future internet

Article
Survey and Perspectives of Vehicular Wi-Fi versus
Sidelink Cellular-V2X in the 5G Era
Alessandro Bazzi * , Giammarco Cecchini , Michele Menarini, Barbara M. Masini and
Alberto Zanella
National Research Council of Italy (CNR), Institute of Electronics, Computer and Telecommunication
Engineering (IEIIT), Viale Risorgimento, 2, 40136 Bologna, Italy; [email protected] (G.C.);
[email protected] (M.M.); [email protected] (B.M.M.); [email protected] (A.Z.)
* Correspondence: [email protected]




Received: 30 April 2019; Accepted: 25 May 2019; Published: 29 May 2019


Abstract: The revolution of cooperative connected and automated vehicles is about to begin and
a key milestone is the introduction of short range wireless communications between cars. Given
the tremendous expected market growth, two different technologies have been standardized by
international companies and consortia, namely IEEE 802.11p, out for nearly a decade, and short
range cellular-vehicle-to-anything (C-V2X), of recent definition. In both cases, evolutions are under
discussion. The former is only decentralized and based on a sensing before transmitting access,
while the latter is based on orthogonal resources that can be also managed by an infrastructure.
Although studies have been conducted to highlight advantages and drawbacks of both, doubts
still remain. In this work, with a reference to the literature and the aid of large scale simulations in
realistic urban and highway scenarios, we provide an insight in such a comparison, also trying to
isolate the contribution of the physical and medium access control layers.

Keywords: vehicle-to-anything; connected and autonomous vehicles; vehicular networks; IEEE 802.11p;
IEEE 802.11bd; DSRC; ITS-G5; LTE-V2X; cellular-V2X; sidelink; PC5; 5G; cooperative awareness

1. Introduction
Wireless communications and the wide distribution of the Internet have completely changed
society over the last twenty years. The rapid evolution of communication technology has made it
possible to implement the paradigm of everyone communicating everywhere and any time. The next step,
which has already begun but is still in its infancy, is the interconnection and interoperability of robots,
which will have a drastic impact in many fields, including industry, health, and mobility.
With an unpredictable period of transition, someday vehicles will be all autonomous and
connected, with the promise of no more deaths on the road, of more efficiency of traffic flows, and of
more comfort for all passengers. One of the main first steps along this path is the transformation of
cars and trucks from isolated road-users, adhering common rules, to communicating, coordinated,
and cooperative parts of an ecosystem. Such transformation passes through the capability of vehicles
to connect to each other and to the rest of the world, and a key role is played by wireless short-range
technologies. In fact, short-range is required to allow scalability of the service as well as to avoid risks
of areas where the connectivity is not available.
From an industrial point of view, this represents an enormous opportunity, given that more than 90
million of vehicles are produced every year [1] and that the global connected car market is projected to
grow tremendously in the next years (estimated in [2] to exceed $225,000.0 million by 2025). In this scenario,
two technologies have been designed to be used on board of the vehicles for short-range connectivity and a
tough debate is ongoing to demonstrate that either can be preferred to the other. The point is that vehicles

Future Internet 2019, 11, 122; doi:10.3390/fi11060122 www.mdpi.com/journal/futureinternet

