PMI), while, in an open-loop transmission, the receiver
may only report Channel Quality Indication (CQI) and Rank Indication (RI). In the DL
transmission, as of Rel-16, to maintain the DM-RS signal’s orthogonality, a maximum of 8
layers can be supported for Single-User MIMO (SU-MIMO), and a maximum of 12 layers
can be supported for Multi-User MIMO (MU-MIMO).
For UL transmissions, a maximum of 4 layers can be supported by NR UL. Compared
to LTE UL, which only supports Single-Carrier FDMA (SC-FDMA) waveform, NR UL
supports both SC-FDMA and OFDM with CP waveforms. It is worth noting that the closed-
loop MIMO transmission only supports the OFDM waveform. In addition to the proposed
codebook-based transmission in LTE, the non-codebook approach can also be used in UL
transmissions. For example, in a TDD system, a UE can design its precoding matrix based
on channel reciprocity.
For SL, as of Rel-16, a maximum of two layers is supported in the NR SL and its MIMO
scheme is limited to 2×2. More layers are expected to be supported in future releases [2].
In terms of the CSI reporting and feedback, NR MIMO can support two types of CSI
feedback in the Frequency Division Multiplexing (FDD) system. The type I CSI feedback,
which is also supported in LTE, allows the receiver to feedback a PMI matrix to the trans-
mitter by identifying a preferred precoding matrix in a standardized codebook. Beyond the
type I codebook, NR also supports the type II CSI feedback, which has a higher resolution
than the type I codebook. Type II feedback has the potential to improve the average user
throughput significantly (at least 30 %) over LTE Rel.14, but at the expense of high feed-
back overhead [14]. The goal of type II feedback is to improve MU-MIMO performance
with an increased number of transceivers.
NR expands its supported carrier frequency to cover a much higher frequency spectrum,
such as Millimeter Wave (mmWave). In order to compensate for the large propagation
loss occurring during the high-frequency transmission, antennas with a large number of
elements are often used to beamform the signal and provide additional directional gain. On
the other hand, thanks to the high frequency, antenna size can be significantly reduced, and
it is feasible to pack more antennas in a small form factor.
Based on the limited study conducted by 3GPP RAN1, beamforming is beneficial to
SL communication, especially in FR2 [15]. When a UE is equipped with one or multiple
antenna arrays in direct communication mode, a simple phase shift or a complex gain value
can be applied to each antenna element to form a beam or multiple beams. A larger antenna
size will lead to higher directional gain, but the coverage angle will be reduced. Therefore,
to provide a good coverage for the area, multiple beams may be required. In order to main-
tain the communication, the beam needs to be monitored, and beam failure and recovery
need to be addressed in a timely manner. The beam management procedure, a set of L1
23
______________________________________________________________________________________________________
and L2 procedures to acquire and maintain beams between Transmitter and Receiver Points
(TRPs) and UEs for DL and UL communications is defined in [16]. The beam manage-
ment procedure covers beam determination, beam measurement, beam reporting, as well
as beam sweeping. Beamforming and MIMO study continues to be carried out in Rel-17,
which includes enhancements to multi-beam operation, distributed multiple antennas at the
transmitter and receiver, the Sounding Reference Signal (SRS), and the CSI measurement
and reporting [17].
4.5 Summary
In summary, NR SL introduces numerous new PHY layer features, including configurable
numerology to allow resource usage to be more flexible in both time and frequency domain.
NR also supports new coding schemes, i.e., LDPC for data channels that reduces decoding
complexity and polar code for control channels that efficiently handles small-size block
transmission. In terms of physical channels and physical reference signaling, NR adds a
new physical layer feedback channel, i.e., PSFCH, which can reduce blind transmissions
and improve spectrum efficiency. With respect to power efficiency, NR introduces BWP
such that devices do not have to monitor the whole carrier bandwidth for the control sig-
nals. MIMO and beamforming are important areas for NR, but more studies are needed to
determine under which conditions they could be used for SL communication..
5. Capacity
5.1 Introduction
One major performance metric for SL communications is capacity, which in this context
refers to the maximum achievable data rate. Compared with LTE SL, NR SL supports mul-
tiple features that are expected to improve capacity significantly. Some major ones are up
to two-layer spatial multiplexing, higher modulation and coding schemes, and configurable
HARQ transmissions. In addition, NR also supports a flexible OFDM numerology to meet
different types of service requirements.
In this section we provide our initial investigation on NR SL capacity and its com-
parison with LTE. We first look at LTE SL capacity, then study NR SL capacity and its
improvement brought by different features. The public safety band, Band 14 / Band n14
with 10 MHz bandwidth, is selected for numerical analysis. While NR SL supports unicast,
multicast, and broadcast, in this analysis we focus on unicast and will extend the work to
multicast and broadcast in the future when the related standards are relatively mature.
5.2 LTE Capacity

Communication over LTE SL is periodic in the time domain. An SL period can be config-
ured to last from 40 ms to 320 ms, with a set of predefined values by the RRC layer [18,
Section 6.3.8]. The PSCCH is used to indicate how, where, and when transmissions will
24
______________________________________________________________________________________________________
occur in the PSSCH. The PSCCH and PSSCH are multiplexed in the time domain, with the
PSCCH always preceding the PSSCH, as depicted in Fig. 9.
Fig. 9. LTE’s SL periodic transmission of PSCCH and PSSCH [19].
Both, the PSCCH and PSSCH, have a set of resources configurable in the time and
frequency domain. Within the SL period, the PSCCH can be configured to last from 2 ms
and up to 40 ms, leaving the remaining time available for the PSSCH.
To compute the maximum SL capacity, without considering overhead from retransmis-
sion protocols (e.g., HARQ), the shortest PSCCH (2 ms) is selected. This maximizes the
time available for the PSSCH. For example, in a 40 ms SL period, 2 ms would be reserved
for PSCCH and the remaining 38 ms would be for the PSSCH. Assuming a 10 MHz band-
width is available, 50 PRBs, and no time restriction (i.e., all subframes1 are made available
for SL transmissions) within the PSSCH, we compute the SL capacities for all the possible
SL period lengths. The highest MCS supported by SL, MCS 20, is used in the analysis.
The capacities at the physical layer are presented in Fig. 10. A clear trend is observed,
with longer SL period lengths there is less overhead incurred by fewer transmissions of
PSCCHs, hence the higher capacities. It is worth noting that, even when the difference in
Mbit/s is visually notorious given the scale used, the relative difference is not that signifi-
cant across all SL periods analyzed. A delta of 0.94 Mbit/s is observed from the shortest to
the longest SL period, which provides a capacity increase that is less than 5 %.
LTE SL uses HARQ with no feedback. Four blind HARQ processes are employed for
every TB transmission. The transmission of a TB would only occur if there are enough time
resources for all the HARQ processes associated with the TB to be transmitted, i.e., four
transmission opportunities per TB. After considering the overhead created by HARQ blind
retransmissions in SL, we can compute the actual capacity for SL, presented in Fig. 11.
A delta of 0.47 Mbit/s is observed from the shortest to the longest SL period, which is
1 In LTE, a subframe is 1 ms long.
25
______________________________________________________________________________________________________
Fig. 10. LTE SL capacity, without considering overhead from HARQ, achieved for each period
length, with a 2 ms long PSCCH.
Fig. 11. LTE SL capacity, considering overhead from HARQ, achieved for each period length,
considering HARQ blind retransmissions. The PSCCH is 2 ms long and the PSSCH lasting the
remaining period length, respectively.
26
______________________________________________________________________________________________________
just under a 10 % capacity increase. The capacities are also affected by the length of the
PSSCH not being divisible by a factor of four (the number of blind HARQ transmissions),
which leaves unused transmission opportunities. With our configuration assumptions, only
the 70 ms period with a 68 ms long PSSCH leaves no unused transmission opportunities,
making this configuration stand out.
5.3 NR Capacity
Different from LTE SL, NR SL introduces a number of features that improve the capacity
significantly, such as spatial multiplexing, higher modulation and coding schemes, and a
configurable HARQ. NR SL also supports different numerologies, similar to the NR Uu
interface. With these factors considered and together with the total available resources, NR
SL capacity can be calculated as shown below:
BW,µ
−6 N × 12
data rate (in Mbit/s) = 10 · νlayers · Qm · f · Rmax · PRB µ · (1 − OH), (18)
Ts
where νlayers is the spatial multiplexing number of layers, Qm is the modulation order, f
BW,µ
is the scaling factor, Rmax is the coding rate, NPRB is the number of PRBs for a specific
bandwidth BW and numerology µ, 12 is the number of sub-carriers per physical resource
µ
block as specified by 3GPP, Ts is the symbol duration time in seconds for numerology µ,
OH is the overhead ratio, and the number of transmissions is one.
While Eq. (18) seems the same as the data rate formula in 3GPP Technical Specification
(TS) 38.306 [20, Section 4.1], note that here OH is not necessarily independent of the
bandwidth or numerology, which will manifest in the later overhead analysis. Also, while
Eq. (18) is used as the base of the later analysis and addresses the configuration with the
number of transmissions set to one, it will be extended to include the configuration with
blind retransmissions. In the following we will examine each parameter first, then analyze
overheads and HARQ configurations, followed by the resulting capacity numbers.
5.3.1 Parameter Settings

To evaluate capacity, the value for each parameter in Eq. (18) is selected to maximize the
resulting data rate, as described below.
νlayers : for NR SL unicast transmissions, the maximum number of layers supported for
spatial multiplexing is two [9, Section 8.3.1.3]. Hence we choose νlayers = 2.
Qm and Rmax : NR SL reuses two of the NR downlink MCS tables for its transmis-
sions [12, Section 8.3.1.3], namely 64 QAM table and 256 QAM table, respectively. The
64 QAM table is the default MCS table, while the 256 QAM table may be selected via
sl-MCS-Table from the higher layer based on factors, such as channel conditions and UE
capabilities. Table 3 and Table 4 show part of the two tables in [12]. It can be seen that the
highest MCS has a modulation order of six or eight per MCS table selected, with coding
rate 948/1024. This gives Qm = 6 or 8 and Rmax = 948/1024.
27
______________________________________________________________________________________________________
Table 3. 64 QAM MCS Table [9, Table 5.1.3.1-1].
MCS Index Modulation Order Target code Rate Spectral

IMCS Qm R x [1024] efficiency
0 2 120 0.2344
1 2 157 0.3066
... ... ... ...

