Exploring NR-V2X Dynamic Grant Limitations for Aperiodic Traffic

Brian McCarthy and Aisling O’Driscoll

School of Computer Science and Information Technology, University College Cork, Cork, Ireland

Keywords:

Cellular V2X (C-V2X), LTE-V, Sidelink, NR-V2X, Aperiodic, CAM, DENM, Autonomous Resource

Selection, 4G, 5G, SPS.

Abstract:

The recent 3GPP NR-V2X standard (Rel. 16) has largely built upon its precursor Cellular-V2X (Rels. 14 &

15) but has introduced new approaches for dealing with application traffic exhibiting aperiodic arrival rates

algorithm and evaluate the performance of alternative dedicated scheduling mechanisms specifically designed

for aperiodicity. This paper highlights that the level of aperiodicity exhibited by the application model greatly

impacts scheduling performance, both for the default SB-SPS and dedicated approaches. As such we conclude

that a novel aperiodic scheduling mechanism must be devised, or more promisingly, an approach to enable ap-

plication traffic to mimic periodic characteristics allowing it to co-exist with the existing scheduling approach.

1 INTRODUCTION

The Third Generation Partnership Project (3GPP) has

specified mobile standards to support V2X (vehicle to

everything) communications as an alternative to the

existing wireless standards based on IEEE 802.11p.

facilitate V2V communications can be selected and

managed by the cellular network (Mode 3) or selected

autonomously by the vehicles (Mode 4) using the dis-

tributed scheduling algorithm, Sensing Based Semi-

Persistent Scheduling (SB-SPS). The specification in

Rel. 14 and Rel. 15, known as Cellular-V2X (C-V2X)

has acted as the basis for the New Radio (NR) speci-

fication in 3GPP Release 16 (3GPP, 2020) with Mode

4 forming the basis for NR-V2X Mode 2.

The original focus of the 3GPP Rel. 14 standard

sumption does not typically hold true as ETSI mes-

(e-V2X) applications for connected and automated

vehicles including platooning, extended sensors, ad-

vanced and remote driving that will need to support

aperiodic application traffic patterns.

However, the transmission of aperiodic applica-

tion traffic has a large impact on the operation of

the SB-SPS MAC scheduling mechanism. It was de-

signed to assume periodic arrival rates thereby en-

abling accurate prediction of free radio resources.

Aperiodic traffic which by it’s nature does not have a

and additionally includes some changes to the SB-

SPS historical sensing mechanism. A detailed quan-

titative study is however missing from literature to

evaluate if this can adequately address high levels

of aperiodicity. As such an open research challenge

McCarthy, B. and O’Driscoll, A.

Exploring NR-V2X Dynamic Grant Limitations for Aperiodic Traffic.

DOI: 10.5220/0010937500003191

In Proceedings of the 8th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2022), pages 17-25

ISBN: 978-989-758-573-9; ISSN: 2184-495X

C-V2X MAC scheduling algorithm to more reliably

support such aperiodicity, ideally in co-existence with

the current scheduling algorithm.

This paper provides an in-depth study to deter-

mine the precise reasons and conditions under which

the C-V2X and NR-V2X SB-SPS algorithms exhibit

significant performance degradation when supporting

aperiodic application traffic across a variety of vehic-

ular densities. Next, the paper investigates the perfor-

mance limitations of NR-V2X dynamic grant, high-

paper evaluates the viability of some alternative dedi-

cated scheduling approaches.

The main contributions of this paper are:

of the C-V2X and NR-V2X SB-SPS algorithms

when faced with aperiodic application traffic char-

acteristics.

V2X dynamic grant mechanism in order to gain

deeper insight into how aperiodic application traf-

lows. Section 2 describes the related literature with

Section 3 describing the operation of the PHY &

MAC layers of C-V2X and NR-V2X. Subsection 4.1

provides a detailed simulation study of the challenges

faced by SB-SPS in supporting aperiodic application

traffic for an ETSI CAM application model. Subsec-

tion 4.2 highlights the shortcomings of the NR-V2X

dynamic grant mechanism in supporting fully aperi-

odic services, with Subsection 4.3 highlighting the

impact of considering hybrid periodic vs aperiodic ap-

beneficial future solution.