Future Internet 2019, 11, 122 2 of 20

need to use the same technology to be able to communicate, thus if vehicles from different manufacturers
are equipped or use only one of them without agreements, it would have detrimental effects.
The two technologies, largely discussed in the following, are the Wi-Fi solution, often referred to as
IEEE 802.11p from the name of the first standard designed to this scope, and the cellular solution, named
long-term-evolution-vehicle-to-anything (LTE-V2X) as part of the fourth generation of Third Generation
Partnership Project (3GPP) standards and under the broader umbrella of the so-called cellular-V2X (C-V2X).
Whereas IEEE 802.11p has a relatively long history, with very large field trials performed worldwide,
LTE-V2X is very recent. This gives to the latter solution a potential advantage in the improved physical (PHY)
layer design, but makes its effective reliability and scalability still to be completely verified.
Several works have recently discussed the two technologies from various perspectives and a
number of performance comparisons have been provided both in white [3–6] and peer-reviewed
papers [7–15]. A common conclusion is normally that, given a target performance, the more advanced
PHY layer of LTE-V2X allows for reaching a longer distance. However, as deeply discussed in
the following, there are specific scenarios and settings where IEEE 802.11p provides similar or
even better results. It is also worth noting that most papers showing a comparison assume given
modulation and coding schemes (MCSs), which are not optimized and often penalize IEEE 802.11p.
In addition, no specific discussion has been done about the role played by the medium access
control (MAC) protocol.
In this work, we aim at revising the performance comparison between IEEE 802.11p and sidelink
LTE-V2X, first summarizing the related work and then providing original results in various realistic
scenarios and optimizing the adopted MCSs. More specifically, the rest of the paper is organized into
the following two main parts:

• The first part is related to the state of the art and corresponds to Section 2: specifically, we provide
an overview of the two technologies and their evolutions in Sections 2.1 and 2.2, followed by a
discussion of the related work on their comparison in Section 2.3;
• The second part reports original results obtained through large-scale simulations; the simulation
tool and the adopted settings are detailed in Section 3 and the outputs are analysed in Section 4.

After these parts, we dedicate Section 5 to a summary of the main findings that derive from both
literature and new simulations, followed by final considerations in Section 6.

2. Overview of the Technologies and Comparative Discussion

After the studies on short-range communications between vehicles carried out in the eighties
and nineties, probably the first milestone was the allocation of dedicated bandwidth in the US in
1999. In particular, seven channels of 10 MHz each were reserved around 5.9 GHz for what was
called dedicated short range communication (DSRC) (for this reason, DSRC is often used to denote
IEEE 802.11p and the related American standards). A similar allocation was issued in Europe in
2008 (similar frequencies, with only five channels [16]). In both cases, the allocations were implicitly
addressing IEEE 802.11p.
Since the first allocations in the US, for more than one decade all remained confined to research
efforts and there was no clear program for a real deployment. Everything has instead accelerated in the
latest years, after large scale tests in the US and the so-called European Telecommunications Standards
Institute (ETSI) Plugtests in Europe have proved the feasibility of massive use of short-range V2X. As a
consequence, various governments are working at the opportunity to mandate or at least encourage
the installation on new vehicles.
In the meantime, 3GPP have recognized the relevance of this vertical and the urgency of specific
effort inside the cellular ecosystem. For this reason, in 2016, it created the so-called C-V2X within
LTE Release 14, frozen in 2017, and included a short-range interface that can be used also outside the
cellular coverage. The new short-range LTE-V2X interface clearly appears as an alternative to IEEE
802.11p. The main events are also summarized in Table 1, where it is possible to observe at a glance
Future Internet 2019, 11, 122 3 of 20

that, while the IEEE 802.11 working group spent about 15 years to make commercial products available
on the market, 3GPP has taken a leap forward and promises to have LTE-V2X devices soon available
(despite the decidedly lower series of experimentations).
In this section, the main characteristics of the two technologies, the planned evolutions, and a
comparative discussion based on related work are reported.

Table 1. Chronology of the main facts in the US and Europe.

1999 Frequencies are allocated in the US General

2004 IEEE 802.11p Task Group is formed IEEE 802.11p
2008 Frequencies are allocated in Europe General
2010 IEEE 802.11p is approved IEEE 802.11p
2012 The Ann Arbor experiment starts with thousands of devices IEEE 802.11p
2016, October The first part of 3GPP Release 14 is published C-V2X
2016, November The Livorno Plug test: ITS-G5 is declared ready IEEE 802.11p
2016, November The European Commission publishes the “Strategy” General
2017, January The proposed rule is issued in the US to mandate short-range General
2017, March 3GPP Release 14 is frozen C-V2X
2018, May IEEE 802.11 Next Generation V2X is announced (IEEE 802.11bd) IEEE 802.11p
2019, March 3GPP Release 15 is frozen C-V2X