27 6 910 5.3320
28 6 948 5.5547
29 2 reserved
... ... ...
Table 4. 256 QAM MCS Table [9, Table 5.1.3.1-2].
MCS Index Modulation Order Target code Rate Spectral

IMCS Qm R x [1024] efficiency
0 2 120 0.2344
1 2 193 0.3770
... ... ... ...
26 8 916.5 7.1602
27 8 948 7.4063
28 2 reserved
... ... ...
f : the scaling factor was introduced mainly to capture capability mismatch between
the baseband capabilities and Radio Frequency (RF) capabilities, and/or hardware resource
sharing between LTE and NR [21]. Here we consider Band n14 NR. There is no other
component carrier and no sharing between NR and LTE. Hence the scaling factor f = 1.
BW,µ µ
NPRB and Ts : Table 5.3.5-1 in [6], partly shown in Table 5, provides channel band-
widths for each NR band and corresponding SCS. Per this table, for the considered Band
n14 10 MHz, two numerology configurations are supported, 0 and 1, as listed in Table 6.
This leads to normal CP [9, Table 8.2.2-1], and the maximum number of 52 PRBs and 24
PRBs, respectively [6, Table 5.3.2-1]. The corresponding symbol duration, in seconds, can
be calculated from the formula:
Ts = 10−3 /(14 × 2µ );
µ
(19)
where µ is the numerology, 10−3 is the time in seconds for one subframe, 2µ is the number
of slots per subframe, and 14 is the number of symbols per slot.
The resulting values per numerology are also listed in Table 6. Note that a different nu-
merology leads to a different number of PRBs and symbol duration, as well as their ratio
BW,µ
NPRB
Ts
µ , which directly affects the resulting capacity (Eq. (18)).
It is worth mentioning that, as discussed in Section 3.3, resources are allocated to users
28
______________________________________________________________________________________________________
Table 5. NR Channel Bandwidth Table [9, Table 5.3.5-1]
NR band / SCS / UE Channel bandwidth

NR SCS 5 10 15 20 25 30 40 50 60 70 80 90 100
Band kHz MHz MHz MHz MHz MHz MHz MHz MHz MHz MHz MHz MHz MHz
...
15 Yes Yes
n14 30 Yes
60
...
15 Yes Yes10 Yes Yes10 Yes Yes10 Yes10
n38 30 Yes10 Yes Yes10 Yes Yes10 Yes10
60 Yes10 Yes Yes10 Yes Yes10 Yes10
...
15 Yes10 Yes10 Yes10 Yes10
n47 30 Yes10 Yes10 Yes10 Yes10
60 Yes10 Yes10 Yes10 Yes10
...
...
NOTE 10: These UE channel bandwidths are applicable to sidelink operation.
per subchannel, and there is a discrete set for subchannel sizes. This could lead to left-
over RBs inside the resource pool that cannot be used. In our study, however, we use the
maximum number of PRBs within the bandwidth to simplify the context.
Table 6. Configurations for Band n14 10 MHz NR SL

BW,µ
Sub-carrier Number Symbol Duration NPRB
Numerology µ
Spacing of PRBs in microseconds Ts
µ =0 15 kHz 52 71.43 0.73
µ =1 30 kHz 24 35.71 0.67
As a recap, Table 7 summarizes the values of each parameter discussed above. The
numbers will be used in the NR capacity computation later.
Table 7. Parameter Settings for NR Capacity
Number Symbol Duration

µ
Parameter νlayers Qm (MCS Table) f Rmax numerology of PRBs Ts
BW,µ
NPRB in microseconds
6 (64 QAM) µ =0 52 71.43
Value 2 1 948 / 1024
8 (256 QAM) µ =1 24 35.71
5.3.2 Overheads
In addition to data packets, the NR SL also carries other necessary channels and signals for
operations. These channels and signals count towards transmission Overhead (OH), and is
29
______________________________________________________________________________________________________
reflected in Eq. (18) as overhead ratio OH. The OH in NR SL is composed of the following
items:
• PSFCH,
• CSI-RS,
• PT-RS,
• Reserved slots,
• PSCCH,
• 2nd-stage SCI on PSSCH,
• DM-RS for PSSCH,
• AGC symbol,
• Guard symbol, and
• S-SSB.
The ratio OH is determined by the number of REs occupied by these OH items, denoted
µ µ
by REOH,slot , over the total number of REs allocated for NR SL, denoted by REslot . That is:
µ
REOH,slot
OH = µ . (20)
REslot
Thus, to achieve maximum capacity in Eq. (18), the resource allocation for NR SL shall be
maximized at the target band, and OH shall be minimized.
Regarding the total number of REs allocated to the SL, as shown in Table 6, the maxi-
mum number of PRBs at Band n14, with a 10 MHz bandwidth, is 52 for µ = 0 and 24 for
µ = 1. This gives the maximum resource allocation of 8736 REs per slot for µ = 0 and
4032 REs per slot for µ = 1.
Regarding resources occupied by the OH items, OH resource allocation can be opti-
mized in a context where each of the OH items occupies its minimum allowable resource.
Towards this goal, we examine each OH item as shown below.
PSFCH, CSI-RS, PT-RS and Reserved Slots: in NR SL, PSFCH and CSI-RS can be
disabled in the RRC, by configuring SL-PSFCH-Config-r16 to sl0 and sl-X-Overhead to n0,
respectively (6.3.5 in [7]). In addition, as the transmission of the PT-RS is only supported
in FR2 [12, Section 8.2.3], for Band n14 under study, it is also excluded from OH. That is,
the minimum resources occupied by these three items is zero. Fig. 12 illustrates a resulting
slot structure with zero occupancy of these three items.
Reserved slots are discussed in Section 3.3. Its period is related to a series of parame-
ters: numerology, S-SSB repetition, number of non-sidelink slots, and sl-TimeResource-r16
30
______________________________________________________________________________________________________
bitmap length [12, Section 8]. By configuring these parameters, the period of the reserved
slots can be high enough that the resulting overhead is negligible. Therefore, its impact is
excluded from our study.
Here we disable the PSFCH in order to minimize OH and maximize data rate. This
maps to no feedback-based HARQ. In practice when feedback-based HARQ is requested,
the PSFCH is enabled. Accordingly, in the slots that the HARQ requests are placed in, three
OFDM symbols that were supposed to be allocated for the PSSCH are instead allocated
for the PSFCH, its AGC symbol before it, and the guard symbol that follows. A similar
configuration is illustrated in Fig. 4 in Section 4.3. By comparing Fig. 4 with Fig. 12, we
can see a significant increase of OH resource allocation, which would lead to a higher OH
ratio and eventually a much lower throughput. We concentrate on capacity in this study
and will address this case in our future work.
Fig. 12. A slot structure configuration with minimal OH.
PSCCH: The PSCCH carries the 1st-stage SCI. Its resources are allocated at the RRC
[7, Section 6.3.5], with related configurations listed as:
• sl-TimeResourcePSCCH-r16: n2, n3,
• sl-FreqResourcePSCCH-r16: n10, n12, n15, n20, n25.

It shows that by configuring sl-TimeResourcePSCCH-r16 and sl-FreqResourcePSCCH-r16
in SL-PSCCH-Config-r16, two or three OFDM symbols can be allocated to the PSCCH in
the time domain, and 10, 12, 15, 20, or 25 RBs can be allocated in the frequency domain,
respectively. Therefore, a minimum of 240 REs can be allocated to the PSCCH.
2nd-Stage SCI: The 2nd-stage SCI is carried on the PSSCH. Its number of bits is deter-
mined by the 2nd-stage SCI formats, which are: 35 bits for 2-A, and 48 bits for 2-B [11,
Section 8.4.1.1]. These SCI bits are then attached with 24 CRC bits [11, Section 7.3.2].
After performing channel coding and rate matching on this concatenated bit sequence, the
coded bits are modulated with QPSK and multiplexed with the coded bits for the SL-SCH
[11, Sections 8.4.3, 8.4.4, 8.4.5, and 8.2]. These procedures jointly determine the number
of REs, which is expressed in Eq. (21) [11, Section 8.4.4]:
31
______________________________________________________________________________________________________
(& ' )
SCI2
(OSCI2 + LSCI2 ) · βoffset l PSSCH −1
Nsymbol
m
0 SCI2
QSCI2 = min SCI2
, α ∑l=0 MSC (l) + γ, (21)
Qm · R
where the first parameter in min() is the number of REs required for the 2nd-stage SCI after
QPSK, and the second parameter is the total number of available REs in the PSSCH, which
is the greater one in our context; OSCI2 = 35 is the number of bits for the 2nd-stage SCI
format 2-A; LSCI2 = 24 is the number of CRC bits; QSCI2 m = 2 is the modulation order of
QPSK for the 2nd-stage SCI; R is the target coding rate for the PSSCH, which is 948/1024
SCI2 is 1.125, which are
from the previous analysis; the minimum value of the coefficient βoffset
jointly determined by the beta offset indicator from the 1st-stage SCI [11, Section 8.3.1.1]),
sl-BetaOffsets2ndSCI-r16 in the RRC [7, Section 6.3.5], and in [10, Table 9.3-2]; and γ is
the number of vacant REs in the last RB of the 2nd-stage SCI [11, Section 8.4.4].
By plugging all these numbers into Eq. (21), the minimum number of REs required for
the 2nd-stage SCI is 36.
DM-RS: The resource occupied by the PSSCH DM-RS can be viewed from the time
domain and the frequency domain. In the time domain, resource allocation for the PSSCH
DM-RS consists of the number of OFDM symbols occupied and their positions in a slot.
The number of OFDM symbols is associated with the time pattern value, which is config-
ured by sl-PSSCH-DMRS-TimePatternList-r16 in the RRC [7, Section 6.3.5] as:
• sl-PSSCH-DMRS-TimePatternList-r16: SEQUENCE (SIZE (1..3)) OF

INTEGER 2..4.
It is then interpreted as the number of DM-RS symbols at the PHY. It also indicates that
the minimum number of PSSCH DM-RS symbols that can be configured is two.
Following the number of OFDM symbols occupied in a slot, Table 8 indicates the po-
sitions of the DM-RS symbols in a slot. The positions are jointly determined by 1) the
number of PSCCH symbols, which is configured to 2 in our context, and 2) the number
of OFDM symbols for the PSCCH and PSSCH, ld , including the AGC symbol [9, Sec-
tion 8.4.1.1.2]. As the maximum value of ld is 13, the two DM-RS symbols are positioned
at Symbols 3 and 10, which are illustrated in Fig. 12.
Meanwhile, in the frequency domain, resource allocation for the PSSCH DM-RS fol-
lows Configuration Type 1 (8.4.1.1.2 in [9]), in which each two adjacent REs for the PSSCH
DM-RS are spaced by an RE for PSSCH data or the 2nd-stage SCI. Mathematically, this
relation is expressed as in [9, Section 6.4.1.1.3]:
( p̃ )
ãk,l j,µ = w f (k0 )wt (l 0 )r(2n + k0 ),
k = 4n + 2k0 + ∆, (22)
k0 = 0, 1,
n = 0, 1, ...,
32
______________________________________________________________________________________________________
Table 8. PSSCH DM-RS Time-domain Location [9, Table 8.4.1.1.2-1].
DM-RS position l¯
ld in PSCCH duration 2 symbols PSCCH duration 3 symbols
symbols Number of PSSCH DM-RS Number of PSSCH DM-RS
2 3 4 2 3 4
6 1, 5 − − 1, 5 − −
7 1, 5 − − 1, 5 − −
8 1, 5 − − 1, 5 − −
9 3, 8 1, 4, 7 − 4, 8 1, 4, 7 −
10 3, 8 1, 4, 7 − 4, 8 1, 4, 7 −
11 3, 10 1, 5, 9 1, 4, 7, 10 4, 10 1, 5, 9 1, 4, 7, 10
12 3, 10 1, 5, 9 1, 4, 7, 10 4, 10 1, 5, 9 1, 4, 7, 10
13 3, 10 1, 6, 11 1, 4, 7, 10 4, 10 1, 6, 11 1, 4, 7, 10
where k is the RE index. Since ∆ = 0 (in [9, Table 6.4.1.1.3-1]), then k = 0, 2, 4, 8, ..., which
are for all even RE indexes. Therefore, the number of REs per RB for a PSSCH DM-RS is
six in each DM-RS symbol.
By combining the time domain of two DM-RS symbols per slot and the frequency
domain of six REs per RB in each DM-RS symbol, and noting the total number of RBs
for different numerologies in Table 6, the total number of REs for the PSSCH DM-RS is
624 for µ = 0 and 288 for µ = 1. Note that the resulting number of REs occupied by the
DM-RS is a function of channel bandwidth as well as the numerology employed. That is,
BW,µ
the OH in Eq. (18) is not independent of channel bandwidth and numerology, or NPRB and
µ
Ts .
AGC and Guard Symbols: NR SL employs AGCs and guard symbols. The first symbol
in each slot is the duplication of the first multiplexed symbol of the PSCCH and PSSCH
[9, Sections 8.3.1.5 and 8.3.2.3]. This duplicated symbol serves as an AGC symbol. Mean-
while, in case the PSFCH is enabled, its first mapped OFDM symbol, including the DM-
RS, PT-RS, or CSI-RS, is duplicated in the immediate preceding symbol. This duplicated
symbol serves as the AGC symbol for the PSFCH [9, Section 8.3.4.2.2]. Additionally, a
guard symbol follows the last symbol of the PSSCH, PSFCH, or S-SSB [9, Section 8.2.1].
As mentioned previously, in this study the PSFCH is not enabled. Consequently, one AGC
and one guard symbol are allocated in each slot, as illustrated in Fig. 12, and the number
of REs for each of them are 624 when µ = 0 and 288 when µ = 1.
S-SSB: The S-SSB carries the PSBCH and Sidelink Primary/Secondary Synchroniza-
tion Signals. The transmission period of the S-SSB is 16 frames, and the number of S-SSBs
within a period can be configured by sl-NumSSB-WithinPeriod-r16 with a minimum value
of 1 [7, Section 6.3.5]. That is, the S-SSB can be transmitted at a minimum rate of once
per 16 frames. Mapping this rate to slots under different numerologies leads to once every
160 slots for µ = 0 and once every 320 slots for µ = 1, respectively [9, Section 4.3.2].
As the set of slots for the resource pool, i.e., for transmission or reception of the PSSCH,
33
______________________________________________________________________________________________________
does not include the slots for the S-SSB [12, Section 8], the un-allocated resources in the
slots that contain S-SSBs are not intended for the PSSCH. Therefore, the overhead due to
S-SSBs comes mainly from its occupancy in the time domain. In the time domain, each
S-SSB occupies 13 OFDM symbols (under normal cyclic prefix) plus one guard symbol [9,
Section 8.4.3.1].
By combining the S-SSB rate and the size of each slot that contains an S-SSB, the OH
due to S-SSBs can be averaged to the slot level as
µ=0 (13 + 1) × 52 × 12
REOH,S-SSB = = 54.6 REs/slot (23)
160
for µ = 0, and
µ=1 (13 + 1) × 24 × 12
REOH,S-SSB = = 12.6 REs/slot (24)
320
for µ = 1.
In summary, Table 9 lists the number of REs each OH item occupied per slot, in
Columns 2 and 3. The total number of REs occupied by all items is calculated in the last
row. In addition, the percentage of each OH item over the total OH is listed in Columns
3 and 4. It shows that OH mainly comes from three items—DM-RS, AGC, and guard
symbols—with each contributing almost 30 % and in total around 85 % of the overall OH.
Following the three is the PSCCH, which contributes around 10 %. The contribution of
other items, i.e., 2nd-stage SCI (labeled as SCI2 on the table) and S-SSB, is minimal.
The last two columns of Table 9 shows the overhead over total number of REs, both
from each OH item and their sum, which maps to the OH term in Eq. (18). It shows that
the OH occupies around 1/4 of the overall resources. Note that while numerology 1 uses
a larger SCS and supports more slots, it also introduces ∼ 3 % higher OH compared with
numerology 0. This is because a larger SCS leads to a lower number of total REs per
slot. For the OH item DM-RSs, AGCs, guard symbols and S-SSBs do not contribute to
the change of the overhead ratio since they scale with the total number of REs. However,
for the PSCCH and 2nd-stage SCI, their sizes stay the same, which results in an increased
overhead ratio when the total number of REs is lower. Later these total OH ratios will be
substituted into Eq. (18) for a capacity calculation.
5.3.3 Blind Transmissions