Existing literature has not quantitatively evaluated

whether Rel. 16 NR-V2X, including dynamic grant,

can adequately support aperiodic traffic, with evalua-

tions largely focusing on the Rel. 14 C-V2X SB-SPS

standard. Such a study is vital as ETSI CAM and

DENM transmissions are fully aperiodic and these

will form the basis for future e-V2X services to sup-

port increased autonomy.

Existing studies have stated that SB-SPS has

els, as well as variable packet sizes. They further

study adapting the resource reservation interval by

balancing re-selections against occurrences of wasted

resources using three strategies. Notably, the authors

suggest that the SB-SPS mechanism is fundamentally

counter-productive for aperiodic traffic and highlight

that further dedicated schemes are required. They do

not consider grant breaking or dynamic grant in their

evaluation.

Similarly, Bartoletti et. al (Bartoletti et al., 2021)

sequent collision being maintained for the duration

of the grant, thereby impacting successful receipt of

the DENM. However, the authors consider a one time

reservation to mimic the Dynamic Grant mechanism

but this operates using the C-V2X Mode 4 SB-SPS

standard i.e. includes RSSI filtering and RSRP aver-

aging. As such it is does accurately model NR-V2X

Mode 2.

Finally, Lusvarghi et. al (Lusvarghi and Merani,

2020) investigate a hybrid application model using

1000 Bytes, due to high subchannel occupation.

ing adjacent or non-adjacent schemes. In the adja-

cent scheme, the PSCCH and PSSCH are transmit-

ted in contiguous RBs. In the non-adjacent scheme,

the PSCCH and PSSCH are transmitted in different

RB pools. In terms of occupancy, the PSCCH re-

quires 2 RBs, while the number of RBs required by

the PSSCH is variable and depends on the size of the

TB. It is worth noting that the PSSCH and PSCCH

are always transmitted in the same subframe indepen-

dently of the transmission scheme.

number of subchannels, the number of recurrent slots

for which the subchannels will be reserved, and the

periodicity between transmissions. If a grant has al-

ready been created at the time a packet is received

est levels of interference are considered for selection.

The remaining CSRs are passed to the MAC layer,

where a single CSR is selected at random to reduce

the probability of multiple vehicles choosing the same

CSR. The CSR is selected for a number of recurrent

slots defined by the RRC, whose value is decreased

by one after each transmission.

This introduces a multi-stage SCI format to enable

these communication patterns. More information re-

lating to NR-V2X PHY changes can be found in (Gar-

cia et al., 2021).

At the MAC layer, we highlight only the changes

that are significant with respect to impacting SB-SPS

scheduling performance. The first relates to the RSRP

filtering stage; rather than using an average RSRP

value for the CSR in the sensing history as is the case

in Mode 4, Mode 2 instead uses only the RSRP of the

5G NR-V2X to deal with aperiodic traffic patterns. It

operates by using the same underlying SB-SPS algo-

rithm but assuming a single transmission. The SCI

message for this transmission will highlight that the

resource will not be maintained in the future. This

does not apply when retransmissions are enabled, in

which case a grant is maintained for said retransmis-

sions. We do not consider retransmissions in this

study.

In this section, the causative effects of declined SB-

SPS performance when faced with aperiodic applica-

tion traffic are quantitatively explored. We next in-

vestigate the limitations of the latest standardised ap-

proach to address this i.e. NR-V2X dynamic grants,

when application traffic is highly aperiodic. Finally,

we show that application traffic with hybrid periodic

and aperiodic characteristics can perform adequately,

motivating the need for a dedicated scheduling mech-

Table 1: Simulation Parameters.