2.1. IEEE 802.11p

The first Wi-Fi based standard specifically designed for vehicular communications was approved
under the name IEEE 802.11p in 2010. IEEE 802.11p was later included in the IEEE 802.11-2012, now
superseded by the IEEE 802.11-2016. The peculiarities of the vehicular based technology with respect
to the other associated standards are mainly a smaller bandwidth of 10 MHz (less noise and longer
cyclic prefix) and the possibility to work without the need to join a basic service set (BSS), which is
called outside of the context of a BSS (OCB) mode.
IEEE 802.11p defines the PHY and MAC layer protocols. Above IEEE 802.11p, a number of other
standards have been defined, creating two different pillars in the US and in Europe [17]. In the US,
IEEE 1609 standards focus on the protocols below the application layer, which is instead defined by
the Society of Automotive Engineers (SAE). In Europe, IEEE 802.11p was adopted by ETSI under the
name ITS-G5, together with a large number of other documents dealing with all layers above it.
At the PHY layer, IEEE 802.11p adopts orthogonal frequency division multiplexing (OFDM)
with 48 useful and four pilot subcarriers. The OFDM symbol lasts 8 µs and the subcarrier spacing
is 156.25 kHz, bringing to a raw bandwidth of 10 MHz. The packet is preceded by a preamble
for synchronization and training. Eight MCSs are possible, with modulations going from binary
phase shift keying (BPSK) to 16-quadrature amplitude modulation (QAM) and (possibly punctured)
convolutional coding. The eight combinations are reported in Table 2.
At the MAC layer, carrier sensing multiple access with collision avoidance (CSMA/CA) is
adopted. A node needing to transmit senses the medium to check if it is idle, and a mechanism
based on random backoff is performed to reduce the probability of collisions. In the vehicular scenario,
the acknowledgement is not foreseen and thus the exponential backoff introduced in other versions
for retransmissions does not apply.
The CSMA/CA mechanism introduces significant overhead, especially with short packets and
high level modulations. In addition, it is known that communications are severely affected by collisions
when the channel occupation raises 50–60% [18], thus mechanisms to avoid overloads are necessary.
This indeed brought both SAE through SAE J2945/1 and ETSI through decentralized congestion
control (DCC) to add algorithms specifically addressing congestion control.
Future Internet 2019, 11, 122 4 of 20

2.1.1. Next Generation: IEEE 802.11bd

In May 2018, IEEE and the IEEE Standards Association announced the formation of a new study
group focused on the evolution of 802.11 technology for next-generation V2X communications, named
IEEE 802.11 Next Generation V2X (NGV) and now preparing the amendment IEEE 802.11bd.
This activity was also strongly encouraged by the CAR 2 CAR Communication Consortium, which
in its early studies towards an ITS-G5 Release 2, noted that IEEE 802.11ac is already the state of the art
for Wi-Fi technologies. Therefore, the idea was to take this version of the standard as a starting point
and adapt it to high mobility and OCB mode, while guaranteeing backward compatibility with legacy
devices. Emphasis was posed, in particular, to multiple input multiple output (MIMO) techniques,
improved channel coding, and better pilots placing (specifically designed for the vehicular wireless
channel). Expected gains have been estimated in about 6 dB by low density parity check (LDPC)
(advanced channel coding) and 2–3 dB by space time block codes (STBC) (MIMO space diversity).
The IEEE study group is presently in its initial stage [19], with the goal to investigate the
usage of more advanced PHY technologies in amendments after 802.11p, in order to enhance the
throughput (one transmission mode that at least doubles the throughput) and the transmission
range (one transmission mode providing at least 3 dB lower sensitivity). In particular, the following
technology candidates have been listed [19,20]:

• OFDM numerology re-design: definition of new optimized tone spacing and guard interval
duration to better cope with 5.9 GHz frequencies and high-speed mobility;
• LDPC codes to improve coding effectiveness;
• MIMO diversity through STBC codes or cyclic shift diversity (CSD);
• Addition of midambles (control sequences in the middle of the packet) to improve Doppler recovery;
• Dual carrier modulation (DCM) and 20 MHz channels.