As mentioned previously, one major difference between NR SL and LTE SL is that NR
supports a configurable HARQ while LTE uses blind transmissions only. The configurable
HARQ in NR allows for efficient use of bandwidth by eliminating unnecessary retransmis-
sions. It works especially well in capacity under study, as capacity is achieved when the
channel quality is good and no retransmissions are necessary. Nevertheless, to adapt to
different scenarios, NR also supports the configuration of blind transmissions, with up to
32 total transmissions (including both the initial transmission and retransmissions). In this
section we consider capacity under this configuration.
34
______________________________________________________________________________________________________
Table 9. Resource Allocation of OH Items
REs/Slot REs/Total OH REs REs/Total REs

µ =0 µ =1 µ =0 µ =1 µ =0 µ =1
PSFCH 0 0 0 0 0 0
CSI-RS 0 0 0 0 0 0
PT-RS 0 0 0 0 0 0
PSCCH 240 240 10.90 % 20.82 % 2.75 % 5.95 %
SCI2 36 36 1.63 % 3.12 % 0.41 % 0.89 %
DM-RS 624 288 28.33 % 24.99 % 7.14 % 7.14 %
AGC 624 288 28.33 % 24.99 % 7.14 % 7.14 %
Guard 624 288 28.33 % 24.99 % 7.14 % 7.14 %
S-SSB 54.6 12.6 2.48 % 1.09 % 0.63 % 0.31 %
Total 2202.6 1152.6 100 % 100 % 25.21 % 28.59 %
When blind transmission is configured, all the transmissions of a TB contain the same
information bits. Note that the contribution from the S-SSB to OH is minimal. Hence, to
calculate the maximum capacity and in order to reuse Eq. (18) for an approximation, we
could take all the retransmissions as overhead in the calculation and include them into OH
in Eq. (18). To do so, since OH in Eq. (18) is per slot while blind retransmissions can
occupy multiple slots, we could map existing OHs, plus the retransmission overhead into
one slot by averaging them over the slots for the same TB, as below:
µ µ
µ,N REOH,slot + REslot · (N − 1)
OH slot = µ , (25)
REslot · N
µ,N
where OH slot is the average overhead ratio including the blind retransmissions, and N is
the configured number of transmissions per TB.
Further manipulation of the above equation yields:
µ,N 1 − OH
1 − OH slot = , (26)
N
µ
REOH,slot
where OH = REslot
µ as in Eq. (20).
Eq. (26) tells us that the capacity under the blind transmission configuration can be
captured by the revised OH. With other parameters unchanged and by noting Eq. (18), we
then have
35
______________________________________________________________________________________________________
data rate with blind transmissions

BW,µ
−6 N · 12 1 − OH
= 10 · νlayers · Qm · f · Rmax · PRB µ · (27)
Ts N
data rate without blind transmissions

= .
N
That is, Eq. (27) can be used to calculate the capacity with blind transmissions, which is N
times less than the capacity without blind transmissions.
Table 10 provides a few examples of transmission numbers and their relationships to
µ,N µ,N
REs/Slot and OH slot . It is evident that a larger N results in a significantly higher OH slot .
The capacity changes caused by blind transmissions are discussed in the following section.
Table 10. Resource Allocation with Blind Transmissions
µ,N
No. of REs/Slot REs/Total REs (OH slot )
Transmissions µ =0 µ =1 µ =0 µ =1
1 2202.6 1152.6 25.21 % 28.59 %
... ...
4 7102.65 3312.15 81.30 % 82.15 %
... ...
32 8531.83 3942.02 97.66 % 97.77 %
5.3.4 NR Capacity Results

For NR SL, in the previous two sections, Section 5.3.1 and 5.3.2, we analyzed each param-
eter and the overhead item required in Eq. (18), and derived capacity formula Eq. (27) for
a blind transmission configuration. The resulting numbers were summarized in Tables 7
and 10. By examining the two tables, we can see that different parameter combinations can
actually be narrowed down to three independent configurations: numerology, MCS table
employed, and HARQ configuration. Categorized by these three dimensions, the resulting
NR capacity is listed in Table 11 and also plotted in Fig. 13. Note that for the number
of transmissions, three numbers are highlighted for illustration purposes: 1 (equivalent to
no retranmissions), 4 (mapping to LTE for comparison purpose) and 32 (maximum NR
transmissions per 3GPP).
From Table 11 and Fig. 13 we can read the impacts of the three independent configu-
rations on the resulting capacity. For numerology, we can see that numerology 0 gives a
higher capacity number than numerology 1, and the ratio is fixed to be around 1.14 : 1. The
BW,µ
NPRB
differences come mainly from the term Ts
µ (Table 6) and OH (Table 9). As discussed in
BW,µ
NPRB
previous sections, numerology 0 has a smaller SCS, which results in both a higher Ts
µ
and a lower OH.
36
______________________________________________________________________________________________________
In addition, depending on the device capabilities, the maximum modulation order one
UE supports can be 6 or 8. This leads to either the 64 QAM or the 256 QAM MCS
table being employed without changing the values of other parameters. Consequently, the
capacity ratio between the employed two MCS tables is 3 : 4, which is consistent with the
numbers in Table 11.
Furthermore, capacity decreases significantly with an increasing number of transmis-

sions. This is consistent with the derived Eq. (27), which shows that capacity follows an
inversely proportional trend when compared to the number of transmissions.
Table 11 also shows that the maximum NR SL capacity is around 97 Mbit/s, which, as
expected and analyzed above, is achieved under configuration with numerology 0, 256 QAM
MCS table, and no HARQ retransmissions. This is significantly higher than the LTE ca-
pacity, which is discussed in the next section.
Table 11. NR SL Capacity in Mbit/s
No. of 64 QAM table 256 QAM table

Transmissions µ =0 µ =1 µ =0 µ =1
1 72.58 63.98 96.78 85.30
... ...
4 18.15 15.99 24.19 21.33
... ...
32 2.26 2.00 3.02 2.67
5.4 LTE and NR Capacity Comparison

In Section 5.2 we obtained an LTE capacity of 5.28 Mbit/s. In Section 5.3 we see that NR
capacity varies per configuration with a maximum of 96.78 Mbit/s, which is significantly
higher than LTE (around 18 times greater) and is achieved with numerology 0, 256 QAM
MCS and no retransmissions. Eq. (18) and (27) provide some insights on where the gains
come from.
The first improvement results from the use of 2-layer spatial multiplexing in NR versus
no MIMO in LTE. This gives NR two times the data rate of LTE. The second is the higher
NR MCS. NR can use 256 QAM with a 948/1024 coding rate, compared to 16 QAM
with lower coding rate in LTE. The most significant NR improvement comes from the
configurable HARQ. Considering the best NR case with no retransmissions versus LTE ,
which requires four transmissions, this gives NR approximately four times the achievable
data in LTE (Eq. (27)). Added together, these three factors contribute to approximately 18
times the data rate of LTE.
Another perspective is that in order to adapt to different deployment scenarios and
device capabilities, NR supports various configurations with different capacity numbers
as described in Section 5.3. As shown in Fig. 13, the factors discussed above still work for
various configurations except with blind retransmissions. As shown, the NR improvement
37
______________________________________________________________________________________________________
2
10
Q max = 6, =0
Q max = 6, =1
Q max = 8, =0
1, 4, 32 Transmissions
Q max = 8, =1
Capacity (Mbit/s)
101
LTE with 4 Transmissions
100
0 5 10 15 20 25 30
Number of NR HARQ Transmissions
Fig. 13. Capacity - NR vs LTE.
decreases significantly with an increasing number of transmissions, and eventually reaches