(6/12 total)

Vehicle Speed 70km/h

Vehicle Mobility SUMO (step-length

= 1ms)

Carrier frequency 5.9 GHz

Channel bandwidth 10 MHz

No. subchannels 3

Subchannel size 16 Resource Blocks

Noise figure 9 dB

Shadowing variance 3 dB

Results are based on the OpenCV2X

(McCarthy and O’Driscoll, 2019) in conjunction with

SUMO

et al., 2015). Table 1 describes the simulation pa-

rameters used. The vehicular scenario was chosen

experiments. Application traffic is modelled as fol-

lows:

This rate is deliberately chosen to allow for com-

parable analysis with the ETSI model with respect

to channel load. Half of all vehicles transmit ev-

ery 200ms with the remainder transmitting every

300ms, selection of 200ms or 300ms rate is done

dynamics i.e. deviation in heading (>4°), posi-

tion (>4m) and speed (>0.5m/s) or at 1s intervals

if these conditions are not satisfied.

We now quantitatively explore why SB-SPS exhibits

declined performance when application traffic follows

aperiodic arrival rates. Both C-V2X (Rel. 14) and

NR-V2X (Rel. 16) are considered. The distinction

between these, as described fully in Section 3.2, is

source use as a consequence of inaccurate knowledge,

with the reasons differing depending on whether grant

breaking is enabled or not. This leads to a rise in colli-

sions and decreased packet delivery rates. Subsequent

results prove this to be the case.

C-V2X (Periodic) 64

NR-V2X (Periodic) 20

C-V2X. A grant break occurs as no TB exists to send

Figure 1: SB-SPS performance of NR-V2X (Mode 2)

and C-V2X (Mode 4) for periodic and ETSI application

models.β = 0.12 veh/m. With and without grant breaking.

in a scheduled slot, thus the resource goes unused as

SB-SPS ends the grant when a missed transmission

occurs. When vehicles frequently break their grants,

they must schedule new resources. Each time re-

sources are reserved there is a probability that a neigh-

bouring vehicle will choose the same set of resources,

resources, which can be exacerbated further as vehi-

cle density increases. This has a severe impact on C-

V2X Mode 4 due to it’s CSR selection mechanism

incorporating RSSI filtering. In contrast, this stage is

removed for NR-V2X which results in a larger CSR

2 which represents the number of instances where

grants collide i.e. where neighbouring vehicles select

the same subchannel in the same subframe for trans-

missions. It can be seen that C-V2X incurs a high

noted as No-GB). The grant is maintained irrespective

of a missed transmission. It can be seen that while

this greatly improves performance for C-V2X when

compared to grant breaking, it still comes at a cost.

Resources are reserved that may not be used. Further-

more these reserved but unused resources introduce

inconsistencies in the CSR selection process where a

vehicle does not send an SCI due to a missed trans-

mission and neighbouring vehicles mistakenly believe

breaking is disabled is worse than the performance of

C-V2X. This indicates that when the grant is main-

tained, RSSI filtering can be a positive feature by dis-

couraging the selection of resources that exhibit high

levels of interference. Furthermore, RSRP averaging

can result in more accurate sensing history.

Irrespective of whether grant breaking is enabled

or not, it has been shown that the performance of SB-

SPS for NR-V2X and C-V2X becomes compromised

due to inefficient resource scheduling. This leads to

Figure 2: Packet collisions for periodic and ETSI applica-

ing, makes use of two selection windows. It works

by sending a reservation signal in the first selec-

tion window, which reserves resources in the sec-

on which to send the reservation signal is chosen

randomly, discounting resources reserved for SCI

and TB pairs. Until the reservation signal is sent,

the vehicle continues to listen in case the reserva-

tion slot becomes reserved by another vehicle. At

the time of sending the reservation signal, a future

free resource is selected if it has not already been

TB pairs.

decremented by the number of free subchannels.

Once the counter reaches 0, a free subchannel is

chosen at random in the next available subframe

where the SCI and TB pair will be transmitted.

It is clear from Fig. 5 that counter demonstrates

negligible improvement over NR-V2X dynamic grant

as it is essentially a random scheduling mechanism.

the reservation signals would become a cause of col-

lisions. While STR has limited for applicability for

(a) NR-V2X PDR when for SB-SPS, dynamic grant & ded-

icated aperiodic scheduling.