It can be noted that all mentioned modifications are at the PHY layer.
In addition, co-existence and interoperability with legacy IEEE 802.11p is remarked as mandatory:
any NGV device must communicate with legacy devices, granting fairness in channel access.

2.2. Sidelink LTE-V2X

Short-range LTE-V2X was defined in Release 14, based on the device-to-device (D2D) of Release 12.
It is also called sidelink (opposed to downlink/uplink) and its communication interface is named PC5.
At the PHY and MAC layers, sidelink LTE-V2X (hereafter simply LTE-V2X) relies on the same
single carrier frequency division multiple access (SC-FDMA) used by LTE uplink: in the frequency
domain, the subcarrier spacing is fixed to 15 kHz and subcarriers are used in groups of 12 (i.e., 180 kHz);
in the time domain, 14 symbols form a subframe of 1 ms, also called transmission time interval (TTI),
and include nine data symbols, four demodulation reference signal (DMRS) symbols, and one empty
symbol for Tx-Rx switch and timing adjustment. LTE-V2X has a high number of MCSs, with 4-QAM
and 16-QAM modulations and an almost continuous coding rate. As an example, the MCSs, with
modulation and coding rate, are listed in Table 3 referring to packets of 300 bytes.
The minimum resource for the allocation of messages in LTE-V2X is the subchannel in the
frequency domain, corresponding to a multiple of the 12 subcarriers groups, and the TTI in the time
domain. One packet normally occupies one or more subchannels in one TTI. In principle, resources are
orthogonal and two transmitters either fully or negligibly interfere, except for in-band emission (IBE)
due to the use of different frequencies in the same TTI.
Regarding the upper layers, LTE-V2X is stated compatible with ETSI-Intelligent Transportation
System (ITS), ISO, SAE and IEEE standards.
Given the time-frequency structure with orthogonal resources, the resource allocation algorithm
plays a crucial role to optimize the spatial reuse. Since LTE-V2X was mainly designed to enable the
cooperative awareness service, which is the transmission of periodic messages by each vehicle to inform
Future Internet 2019, 11, 122 5 of 20

about its status and movements, the resource allocation is normally associated with a semi-persistent
scheduling (SPS) mechanism. Two different approaches are defined by 3GPP for the reservation of
resources, namely Mode 3 and Mode 4, depending on the entity in charge of the allocation.
In Mode 3, also known as scheduled or controlled, the resources to be used by each vehicle
for short-range communications are defined by the network. This strategy, on the one hand, clearly
requires that the vehicles are within the coverage of some eNodeB and that some information is
exchanged between nodes and controllers to allow efficient transmissions. On the other hand, it can
take advantage of large processing capabilities and wider view of the state of the network, thus is
expected to enable improved performance. Mode 3 algorithms are not defined in the specifications
and their implementation is left to the operators. Software-defined networking (SDN) is also expected
to play a relevant role in this case [21].
In Mode 4, also known as autonomous or out-of-coverage, each node selects the resources to use
based on a sensing procedure and an SPS mechanism. The algorithm is defined in details by 3GPP
(see also [7,22]) and can be summarized as follows. At the MAC layer, a resource is randomly selected
within the set received from the PHY and maintained (i.e., the same subchannels at periodic TTIs, with
a period equal to the periodicity of the messages) for a period normally taken within 5 to 15 times the
packet generation interval; after that time, the same resource is kept for another random interval with
probability pk (set by the operator within 0 and 0.8) and changed otherwise.
At the PHY layer, the resources are monitored within a window of 1 s, comparing the received
power measurements with a given threshold and reading sidelink control information (SCI) messages
that advertise future reservations. Among the resources that are assumed not used in the next packet
interval (or a portion of it), the 20% less interfered are selected and passed to the MAC. If less than 20%
resources are estimated free, the mentioned threshold is increased by 3 dB and the process is repeated.
The algorithm presents several parameters, although it has been shown in [22,23] that its performance
is mostly determined by the setting of pk. The choice of pk in most scenarios leads to a trade-off between
reliability and latency: on the one hand, a low value makes reservations changing often and thus the sensing
procedure fail to correctly estimate the occupancy in the next future; on the other, a high value makes
possible incorrect selections to last for long intervals, causing longer bursts of errors.