zero when the number is close to 12, 14, 16, and 18 respectively. NR capacity then falls
below LTE capacity thereafter. The configuration with such a high number of transmissions
is typically used when ultra reliability is required, and high capacity is less critical.
While NR has much higher capacity numbers in general, the numbers map to the max-
imum data rate achievable. In practice, the data may be much lower due to various factors.
One example is the significantly higher OH caused by enabling the PSFCH when feedback-
based HARQ is configured (Section 5.3.2), which would eventually lead to a much lower
data rate.
5.5 Summary
In this section, we analyzed LTE SL and NR SL capacity under Band 14 / Band n14 10 MHz
bandwidth and with a focus on unicast. Our initial quantitative results show that NR’s new
features improve capacity significantly, except when blind transmissions are configured
with a high number of transmissions. The improvement varies by configurations and can
be as high as around 18 times, which can be achieved with the highest configuration, specif-
ically, 256 QAM MCS table with numerology 0 and no HARQ retransmissions. Each of the
new features contributes to a portion of the improvement. Using the highest configuration
as an example, up to two-layer spatial multiplexing contributes two times, the higher order
38
______________________________________________________________________________________________________
of MCS contributes to slightly above two times, and the most improvement comes from the
supported configurable HARQ - by eliminating the four blind transmissions employed by
LTE, the NR capacity is improved by four times.
Note that NR allows flexible configurations to adapt to different deployment scenarios
and device capabilities, hence different capacity numbers are achieved as a balance to meet
service requirements. One extreme case is the configuration with a high number of blind
transmissions. While NR could have less capacity than LTE, this configuration is useful
when ultra reliability is required and a high capacity is relatively not so critical. Also note
that the high NR capacity in general maps to the highest data rate achievable. The actual
data rate could be lower due to various factors such as significantly higher overhead due to
the configured feedback channel.
While our initial capacity study focuses on unicast, one next step is to evaluate capacity
for multicast and broadcast together with related major factors.
6. Quality of Service
6.1 Introduction
5G NR inherits LTE’s framework of handling different QoS requirements, and enhances it
in several aspects.
Over the core network and Uu interface, NR’s QoS handling principles are similar to
LTE’s: the core network specifies QoS requirements, and the radio-access network maps
and handles packets accordingly. The exact mechanisms and terms are somewhat different
between NR and LTE though. Between the NR network and the connected device, there
is one or more Protocol Data Unit (PDU) sessions, which correspond to one or more QoS
flows and DRBs. A general overview of the NR QoS model is provided in Section 6.2
first, to better understand and compare with the QoS operations over NR SL described in
the following sections. Unlike LTE, the mapping between QoS flows to DRBs are done
by a separate layer in the Radio Access Network (RAN), namely, the SDAP layer. A brief
introduction of the SDAP layer is given in Section 6.3, including SDAP’s support for NR
SL QoS handling.
The QoS support for direct mode communication over SL has been improved consider-
ably in NR. Section 6.4 describes the NR SL QoS handling in a nutshell, and points out the
correspondence between LTE ProSe QoS parameters and NR SL QoS parameters. Section
6.5 focuses on the new QoS parameter, Range, that NR provided for direct mode commu-
nication. Although the majority of the standardized NR SL QoS mechanisms so far are
provisions for V2X scenarios, they may be applicable to public safety use cases as well. In
addition, there are ongoing 3GPP discussions on QoS support designed for public safety
applications specifically, as listed in Section 6.6.
39
______________________________________________________________________________________________________
6.2 Overview of New Radio Quality of Service Model

Fig. 14 illustrates NR’s overall QoS architecture briefly. 5G Core Network (5GC) estab-
lishes one or more PDU sessions for each UE. There may be one or multiple QoS flows
within each PDU session. The Next-Generation Radio Access Network (NG-RAN) may
map one or multiple QoS flows to the same DRB. 5GC and NG-RAN map packets in both
directions with proper QoS flows and DRBs to satisfy QoS requirements. Table 12 outlines
UL and DL operations for mapping packets from IP-flows to QoS flows and then from
QoS flows to DRBs. The QoS Flow Identifier (QFI) mentioned in the table is embedded
in the header over the user plane interface (NG-U) between NG-RAN and 5GC. It identi-
fies a QoS Flow within a PDU session, and its value may be dynamically assigned or may
be equal to the 5G QoS Identifier (5QI), a QoS parameter which will be discussed later.
The reflective mapping in the UL is a new function introduced in NR. When it is enabled,
a UE monitors the QoS flow to DRB mapping rule in the DL, and applies it in the UL
accordingly.
Fig. 14. QoS architecture [22, Fig. 12-1].2
2 ©2020 - 3GPP™ deliverables and material are the property of Association of Radio Industries and Busi-
nesses (ARIB), Alliance for Telecommunications Industry Solutions (ATIS), China Communications Stan-
dards Association (CCSA), European Telecommunications Standards Institute (ETSI), Telecommunications
Standards Development Society, India (TSDSI), Telecommunications Technology Association (TTA) and
Telecommunications Technology Committee (TTC) who jointly own the copyright in them. They may be
subject to further modifications and are therefore provided to you ”as is” for information purposes only.
Further use is strictly prohibited.
40
______________________________________________________________________________________________________
DL UL
QoS profiles QoS rules
Characterization of QoS flows @ NAS level
(provided by 5GC to NG-RAN) (provided by 5GC to the UE)
From IP-flows to QoS flows:
Done by NAS level packet filter @5GC Done by NAS level packet filter @ UE
Non-Access Stratum (NAS) level mapping
Follow AS level mapping rules @ NG-RAN. Follow AS level mapping rules which are
From QoS Flows to DRBs:
Based on QoS Flow Identifier (QFI) and signaled to UE either explicitly (RRC) or
AS level mapping
the associated QoS profiles. implicitly (reflective mapping)
Table 12. NR QoS Mapping Operations
As listed in Table 12, NG-RAN uses the QoS profile to determine how to handle packets
over the radio interface according to QoS requirements. The QoS profile of a QoS flow
contains QoS parameters, such as 5QI and Allocation and Retention Priority (ARP). In
addition, depending on whether the QoS flow is a Guaranteed Bit Rate (GBR) flow or not,
more QoS parameters may be included in the QoS profile. For a Non-GBR QoS flow,
the Reflective QoS Attribute (RQA) parameter and other additional QoS Flow Information
may be included. For a GBR QoS flow, the following QoS parameters may be specified
[23]:
• Guaranteed Flow Bit Rate (GFBR) for both UL and DL;
• Maximum Flow Bit Rate (MFBR) for both UL and DL;
• Maximum Packet Loss Rate for both UL and DL;
• Delay Critical Resource Type;
• Notification Control.
Among the above QoS parameters, the 5QI is associated to QoS characteristics such as
priority level, Packet Delay Budget (PDB), Packet Error Rate (PER), averaging window,
and maximum data burst volume. If the 5QI value in a QoS profile is a standardized or
pre-configured one, such as those listed in Table 13, the 5G QoS characteristics of that
QoS flow can be derived from the 5QI value and do not need to be explicitly signaled.
Otherwise, QoS characteristics will be signaled as part of the QoS profile of that QoS flow.
Besides the above QoS requirements, which specify per QoS flow expectations, a few
other parameters may also be configured to guide/limit the interaction and coexistence
among different flows and/or sessions. For example, the Session-Aggregate Maximum Bit
Rate (AMBR) limits the aggregate bit rate that can be provided across all Non-GBR QoS
Flows for a specific PDU Session. The UE-AMBR limits the aggregate bit rate that can be
expected across all Non-GBR QoS Flows of a UE, and it is guaranteed by the RAN. When
the transmission demand from various flows cannot be met simultaneously, prioritization
will be done based on QoS information.
When comparing 5QI defined for NR ([23] Table 5.7.4-1) with QoS Class Identifier
(QCI) defined for LTE ([24] Table 6.7.1-A), we can see that new values 82, 83, 84, 85,
and 86 are introduced in NR, among which 83, 85, and 86 are recommended for V2X
41
______________________________________________________________________________________________________
Table 13. SL related standardized 5QI to QoS characteristics mapping (selected from [23, Table
5.7.4-1])3 .
messages. Also, a couple of additional QoS characteristics are indicated by 5QI for NR,
namely, ”Maximum Data Burst Volume” and ”Averaging Window”.
6.3 SDAP Layer for Quality of Service

The SDAP layer is a new layer in the user plane, responsible for mapping QoS bearers to
radio bearers based on QoS requirements. It is added above the Packet Data Convergence
Protocol (PDCP) layer, and is specified by 3GPP standard TS 37.324 [25]. The main
functions of SDAP layers are [22]:
• transfer of user plane data;
• mapping between a QoS flow and a DRB for both DL and UL;
• mapping between a PC5 QoS flow and an SL-DRB for NR SL communication;
3 ©2020- 3GPP™ deliverables and material are the property of ARIB, ATIS, CCSA, ETSI, TSDSI, TTA and
TTC who jointly own the copyright in them. They may be subject to further modifications and are therefore
provided to you ”as is” for information purposes only. Further use is strictly prohibited.
42
______________________________________________________________________________________________________
• marking QFI in both DL and UL packets;
• marking PC5 QoS Flow Identifier (PFI) in unicast mode of NR SL communication

packets;
• reflective QoS flow to DRB mapping for the UL SDAP data PDUs.
The above functions can be categorized roughly into two types: 1) mapping functions
and 2) data marking and transfer functions.
Among SDAP mapping functions, the reflective QoS flow to DRB mapping is a new
function introduced in NR. However, for NR SL communication, reflective PC5 QoS flow
to SL-DRB mapping is not supported. Regarding other mapping functions, as illustrated in
Fig. 15, a single SDAP entity is configured for each PDU session. Multiple-to-one mapping
is allowed between QoS flows and DRBs, while one QoS flow is mapped onto only one
DRB at a time in the UL. The mapping rules are configured in the device using RRC
signaling. For NR SL communications, similarly, there is a one-to-one mapping between
an SDAP entity and each Destination Layer-2 ID and cast type in the UE. Multiple-to-one
mapping is allowed between PC5 QoS flows and SL-DRBs, while one PC5 QoS flow is
mapped onto only one SL-DRB at a time for the SL transmission.
Fig. 15. SDAP sublayer structure view [25, Fig. 4.2.1-1]4 .
The SDAP data marking and transfer functions are supported by five data PDU formats
and one control PDU format defined in [25], among which, three PDU formats are relevant
43
______________________________________________________________________________________________________
to NR SL communications. Groupcast and broadcast of NR SL communication only sup-

port the SDAP data PDU format without an SDAP header ([25, Figure 6.2.2.2-1] ). Unicast
SL communication supports the SDAP data PDU format with an SDAP header ([25, Figure
6.2.2.4-1]). UE’s SDAP entity uses End-Marker control PDU ([25, Figure 6.2.3-1]) to stop
the mapping from the SDAP Service Data Unit (SDU) of the QoS flow indicated by the
QFI/PFI to the DRB/SL-DRB on which the End-Marker PDU is transmitted.
6.4 Quality of Service over New Radio Sidelink

For NR SL communication towards the same destination identified by a Destination Layer-
2 ID, the PC5 QoS Flow is the finest granularity of QoS differentiation. Each QoS flow is
identified by a PFI which is unique within the same destination, and the same traffic for-
warding treatment (e.g., scheduling and admission threshold) applies to all User Plane (UP)
traffic with the same PFIs. Note that the role of PFIs for NR PC5 communication is quite
similar to the role of QFIs for NR communication over the Uu interface. As illustrated in
Fig. 16, multiple radio bearers may be established between a pair of source and destination
Layer-2 IDs, and each radio bearer may correspond to a different PC5 QoS level. Multiple
PC5 QoS Flows may be mapped to the same radio bearer, as determined by the AS layer
based on the information provided. For broadcast and groupcast mode communication, the
L2 link goes to all UEs in proximity identified by the destination Layer-2 ID. Each PC5
QoS Flow is associated with a PC5 QoS Rule and the following PC5 QoS parameters [26]:
• PC5 5QI (PQI);
• PC5 Flow Bit Rates: for GBR QoS flow only, and may include GFBR and MFBR;
• PC5 Link Aggregated Bit Rates: for unicast links and non-GBR flows, and may
include per link AMBR (PC5 LINK-AMBR);
• Range: for groupcast type only;
• Default Values: will be used if the corresponding PC5 QoS parameter is not provided
by the upper layer.
Among the PC5 QoS parameters above, the Range parameter is a new and unique QoS
parameter introduced for NR SL groupcast, and its definition and usage are further studied
in Section 6.5. A PQI is a special 5QI, and there is no overlapping between the values of
5QIs and PQIs. The association between standardized PQI values and PC5 QoS character-
istics is shown in Table 14. It is also possible to pre-configure PC5 QoS characteristics and
indicate them through the PQI value. In addition, non-standardized PC5 QoS characteris-
tics may override the standardized or pre-configured value of PC5 QoS characteristics with
restrictions.
Note that when a UE is out of coverage, pre-configuration may be used for NR SL
communication. In the situation when the UE’s RRC state changes but the updated SL
44
______________________________________________________________________________________________________
Fig. 16. Handling of PC5 QoS Flows based on PC5 QoS Rules [26, Fig. 5.4.1.1.3-1]5 .
DRB configuration(s) hasn’t been received for the new RRC state, UE may continue SL
data transmissions and receptions using the current configuration of the previous RRC state
until getting the configuration for the new RRC state.
For LTE based direct mode communication, i.e., ProSe, the QoS handling is based on
ProSe Per-Packet Priority (PPPP) and ProSe Per-Packet Reliability (PPPR) as defined in
[27]. For NR based direct mode communication over the PC5 SL, the QoS model is more
comprehensive and is quite similar to the QoS model defined for non-SL communication
over the Uu interface, with additional parameters such as Range (see Section 6.5). Within
the NR SL QoS framework, two QoS characteristics can be viewed as the NR counterpart
for the LTE ProSe QoS parameter, PPPP: Priority Level and PDB [26]. NR QoS’s Priority
Level over NR PC5 has the same format and meaning as LTE PPPP’s Priority value over
LTE PC5. In addition, NR SL’s PDB is derived from the PQI value, e.g., through Table 14,
while LTE SL’s PPPP value indicates the latency requirement and determines the PDB.
When compared with the NR QoS model over the Uu interface in Section 6.2, it can
be seen that the QoS characteristics for SL communication reuse those defined for non-
SL communication with minor modifications. In addition, the QoS parameters defined for
communications over NR PC5 are similar to those defined for NR Uu in general, with some
differentiation, such as the range parameter that is specific to PC5 and the RQA parameter
45
______________________________________________________________________________________________________
Table 14. Standardized PQI to QoS characteristics mapping [26, Table 5.4.4-1]6 .
that is specific to the Uu interface.
6.5 Minimum Communication Range