(b) NR-V2X CBR for SB-SPS, dynamic grant & dedicated

This paper provides an in-depth study on the limi-

tations of both C-V2X and NR-V2X SB-SPS algo-

rithms in reliably supporting application traffic with

aperiodic characteristics. It further provides im-

portant insights into the limitations of the NR-V2X

dynamic grant scheduling mechanism, and is the

first paper to do so while considering the Rel. 16

MAC with removed RSSI filtering and RSRP aver-

aging. Ultimately it finds that neither approach is ad-

equate for dealing with aperiodic traffic, either per-

also been shown to be inadequate resembling random

scheduling. It is shown that to perform adequately,

dynamic grant must have a hybrid of periodic and ape-

riodic traffic to allow for accurate resource filtering.

It should be noted that dynamic grant performance

may improve with retransmissions although overhead

may be prohibitive. Alternative dedicated aperiodic

scheduling mechanisms were also explored but have

their own drawbacks such as increased channel usage,

which can limit their applicability.

munications and the current standards are not ade-

quate to effectively manage these patterns. It is our

conclusion that new approaches will need to be im-

with existing SB-SPS scheduling or as a standalone

approach. Alternatively, and more promisingly, an

approach that provides higher layer traffic shaping

or more intelligent dynamic SB-SPS RRI selection

would ensure that aperiodic traffic is scheduled sim-

ilarly to periodic traffic, thus eliminating wasted re-

sources and inconsistent sensing history. The authors

in (Yoon and Kim, 2021) sought to address this, how-

ever the solution is based on an empirical data-set

and cannot be generalised to predict application layer

fic that forms the backbone of vehicular communica-

tions to mimic periodic arrival rates. Adopting such a

mechanism would allow the NR-V2X dynamic grant

to only be used for low frequency aperiodic traffic

such as DENMs. This was shown in Fig. 4a to per-

form reliably. An investigation into the feasibility and

performance of such a mechanism will form part of

future work.

This publication has emanated from research [ con-

ducted with the financial support of Science Foun-

3GPP (Dec. 2020). Release 16 description; summary of

3GPP (Jul. 2016). Study on lte-based v2x services (v14.0.0,

3GPP (Sep. 2019). Release 15 description; summary of rel-

Bartoletti, S., Masini, B. M., Martinez, V., Sarris, I., and

ETSI (April. 2019b). Etsi en 302 637-3; intelligent transport

ETSI (April. 2019c). Etsi tr 103 562; intelligent transport

ETSI (Jan. 2019a). En 302 637-2 intelligent transport sys-

Gonzalez-Mart

´

ın, M., Sepulcre, M., Molina-Masegosa, R.,

LG-Electronics (2019). Discussion on resource allocation

Lopez, P. A., Behrisch, M., Bieker-Walz, L., Erdmann, J.,

to-Everything. IEEE Access, 8:207076–207088.

McCarthy, B. and O’Driscoll, A. (2019). Opencv2x mode

4: A simulation extension for cellular vehicular com-

munication networks. In 2019 IEEE 24th Interna-

tional Workshop on Computer Aided Modeling and

Design of Communication Links and Networks (CA-

MAD), pages 1–6.

Molina-Masegosa, R., Gozalvez, J., and Sepulcre, M.

(2020). Comparison of ieee 802.11p and lte-v2x: An

evaluation with periodic and aperiodic messages of

constant and variable size. IEEE Access, 8:121526–

456.

Romeo, F., Campolo, C., Molinaro, A., and Berthet, A. O.

(2020a). Asynchronous traffic on the sidelink of 5g

v2x. In 2020 IEEE International Conference on Com-

munications Workshops (ICC Workshops), pages 1–6.

Romeo, F., Campolo, C., Molinaro, A., and Berthet, A. O.

(2020b). Denm repetitions to enhance reliability of the

autonomous mode in nr v2x sidelink. In 2020 IEEE

91st Vehicular Technology Conference (VTC2020-

Spring), pages 1–5.

Todisco, V., Bartoletti, S., Campolo, C., Molinaro, A.,

Berthet, A. O., and Bazzi, A. (2021). Performance

Analysis of Sidelink 5G-V2X Mode 2 through an