A new feature introduced in NR SL group communication is the ability to control the area
where the QoS needs to be maintained. As shown in Fig. 17, a transmitting device can
specify the communication range requirements associated with a packet transmission. This
information is communicated inside the SCI format 2-B control message in the form of
an index that maps to a SL pool parameter, which ultimately represents the distance, in
meters, where the QoS parameters apply. The UE determines the value to apply based
on the longest communication range (defined by the PC5 QoS parameter) of the logical
channels in the Medium Access Control (MAC) PDUs to transmit. A user (e.g., RX 1)
that is within the communication range will send a NACK message back to the transmitter
if there is a decoding error. For devices situated outside the communication range (e.g.,
46
______________________________________________________________________________________________________
RX 2), the communication is best effort, which means that while they may still be able to
receive the information, they will not send a NACK message even if there is a decoding
error. Since the feature is based on the distance between two devices, the devices’ locations
are needed to operate properly. In the event that the transmitter does not have its location,
the feature cannot be used even if a logical channel was configured to use it. When it
is a receiver device (e.g., RX 3) that cannot determine its location, 3GPP has defined a
conservative approach and the device assumes that it is within range and will therefore
send a NACK if there is a decoding error. The transmitter location is also indicated in
the SCI format 2-B using a zone identifier that is derived from its latitude and longitude
information. A receiver device then takes the center of the zone as an approximation for
the transmitter’s location to calculate its distance to the transmitter.
Fig. 17. Overview of Minimum Communication Range (MCR) for group communication.
This new capability enhances the reliability of the communication in group scenarios
by allowing retransmissions until all members of the group decode the transmission, up
to the maximum number of retransmissions allowed, without adding overhead to create
groups. Introduced to support V2X scenarios, it is very useful since vehicles come in and
out of range of each other due to traffic patterns and user itineraries. Since first responders
heavily rely on voice, group communication, they could certainly benefit from it. However,
research is still necessary to understand and optimize the configuration parameters. For ex-
ample, we need to evaluate the impact of the communication range parameter on the group
47
______________________________________________________________________________________________________
performance. If the communication range is too high (in Release 16, the maximum range is
1000 m), it may lead to constant retransmissions and may affect overall performance (e.g.,
latency) by trying to reach users beyond the incident area. The range is also likely to be
dependent on the environment or scenario, with a larger range applicable to open spaces
and a shorter range applicable to urban environments.
6.6 Quality of Service Support for Public Safety

As defined in Table 13, separate sets of 5QIs have been assigned to mission critical services
(i.e., 65, 67, 69, and 70) and V2X messages (i.e., 3, 79, 83, 85, and 86) already, among
which, 5QI-65 and 5QI-69, are expected to be used together to provide MCPTT service.
Besides the NR QoS handling already in place for general applications, there are ongoing
3GPP standardization efforts to provide more tailored support to public safety services. In
addition to existing PQI values defined in Table 14, 3GPP working groups are also dis-
cussing the necessity and feasibility of defining new PQI values for public safety services
over NR SL [28]. The new PQIs proposed in [28] for public safety services are:
• New value #1: default priority level 1, PDB 150 ms, and PER 0.01. It may be
assigned to services such as mission critical user plane push to talk voice;
assigned to services such as non-mission-critical user plane push to talk voice;
assigned to services such as mission critical video user plane;
• New value #4: default priority level 1, PDB 120 ms, and PER 0.000 001. It may be
assigned to services such as mission critical delay sensitive signaling;
• New value #5: default priority level 6, PDB 400 ms, and PER 0.000 001. It may be
assigned to services such as mission critical data.
Among the above new PQI values, new values #1, #2, and #3 are defined for GBR resource
type, and the default averaging window is 2000 ms. New values #4 and #5 are defined for
non-GBR resource type.
Although the support for these public safety specific PQI values is still under discussion,
it is hopeful that the QoS requirements associated with these new PQIs can be met by
adapting existing NR SL QoS handling. To be more specific, new values #1, #2, and #3
may be met by PQI = 91 in Table 14 directly. It is also very likely that new value #4 may be
met by adapting PQI = 91 properly, because it may be feasible to improve the packet error
rate by one order of magnitude when the packet delay budget is relaxed by 40 times and the
priority level is promoted. There is some uncertainty about meeting the QoS requirements
of new value #5 considering the downgraded priority level and the stringent packet error
rate. However, combining the QoS handling of PQI = 56 and PQI = 91 may be a promising
starting point.
48
______________________________________________________________________________________________________
Please note that these new PQI values are still under discussion, and have not been
standardized by 3GPP at the time this document was drafted. Please check the latest version
of [26] for the up-to-date, standardized PQI values.
6.7 Summary
As can be seen from above sections, the QoS support for direct mode communication over
SL has been greatly enhanced in NR. In contrast to LTE’s limited and simplified QoS
control for ProSe communications, NR’s QoS mechanism for SL communication is full-
fledged, and is a miniature version of that for communication over the Uu interface. With
the enhanced QoS framework enabled for NR SL communication and the ongoing stan-
dardization efforts to support public safety applications, we expect that 5G NR has great
potential to meet QoS requirements of direct mode communications for public safety ser-
vices.
7. LTE/NR Coexistence
7.1 Introduction
Operators have already rolled out 5G NR in select markets. However, the majority of users
are still equipped with 4G LTE phones and it will take some time before a full nation-
wide 5G network becomes available. In addition, most frequency bands below 2.5 GHz
are FDD and are already occupied, such as Band n14. Refarming, which would require to
completely stop using LTE and switch to NR, would be expensive and cause service dis-
ruption for many users. To address this spectrum limitation and enable a smooth transition,
3GPP has defined mechanisms to support Dynamic Spectrum Sharing (DSS) [29]. In the
following sections, we first describe DSS with an emphasis on the UL then discuss if it can
be leveraged to support SL communications.
7.2 DSS Overview

DSS provides mechanisms for sharing part of the spectrum between LTE and NR to en-
able a smooth transition between the two technologies. This is accomplished by enabling
coordination between eNodeBs and gNodeBs to ensure that specific subframes and RBs
are used by one technology at a time. DSS supports both time multiplexing (for static and
semi-static coordination) and frequency multiplexing (for more dynamic coordination). NR
numerology 0 and 1 can be used for DSS, but each brings its own set of constraints with
regards to how it can be used, such as granularity of time and frequency that can be shared.
Coordination between neighboring cells is done by exchanging two types of information
over the Xn interface: configuration and data allocation.
Configuration information is exchanged via either the XN SETUP procedure or the
NG-RAN NODE CONFIGURATION UPDATE, as shown in Fig. 18. Those procedures
are used to provide or update application-level configuration data needed for two NG-RAN
49
______________________________________________________________________________________________________
nodes to interoperate properly over the Xn-C interface, which is the signaling plane be-
tween NG-RAN nodes. An NG-RAN node is defined as either an ng-eNodeB or a gNodeB.
Fig. 18. Xn setup and Node Configuration messages [29].
When an eNodeB shares its configuration, a neighboring node can compute the list
of resources that will be used to carry cell specific recurring signals, such as DL and UL
control and synchronization signals. A portion of the message is dedicated to providing
Served Cell Information E-UTRA and contains the following information:
• Physical Cell ID (PCI), PLMNs IDs.
• Band information (frequency and bandwidth).
• NarrowBand-Internet of Things (NB-IoT) information.
• TDD allocation
• Antenna port, Physical Random Access Channel (PRACH) configuration, Multicast-

Broadcast Single-Frequency Network (MBSFN) configuration.
• Protected E-UTRA resource indication (i.e., DL and UL control/reference signals

protected and not subject to (data) resource coordination).
– Activation System Frame Number (SFN).

– Physical Downlink Control Channel (PDCCH) and MBSFN control region lengths.
– A list of up to 16 patterns, each containing:
* Type: E-UTRA DL non-CRS, E-UTRA DL CRS, or E-UTRA UL.
* Bitmap of REs occupied by the protected signal within one PRB (if bitmap
contains all ones then it is PRB-level granularity). For example, in LTE
UL, the length of the bitmap would be 84 because there are 12 subcarriers
and 7 symbols per PRB.
* Bitmap with frequency pattern where protected pattern applies inside the
carrier bandwidth. The bitmap size equals the number of PRBs in carrier
bandwidth, e.g., 50 for a bandwidth of 10 MHz.
50
______________________________________________________________________________________________________
* Periodicity of the pattern (e.g., 1 = every slot, 2 = every other slot, etc.)
and start of pattern (in subframe).
The SFN is necessary to indicate when the configuration will be valid in case it changes.
We also notice that the MBSFN configuration is included. This is an important piece to en-
able sharing between LTE and NR because it allows the use of subframes configured for
MBSFN that have a lower overhead compared to normal DL subframes, since there are no
reference signals broadcasted in the data region. More information can be found in [30],
where there is a detailed analysis of DSS in the DL, including its impact on the perfor-
mance. With regards to UL, we observe that all the necessary control and synchronization
signals are supported. In particular, there is the PRACH configuration to determine when
and where the PRACH channel occurs. The Protected E-UTRA resource indication can be
used to specify regions allocated for Physical Uplink Control Channel (PUCCH) and SRS
transmissions. An example is shown in Fig. 19, where a PUCCH Type 2 region and an SRS
region are defined for frames using normal CP on a 10 MHz band. For the PUCCH region,
all the REs in the slot are used, thus the 84 bits (e.g. 12 subcarriers x 7 symbols) are set to
1. The frequency pattern has the first and last three bits set to 1 to indicate that it affects the
first and last three RBs of the slot. The time periodicity is 1 because the PUCCH repeats
every slot, and the start time is 1 to indicate slot 0 of subframe 0 of the SFN specified in the
message. For the SRS region, which occupies only the last symbol of the second slot, only
the last 12 bits of the resource footprint are set to 1. The frequency pattern excludes the
first and last three RBs since those are already occupied by the PUCCH. Finally, the time
periodicity is set to 2, because the SRS occurs every other slot, and the start time is set to 2
to indicate that it starts on slot 1 of subframe 0 of the SFN specified in the message.
Fig. 19. Example of protected resources pattern definitions in UL.
After defining areas that are dedicated to E-UTRA control and synchronization signals,
neighboring cells can coordinate the remaining resources used for data transfer. This is
51
______________________________________________________________________________________________________
achieved by exchanging E-UTRA-NR CELL RESOURCE COORDINATION messages,

as shown in Fig. 20.
Fig. 20. Xn resource coordination messages [29].
Both eNodeBs and gNodeBs can initiate the request, each providing the list of resources
it plans to use via the Data Traffic Resource Indication field containing the following infor-
mation:
• SFN of the receiver node in which the agreement is valid.
• Whether the node is providing resource indication for the UL only or for both the DL
and the UL.
• Bitmap(s) with resources expected to be used for E-UTRA Physical Downlink Shared
Channel (PDSCH) and/or Physical Uplink Shared Channel (PUSCH).
– Size 6 to 17 600, with each bit representing 1 PRB starting with lowest sub-
carrier frequency in the carrier starting with subframe 0 (Minimum size of 6 =
number of RBs in 1 subframe with 1.4 MHz band).
– Bitmap must be multiple of UL/DL bandwidth, i.e., must be multiple of consec-
utive subframes. Therefore with 10 MHz band (50 RBs), the maximum number
of subframes that can be specified is 352. An example of such bitmap and its
visual representation is shown in Fig 21.
• List of reserved subframes where the resource allocation does not hold
Due to the limited number of subframes that can be specified and the dynamic nature of
the traffic, coordination messages have to be exchanged periodically between neighboring
nodes. How often will depend on the operator and the ability to predict future allocations.
Similarly, the algorithms used to converge to an agreed spectrum sharing arrangement, in
case two or more cells decide to use the same resources, will be operator specific.
52
______________________________________________________________________________________________________
Fig. 21. Example use of the Data Traffic Sharing Resource Indication field.
7.3 Sidelink Considerations

As currently defined, DSS provides both DL and UL resource coordination by specifying
broadcast, control, and shared resource information. In order to support deployments with
LTE SL and/or NR SL, additional resource information must be communicated. Table 15
provides a summary of the various channels used by each technology, the frequency of
transmissions, and RB locations.
Table 15. LTE SL and NR SL channels
Channel LTE NR
Periodicity of 16 frames, with up to 64
When activated by transmitting UE, ev-
transmissions within period based on
PSBCH ery 40 ms on the center six RBs of the
numerology (one transmission for nu-
frequency band.
merology 0).
Periodic allocation between 40 ms and
PSCCH Transmitted in same slot as PSSCH.
320 ms, up to 40 ms long.
Periodic allocation between 40 ms and Inside Sidelink BWP. Number of sub-
PSSCH 320 ms. Number of subframes and RBs frames and RBs based on pool config-
based on pool configuration. uration.
Physical
Sidelink Periodic allocation between 320 ms N/A. Not defined in Rel-16. Discovery
Discovery and 10.24 s. Number of subframes and messages may be sent using PSSCH in
Channel RBs based on pool configuration. Rel-17.
(PSDCH)
PSFCH N/A Transmitted in same slot as PSSCH.
There are different ways to address the current limitations depending on the technology
53
______________________________________________________________________________________________________
and allocation type:
• For out-of-coverage operation, which would not require DSS, it is possible to pre-
configure non overlapping resource pools for both LTE and NR SL. Those could be
separated in frequency, time, or both.
• For in-coverage scenarios and UE selected resource pools, it would be beneficial for
the eNodeB or gNodeB to send the resource pool configuration as part of the SETUP
or NG-RAN NODE CONFIGURATION UPDATE in order to limit overhead. This is
because the eNodeB or gNodeB is not aware of the specific resource selection made
by the UEs, and therefore it would be preferable that a neighboring node does not
use any of the resources allocated for SL. This however, would require modifications
to the 3GPP specifications to add SL specific information elements.
• For in-coverage scenarios and scheduled mode, where the eNodeB/gNodeB selects
specific resources for the SL transmissions, it could be possible to use either the
SETUP or NG-RAN NODE CONFIGURATION UPDATE messages, or the CELL
RESOURCE COORDINATION message. In the former case, overhead is reduced
but may lead to lower efficiency if there is not a lot of SL traffic. This is because
the eNodeB/gNodeB would try to completely avoid all resources allocated for SL, as
in the case for UE selected resource pools. With the latter, an eNodeB or gNodeB
could indicate specific resources allocated to SL and provide finer control over the
shared spectrum. If there is no SL traffic, the eNodeB/gNodeB could transmit in the
UL even in resources that are part of the SL resource pools. Either option would also
require modifications to the 3GPP specifications.
7.4 Summary
In this section, we described DSS specified by 3GPP to facilitate coexistence between LTE
and NR and allow efficient use of the spectrum, which is very limited in the lower bands.
Based on the current specifications in Release 16, we noted several limitations with regards
to supporting coexistence when considering both LTE and NR SL communications.
8. Ongoing Standardization
8.1 Introduction
3GPP is currently working on Rel-17 specifications, and a firm timeline for the release has
been approved at the December 2020 plenary meeting: the Stage 2 functional freeze is
planned for June 2021, the Stage 3 protocol freeze for March 2022 and finally, the protocol
coding freeze is aimed for June 2022 [31]. Rel-17 will implement several enhancements
and new features to further support direct mode communications in the 5G NR architecture.
In this section, we provide an overview of the related 3GPP study and work items that were
prioritized for inclusion in Rel-17.
54
______________________________________________________________________________________________________
8.2 Proximity based Services in 5GS work items

Proximity based Services, or ProSe, was the first technology to support direct communi-
cation between UEs for LTE systems. LTE ProSe was introduced in Rel-12 and led to the
creation of the LTE SL, which was later enhanced to support LTE V2X. ProSe use cases
for public safety communications identified in [32] comprise in-network and off-network
operation of device discovery, one-to-one, group, and broadcast traffic exchange between
first responder’s UEs in proximity, and coverage extension with UE-based relays. More re-
cently, commercial ProSe use cases were identified in [33], such as interactive gaming and
augmented reality between UEs in proximity, real-time data exchange between nearby sen-
sors, and lightweight wearable devices connected to computationally capable UEs. Given
that these ProSe use cases are associated to different requirements than those of V2X, and
that ProSe functionalities are key factors to the support of mission critical communications
in 5G System (5GS) within the NR architecture [34], it was important to study their fea-
sibility within Rel-16 NR SL and V2X solutions, and to identify any extension needed to
support them in Rel-17.
8.2.1 Study Item phase

The study item entitled Study on System enhancement for Proximity based Services in 5GS
(FS 5G ProSe) [35] and its resulting technical report TR 23.752 [36] focused on identify-
ing technological gaps and potential solutions to support ProSe on the 5G NR architecture.
The FS 5G ProSe study evaluated direct discovery, direct communication, and UE-based
relay ProSe functionalities, and aimed to find common solutions to satisfy commercial and
public safety requirements. In addition, the Study on NR Sidelink Relay (FS NR SL Relay)
complemented the FS 5G ProSe study by focusing on radio access network aspects of the
UE-based relay ProSe functionalities.
NR ProSe Direct Discovery The definition of direct discovery did not change from LTE
ProSe to NR ProSe. Based on [36], the ProSe direct discovery procedure is still used for
a UE to discover or be discovered by other UE(s) in proximity over the PC5 interface
(as it was the case for LTE ProSe). The UE can discover other UE(s) with interested
application(s) and/or interested group(s) using the ProSe direct discovery procedure. This
feature aims to achieve the common direct discovery procedure for discovering a 5G ProSe
enabled UE, 5G ProSe U2N Relay, or a 5G ProSe UE-to-UE (U2U) Relay.
Multiple aspects (architecture, message, resource pool, protocol stack) of the direct dis-
covery are part of the ongoing work by the standardization. However, certain suggestions
have more approval than others including:
• Direct discovery message would be taken as a new signaling in the ProSe layer,
separate from PC5-S signaling. The ProSe layer would indicate to the AS layer
whether the signaling is a discovery message or PC5-S signaling.
55
______________________________________________________________________________________________________
It is feasible to configure a dedicated priority value for the discovery message, which
is different from the other SL Signaling Radio Bearer (SRB).
• Resource pool to transmit discovery messages can be either shared with or separated
from the data transmission resource pool.
In case of a shared resource pool, a new Logical Channel ID (LCID) is introduced

for discovery messages, i.e., discovery messages are carried by a new SL SRB.
For a separated resource pool, discovery messages are treated equally with each other
during the Logical Channel Prioritization (LCP) procedure. It does not seem neces-
sary to introduce a new LCID for discovery messages since the discovery message
can be identified by a different resource pool.
It is highly expected that new SRB and LCID are introduced for discovery messages
since discovery messages are not a part of PC5-S signaling and they are neither PC5
RRC messages nor SL data.
• Since the direct discovery message is taken as new signaling in ProSe layer sepa-
rately from PC5-S signaling, the protocol stack of direct discovery can be depicted
as Fig. 22, in which the Discovery is located on top of the PDCP, Radio Link Control
(RLC) and MAC sublayers.
Fig. 22. Protocol stack for direct discovery.
NR ProSe Direct Communication Direct communication is the functionality that de-

fines one-to-one and one-to-many direct traffic exchange between nearby UEs for ProSe.
56
______________________________________________________________________________________________________
ProSe direct communication should allow first responders in proximity to use mission crit-
ical applications to communicate even in cases of network unavailability. On the com-
mercial side, the use of direct communication between UEs in proximity, instead of the
network, is expected to provide enhanced experience for interactive applications such as
gaming or augmented reality. Moreover, ProSe direct communication is the base for the
UE-based relaying functionalities and coverage extension use cases that will be presented
in the next section.
Similar to V2X applications over PC5, UEs using ProSe direct communication will
use the SL over the PC5 interface to transmit and receive data and signaling messages,
if applicable. Thus, this functionality was found feasible for 5G NR systems in [36] by
reusing the V2X communication over PC5 reference point mechanisms defined for Rel-16
in [26], which comprise unicast, broadcast, and groupcast direct communication between
UEs using the NR SL. These mechanisms, together with several extensions such as the
support of Internet Protocol version 4 (IPv4), Address Resolution Protocol (ARP), Ethernet
and Unstructured data unit types, and the support of enhanced QoS handling, are envisioned
in [36] to fully support ProSe direct communication use cases.
NR Sidelink Relays Two UE-based relay capabilities were studied in [36]: UE-to-Network
(U2N) relay, where a UE extends the network connectivity to another nearby UE by using
direct communication; and UE-to-UE (U2U) relay, where a UE uses two direct communi-
cation links to connect two UEs in its proximity that otherwise are not able to communicate.
The U2U relay functionality was not part of the LTE ProSe specification, and its in-
clusion on NR ProSe can be beneficial for public safety communications range extension
for both in-network and off-network use cases. U2N relay functionality is fundamental for
network coverage extension for public safety interventions in remote areas, as well as, for
wearable devices tethering in commercial use cases (e.g., sensors, virtual reality headsets).
The Study on NR Sidelink Relay (FS NR SL Relay) and resulting technical report TR
38.836 [37] complemented the FS 5G ProSe study focusing on radio access network as-
pects of the UE-based relay ProSe functionalities. Both, the U2N and U2U relay function-
alities were found feasible by reusing some of the LTE ProSe and NR V2X components,
and defining new architectures and protocols.
LTE U2N relay functionality uses a Layer 3 (L3) architecture in which the relay of
data packets in the PC5 interface is performed at the network layer, and UEs connected
to a L3 U2N relay are transparent to the network. For NR UE-based relay solutions, two
architectures were found feasible in [36] and [37]: the L3 architecture as in LTE, and a
newly defined Layer 2 (L2) architecture in which the relaying in the PC5 interface occurs
within the L2, over the RLC sublayer. A UE connected to an L2 U2N relay is expected to
be seen by the network as a regular UE (i.e., as if it was directly connected to the network),
which gives the network control of the connection and services, but requires the definition
of several new mechanisms not present or needed in the L3 architecture. These would
include, at minimum, a PC5 to Uu adaptation layer for RLC channels and bearer mapping,
indirect paging and system information forwarding, and network controlled path switching.
57
______________________________________________________________________________________________________
Architecture-independent components from previous releases were agreed to be used

as baselines for Rel-17. The relay discovery models and the relay selection criteria and
procedures will be based on the LTE ProSe solutions. Rel-16 V2X unicast communication
protocol will be used as a baseline for both architectures. This includes the establishment,
maintenance, modification, and release of a connection between a UE and UE-based relay,
and the corresponding data exchange using direct communication. Finally, QoS aspects of
Rel-16 will be reused and extended to support mapping between PC5 and Uu QoS param-
eter and rules to ensure end-to-end QoS for a UE connected to a relay UE.
8.2.2 Work Item phase

As shown in the previous section, the technical working groups involved in FS 5G ProSe
and FS NR SL Relay studies found the ProSe functionalities feasible to be implemented
within the 5G NR architecture in Rel-17. Currently, the corresponding technical working
groups are developing the normative specification of the identified solutions under two
work items:
• 5G ProSe work item, with the current scope described in [38], and
• NR SL Relay work item, with the current scope described in [39].
5G ProSe technical specifications are being documented in TS 23.304 [40], and technical
solutions are expected to impact other specification documents as listed in [38] and [39]. At
the time of writing this document, ProSe discovery and ProSe communication normative
work is ongoing, while both L2 and L3 architectures for U2N relay functionality were
recently approved and the normative work should start shortly. Due to time and workload
constraints, the U2U relay functionality was excluded from the current Rel-17 work items
scopes.
8.3 NR sidelink enhancement work item

The NR sidelink enhancement (NR SL enh) work item [41] is ongoing at the time of writing
this document and it aims to specify radio solutions for Rel-17 that can enhance the NR SL
for all services expected to use it, including V2X, public safety, and other commercial use
cases.
3GPP radio access network working groups identified important NR SL topics that
were not fully covered by the solutions specified in Rel-16, and that could be addressed in
Rel-17 [42]. This includes power savings, enhanced reliability, reduced latency, increased
data rate, support of new carrier frequencies and operation scenarios, increased spectral
efficiency, and enhanced coverage. The topics related to UE power saving and to enhanced
reliability and reduced latency were selected to be the main focus of the NR SL enh work
item specifications for Rel-17, although objectives related to the support of new carrier
frequencies and restricted use of frequency ranges are also present in the work item. It was
also determined that extensions to the NR SL evaluation methodology should be discussed
58
______________________________________________________________________________________________________
to include public safety and commercial use cases in addition to existing V2X evaluation
scenarios. The current scope of the NR SL enh work item can be found in [41]. The topics
related to coverage enhancements are to be covered by work items specifying UE-based
relay technologies (i.e., FS NR SL Relay and 5G ProSe), and the specification of the other
topics in future releases is still an open subject at the time of writing.
Some of the NR SL use cases expected to be supported in Rel-17 have strict reliability
and latency requirements. The RAN working groups found that the solutions specified in
Rel-16 may not satisfy those requirements in all conditions, specially when using resource
allocation mode 2, i.e., when each UE autonomously selects its resources, and during busy
channel conditions [42] . Thus, the NR SL enh work item currently focuses on study-
ing and specifying solutions that can further improve Rel-16 reliability and reduce latency
within resource allocation mode 2. In particular, inter-UE resource allocation coordina-
tion schemes are being considered, in which UEs exchange information that can assist the
resource selection of the others. This can further reduce resource collision and thereby
improve reliability.
The power savings enhancements are motivated by Rel-17 uses cases for the NR SL,
which go beyond the vehicular communication defined on Rel-16, in which devices do
not have strong power limitations. In Rel-17, battery-constrained portable devices such as
smartphones, head or backpack mounted devices, and wearables, are expected to use the
NR SL in order to support advanced V2X use cases, such as Vulnerable Road User (VRU)
warning, and also to enable ProSe use cases, such as direct communication for public safety
mission critical operations and commercial interactive applications. Two main aspects were
identified in the work item to enhance UE power savings: reducing sensing active time used
for resource allocation, e.g., by using partial sensing or no sensing at all; and introducing
SL Discontinuous Reception (DRX) in which the devices are active for data and control
reception only when deemed necessary. The random resource selection with no sensing
and the partial sensing schemes of LTE SL, introduced in Rel-14, are going to be used as a
baseline for the power efficient resource allocation schemes of Rel-17. How to adapt them
to the NR SL and whether another solution may be needed are under discussion at the time
of writing. The NR SL DRX will reuse the Uu interface design aspects and it is expected to
be supported in all direct communication cast types, i.e., broadcast, groupcast, and unicast,
and for all coverage scenarios. Active and inactive patterns, durations and mechanisms,
together with active time alignment for UE communication with each other, and with the
Uu in the in-network case, will be studied and specified within the NR SL enh work item.
8.4 Enhanced V2X services work items

Two study items related to enhancing V2X services are present in the Rel-17 timeline:
• Study on architecture enhancements for 3GPP support of advanced V2X services –
Phase 2 (FS eV2XARC Ph2) [43] with findings being reported on TR 23.776 [44]
• Study on enhancements to application layer support for V2X services (FS eV2XAPP)
[45] with findings being reported in TR 23.764 [46].
59
______________________________________________________________________________________________________
The FS eV2XARC Ph2 study item aims to identify and analyze solutions that enhance
Rel-16 V2X specifications to support V2X services for pedestrian UEs in Rel-17, which
may have power and computation limitations with respect to the vehicular UEs supported
in Rel-16. In particular, the study examines how SL DRX, as studied in NR SL enh work
item for power saving (described in Section 8.3), can affect the V2X services that will run
over the NR PC5 and used by pedestrian UEs. To judge the feasibility of SL DRX for V2X
applications, the study will evaluate solutions to determine by what means DRX should
be configured, how to maintain QoS requirements while using it, and how to coordinate
simultaneous V2X services for maximizing power saving.
The FS eV2XAPP study item aims to analyze Rel-16 specifications related to the sup-
port of V2X in 5GS, and to identify the impact they have on the V2X application layer. The
focus is on identifying key issues and proposing solution recommendations to enable the
application layer support for V2X services in 5GS for both NR Uu and NR PC5 interfaces
usage. At the time of writing, 16 key issues have been identified in [44], and many of them
have implications related to the use of PC5 and direct communication for V2X services:
i) application layer support for PC5 QoS parameters provisioning, monitoring, and adap-
tation; ii) NR PC5 session handling for direct communication in support of tele-operated
driving; iii) NR PC5 provisioning in multi-operator V2X scenarios; iv) usage, selection,
provisioning, and configuration of the different cast modes (unicast, broadcast or group-
cast) by the application layer; and v) switching modes between NR Uu and NR PC5 for
V2V communications. The corresponding solutions to be used as baselines for normative
work are described in [44] as well, and the eV2XAPP work item is currently ongoing to
specify the features identified in the FS eV2XAPP study item. The current scope can be
found in [47].
8.5 Summary
The NR SL was defined in Rel-16 to support direct mode communication for several V2X
services. In Rel-17, NR SL specifications are meant to expand to support additional V2X
services (e.g., VRU) and ProSe features, including public safety direct communication. In
this section, we described Rel-17 study and work items related to the enhancement of the
NR SL to support those services (NR SL enh), the definition of NR ProSe and its evolu-
tion from LTE ProSe (FS 5G ProSe, 5G ProSe, FS NR SL Relay, and NR SL Relay) and
the impact and enhancements needed in the V2X architecture to support the new services
(FS eV2XARC Ph2) and to work within 5G NR systems (FS eV2XAPP and eV2XAPP).
60
______________________________________________________________________________________________________
9. Conclusion
In this study, we evaluated the 3GPP specifications enabling D2D communication in 5G

NR in order to understand the improvements and limitations to fully support direct mode
communication for first responders.
We started by looking at how resources are being allocated and services provisioned
for V2X in Rel-16. We focused on configurations related to the PC5 interface for SL
communication. Since V2X services can operate both on- and off-network, it provides
a good foundation for how ProSe services could be managed. Through the analysis of
the specifications, we highlighted possible issues that will need to be addressed especially
regarding the co-existence of V2X and non-V2X applications and how to deal with possible
conflicting resource allocations. In the Physical Layer section, we further described the
mechanisms used to transmit and receive packets over the SL channels. We highlighted
the flexible numerologies, detailed the various SL channels, and described new features
such as the feedback channel, enhanced coding, and MIMO, which are expected to increase
performance compared to LTE. The increase of capacity is discussed in the capacity section
where we conducted an initial analysis of NR SL for the unicast case. We derived formulas
for the various overheads and studied the impact of numerology, modulation, and blind
retransmissions. The results showed that there is a potential capacity increase for NR SL
compared to LTE SL with the actual gain dependent on the system configuration. In the
Quality of Service section, we described the mechanisms defined for the Uu interface as
well as the PC5 interface in NR. The analysis of the specifications show that there are
significant enhancements compared to LTE to handle QoS, including special PQI values for
public safety traffic. We also highlighted the MCR capability that defines the geographical
area where QoS requirements apply for groupcast communication, a feature that can be
beneficial for first responder communications using MCPTT group calls. In the LTE/NR
Coexistence section, we described DSS, which allows the sharing of resources between
LTE and NR, and showed that the current specifications need to be improved in order to take
into account SL communication. Finally, we addressed the continuous discussions by the
3GPP groups regarding discovery (architecture, protocol stack, message content, resource
pool), relays (types, layers, architecture, selection criteria) and other NR enhancements
(e.g., power consumption and battery limitations, NR Uu and NR PC5 interoperability). It
is expected to have those features added to Rel-17 after further review and approval.
The findings of this study confirm that NR SL can provide improved performance for
D2D communication which can benefit direct mode communication for first responders.
However, as discussed throughout this document, there are several limitations in the current
specifications that need to be addressed to support non-V2X applications. In addition,
there is a need to provide measurement tools to quantify the performance improvement
and determine the best set of parameters for the various situations facing first responders.
We plan to continue working with FirstNet and public safety agencies to prioritize further
research based on the findings shown in this document and recognize that other topics, such
as privacy, security, and range must also be considered.
61
______________________________________________________________________________________________________
Acknowledgments
The authors would like to thank Erickson Trejo-Reyes at FirstNet for providing additional
insights and perspectives related to challenges facing first responders and network opera-
tions.
References
[1] 3GPP (2020) NR; User Equipment (UE) radio transmission and reception; Part 2:
Range 2 Standalone (3rd Generation Partnership Project (3GPP)), Technical
Specification (TS) 38.101-2. Version 16.5.0 Available at https://portal.3gpp.org/
desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=3284.
[2] Lien S, Deng D, Lin C, Tsai H, Chen T, Guo C, Cheng S (2020) 3GPP NR sidelink
transmissions toward 5G V2X. IEEE Access 8:35368–35382. https://doi.org/10.1109/
ACCESS.2020.2973706
[3] 3GPP (2020) Vehicle-to-Everything (V2X) services in 5G System (5GS); Stage
3 (Release 16) (3rd Generation Partnership Project (3GPP)), Technical Specifica-
tion (TS) 24.587. Version 16.3.0 Available at https://portal.3gpp.org/desktopmodules/
Specifications/SpecificationDetails.aspx?specificationId=3640.
[4] 3GPP (2020) Vehicle-to-Everything (V2X) services in 5G System (5GS); User
Equipment (UE) policies; Stage 3 (Release 16) (3rd Generation Partnership
Project (3GPP)), Technical Specification (TS) 24.588. Version 16.4.0 Avail-
able at https://portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.
aspx?specificationId=3641.
[5] ISO (2021) Intelligent transport systems — cooperative systems — globally
unique identification. Available at https://standards.iso.org/iso/ts/17419/TS17419%
20Assigned%20Numbers/TS17419 ITS-AID AssignedNumbers.pdf.
[6] 3GPP (2020) NR; User Equipment (UE) radio transmission and reception; Part 1:
Range 1 Standalone (Release 16) (3rd Generation Partnership Project (3GPP)), Tech-
nical Specification (TS) 38.101-1. Version 16.5.0 Available at https://portal.3gpp.org/
[7] 3GPP (2020) NR; Radio Resource Control (RRC) protocol specification (Release 16)
(3rd Generation Partnership Project (3GPP)), Technical Specification (TS) 38.331.
Version 16.2.0 Available at https://portal.3gpp.org/desktopmodules/Specifications/
SpecificationDetails.aspx?specificationId=3197.
[8] 3GPP (2020) Overall description of Radio Access Network (RAN) aspects for
Vehicle-to-everything (V2X) based on LTE and NR (Release 16) (3rd Generation
Partnership Project (3GPP)), Technical Report (TR) 37.985. Version 16.0.0 Avail-
[9] 3GPP (2020) NR; Physical channels and modulation (Release 16) (3rd Gen-
eration Partnership Project (3GPP)), Technical Specification (TS) 38.211. Ver-
62
______________________________________________________________________________________________________
sion 16.3.0 Available at https://portal.3gpp.org/desktopmodules/Specifications/

[10] 3GPP (2020) NR; Physical layer procedures for control (Release 16) (3rd
Generation Partnership Project (3GPP)), Technical Specification (TS) 38.213.
[11] 3GPP (2020) NR; Multiplexing and channel coding (Release 16) (3rd Gen-
eration Partnership Project (3GPP)), Technical Specification (TS) 38.212. Ver-
sion 16.3.0 Available at https://portal.3gpp.org/desktopmodules/Specifications/
[12] 3GPP (2020) NR; Physical layer procedures for data (Release 16) (3rd
[13] 3GPP (2018) Revised WID: Enhancements on MIMO for NR (3rd Genera-
tion Partnership Project (3GPP)), Work Item Description RP-182863. Available
at https://portal.3gpp.org/ngppapp/CreateTDoc.aspx?mode=view&contributionUid=
RP-182863.
[14] Vook FW, Hillery WJ, Visotsky E, Tan J, Shao X, Enescu M (2018) System level
performance characteristics of sub-6ghz massive mimo deployments with the 3gpp
new radio. 2018 IEEE 88th Vehicular Technology Conference (VTC-Fall), , pp 1–5.
https://doi.org/10.1109/VTCFall.2018.8690560
[15] 3GPP (2019) Study on NR Vehicle-to-Everything (V2X) (3rd Generation Part-
nership Project (3GPP)), Technical Report (TR) 38.885. Version 16.0.0 Avail-
[16] 3GPP (2017) Study on new radio access technology physical layer aspects
(3rd Generation Partnership Project (3GPP)), Technical Report (TR) 38.802.
[17] 3GPP (2019) New WID: Further enhancements on MIMO for NR (3rd Gener-
ation Partnership Project (3GPP)), Work Item Description RP-193133. Available
RP-193133.
[18] 3GPP (2020) Evolved Universal Terrestrial Radio Access (E-UTRA); Radio Re-
source Control (RRC); Protocol specification (Release 16) (3rd Generation Partner-
ship Project (3GPP)), Technical Specification (TS) 36.331. Version 16.2.0 Avail-
[19] Cintron FJ (2018) Performance evaluation of LTE device-to-device out-of-coverage
communication with frequency hopping resource scheduling (National Institute of
Standards and Technology, Gaithersburg, MD), NIST IR 8220. https://doi.org/10.
63
______________________________________________________________________________________________________
6028/NIST.IR.8220
[20] 3GPP (2020) NR; User Equipment (UE) radio access capabilities (Release 16)
[21] 3GPP (2018) Discussion on UE peak data rate (3rd Generation Partnership
Project (3GPP)), TDoc R1-1803648. Available at https://portal.3gpp.org/ngppapp/
CreateTdoc.aspx?mode=view&contributionId=879805.
[22] 3GPP (2020) NR; NR and NG-RAN Overall Description; Stage 2 (Release 16)
[23] 3GPP (2020) System architecture for the 5G System (5GS); Stage 2 (Release 16)
[24] 3GPP (2020) Policy and charging control architecture (Release 16) (3rd
[25] 3GPP (2020) E-UTRA and NR; Service Data Adaptation Protocol (SDAP) specifica-
tion (Release 16) (3rd Generation Partnership Project (3GPP)), Technical Specifica-
tion (TS) 37.324. Version 16.2.0 Available at https://portal.3gpp.org/desktopmodules/
Specifications/SpecificationDetails.aspx?specificationId=3282.
[26] 3GPP (2020) Architecture enhancements for 5G System (5GS) to support Vehicle-to-
Everything (V2X) services (Release 16) (3rd Generation Partnership Project (3GPP)),
Technical Specification (TS) 23.287. Version 16.5.0 Available at https://portal.3gpp.
org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=3578.
[27] 3GPP (2020) Architecture enhancements for V2X services (Release 16) (3rd
[28] 3GPP (2020) LS on new PQI support for PC5 communication (3rd Generation Part-
nership Project (3GPP)), Liaison Statement R2-2008761. Available at https://portal.
3gpp.org/ngppapp/CreateTdoc.aspx?mode=view&contributionId=1157530.
[29] 3GPP (2021) NG-RAN; Xn application protocol (XnAP) (Release 16) (3rd
[30] MediaTek (2020) 5G NR and 4G LTE Coexistence; A Comprehensive
Deployment Guide to Dynamic Spectrum Sharing, . White paper. Avail-
able at https://d86o2zu8ugzlg.cloudfront.net/mediatek-craft/documents/
64
______________________________________________________________________________________________________
Dynamic-Spectrum-Sharing-WhitePaper-PDFDSSWP-031320.pdf.
[31] 3GPP (2020) Release 17 timeline agreed. Press Release. Available at https://www.
3gpp.org/news-events/2145-rel-17 newtimeline.
[32] 3GPP (2013) Feasibility study for Proximity Services (ProSe) (3rd Generation
[33] 3GPP (2019) Study on Network Controlled Interactive Services (3rd Generation
[34] 3GPP (2019) LS on D2D enabler for public safety communications in 5GS (3rd
Generation Partnership Project (3GPP)), Liaison Statement S6-190280. Available
S6-190280.
[35] 3GPP (2019) Revised SID: Study on System enhancement for Proximity based Ser-
vices in 5GS (3rd Generation Partnership Project (3GPP)), Work Item Description SP-
190443. Available at https://portal.3gpp.org/ngppapp/CreateTdoc.aspx?mode=view&
contributionUid=SP-190443.
[36] 3GPP (2021) Study on system enhancement for Proximity based Services (ProSe)
in the 5G System (5GS) (Release 17) (3rd Generation Partnership Project (3GPP)),
Technical Report (TR) 23.752. Version 17.0.0 Available at https://portal.3gpp.org/
[37] 3GPP (2021) Study on NR sidelink relay (Release 17) (3rd Generation Part-
[38] 3GPP (2021) Revised WID on System enhancement for Proximity based Ser-
vices in 5GS (3rd Generation Partnership Project (3GPP)), Work Item Description
SP-210272. Available at https://portal.3gpp.org/ngppapp/CreateTDoc.aspx?mode=
view&contributionUid=SP-210272.
[39] 3GPP (2021) New WID on NR Sidelink Relay (3rd Generation Partnership Project
(3GPP)), Work Item Description RP-210904. Available at https://portal.3gpp.org/
ngppapp/CreateTDoc.aspx?mode=view&contributionUid=RP-210904.
[40] 3GPP (2021) Proximity based Services (ProSe) in the 5G System (5GS) Release 17
[41] 3GPP (2020) WID revision: NR Sidelink enhancement (3rd Generation Partnership
Project (3GPP)), Work Item Description RP-202846. Available at https://portal.3gpp.
org/ngppapp/CreateTDoc.aspx?mode=view&contributionUid=RP-202846.
[42] 3GPP (2019) Summary of email discussion on Rel-17 sidelink enhancement (3rd
65
______________________________________________________________________________________________________
Generation Partnership Project (3GPP)), Tdoc RP-192745. Available at https://www.

3gpp.org/ftp/tsg ran/TSG RAN/TSGR 86/Docs/RP-192745.zip.
[43] 3GPP (2019) New SID: Study on architecture enhancements for 3GPP support of
advanced V2X services - Phase 2 (3rd Generation Partnership Project (3GPP)),
Work Item Description SP-190631. Available at https://portal.3gpp.org/ngppapp/
CreateTdoc.aspx?mode=view&contributionUid=SP-190631.
[44] 3GPP (2020) Study on architecture enhancements for 3GPP support of ad-
vanced Vehicle-to-Everything (V2X) services; Phase 2 (3rd Generation Part-
[45] 3GPP (2020) Revised SID on enhancements to application layer support for V2X
services (3rd Generation Partnership Project (3GPP)), Work Item Description SP-
200110. Available at https://portal.3gpp.org/ngppapp/CreateTdoc.aspx?mode=view&
contributionUid=SP-200110.
[46] 3GPP (2020) Study on enhancements to application layer support for V2X ser-
vices; (3rd Generation Partnership Project (3GPP)), Technical Report (TR) 23.764.
[47] 3GPP (2020) New WID for enhanced application layer support for V2X services
(3rd Generation Partnership Project (3GPP)), Work Item Description SP-200831.
Available at https://www.3gpp.org/ftp/tsg sa/TSG SA/TSGs 89E Electronic/Docs/
SP-200831.zip.
66

5G-NR Direct Mode Communication

Uploaded by

Copyright:

Available Formats

5G-NR Direct Mode Communication

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

5G-NR Direct Mode Communication

Uploaded by

Copyright:

Available Formats

NISTIR 8372

Study of 5G New Radio (NR) Support

This publication is available free of charge from:

This publication is available free of charge from:

U.S. Department of Commerce

National Institute of Standards and Technology

National Institute of Standards and Technology

This publication is available free of charge from:

6.5 Minimum Communication Range 46

7.2 DSS Overview 49

Table 4 256 QAM MCS Table 28

Fig. 13 Capacity - NR vs LTE. 38

Uu The Radio interface between UTRAN and the User Equipment.

3GPP Third Generation Partnership Project

5GC 5G Core Network

5QI 5G QoS Identifier

AGC Automatic Gain Control

AMBR Aggregate Maximum Bit Rate

ARFCN Absolute Radio Frequency Channel Number

ARIB Association of Radio Industries and Businesses

ARP Address Resolution Protocol

ARP Allocation and Retention Priority

ATIS Alliance for Telecommunications Industry Solutions

BWP Bandwidth Part

CCSA China Communications Standards Association

CQI Channel Quality Indication

CRB Common Resource Block

CRC Cyclic Redundancy Check

CRS Cell Reference Signal

CSI Channel State Information

CSI-RS Channel-State Information Reference Signal

DM-RS Demodulation Reference Signal

DRB Data Radio Bearer

DRX Discontinuous Reception

DSS Dynamic Spectrum Sharing

E-UTRA Evolved Universal Terrestrial Radio Access

eMBB enhanced Mobile BroadBand

ETSI European Telecommunications Standards Institute

FDD Frequency Division Multiplexing

FR2 Frequency Range 2

GBR Guaranteed Bit Rate

GFBR Guaranteed Flow Bit Rate

HARQ Hybrid Automatic Repeat Request

IoT Internet of Things

IPv4 Internet Protocol version 4

LCID Logical Channel ID

LCP Logical Channel Prioritization

LDPC Low Density Parity Check

LTE Long Term Evolution

MAC Medium Access Control

MBSFN Multicast-Broadcast Single-Frequency Network

MCPTT Mission Critical Push-to-Talk

MCR Minimum Communication Range

MCS Modulation and Coding Scheme

MFBR Maximum Flow Bit Rate

MIMO Multiple-Input and Multiple-Output

mMTC massive Machine Type Communication

mmWave Millimeter Wave