2.2.1. Next Generation: 5G and New Radio

In the cellular ecosystem, standardization processes are continuously proceeding and the first
steps towards 5G and new radio (NR) have been done inside Release 15, frozen in March 2019.
After a debate on how to support each V2X use case in the 3GPP standard, the conclusion is that,
while 5G NR will provide enhancements in future releases to support advanced services, Release 14
C-V2X (i.e., LTE-V2X) will be the only core of basic safety communications. Differently from Wi-Fi,
the principle does not have backward compatibility, but rather adds an optional second interface with
improved performance on different channels [20].
5G-V2X modifications designed to enhance sidelink PC5 interface will include:

• Carrier aggregation with support of up to eight bands;

• Use of frequencies above 6 GHz;
• Flexible numerology, with the possibility of subcarrier spacing of 30 kHz and 60 kHz at 5.9 GHz;
• Higher order of MCSs, also including 64-QAM;
• Possibility to transmit over single slots and even portions of slots;
• Addition of a sidelink feedback channel to allow higher reliability and lower latency;
• LDPC and polar codes designed to offer higher robustness without increasing encoding and
decoding complexity;
• Use of MIMO receiving antennas to enable spatial diversity especially useful to mitigate multipath
in urban scenarios;
Future Internet 2019, 11, 122 6 of 20

• A modification of the scheduling Modes in order to enhance the sensing-based procedures and to
deal with aperiodic traffic, also including the possibility of group reservations.

2.3. Related Work

Since the first experiments with IEEE 802.11p, studies to compare its performance to that of
cellular communications have been carried out. Long-range cellular connection is considered in some
early works, both using unicast or multicast transmissions [24–27]. Although this solution appears
possible, it is clear that it suffers from the following two issues: (1) it is not possible to guarantee a
full coverage of the territory; and, (2) with large cells, it has limited scalability and, with small cells,
the complexity increases due to frequent handovers.
The mentioned problems motivated 3GPP to add first D2D capabilities in Release 12 and specific
short-range V2X in Release 14. In Release 12, the target was mainly on public safety, but its applicability
to vehicular scenarios was already considered by a few works. For example, in [28], a non-standard
frame structure is proposed for this aim.
Since Release 14 was defined, several studies have been carried out to compare IEEE 802.11p
and sidelink LTE-V2X, which can be separated into two major groups: (1) white papers published by
those companies that have a direct interest in either technology, such as [3–6]; and (2) research papers
that have gone through a peer-review process and have in most cases no direct commercial interest
(Table 4) [7–15].
Within white papers, some push C-V2X as a more efficient technology, supported by a stronger
ecosystem and a clearer road-map for future evolutions. In [5], comparative experiments with real
devices are shown, demonstrating 1.7x–3.4x range improvement of C-V2X with respect to IEEE 802.11p
and verifying that the latency in C-V2X under congested conditions does not exceed 100 ms. Other
white papers are instead in favour of IEEE 802.11p and against C-V2X, remarking that only the Wi-Fi
technology is truly tested and reliable, and C-V2X products are not expected to be ready for the market
within a few years. In [6], the mentioned results demonstrating the superiority of C-V2X are contested
as biased by methodological weaknesses and issues with the PHY layer design are highlighted.
Peer-review papers mainly use simulations with moving vehicles and compare the two
technologies in terms of reliability [7–9,11–13,15]. Except for [15], addressing a rural area, all of
them consider a highway scenario. Some also include an urban Manhattan case [8,9]. Platooning as a
specific application is addressed in [12]. In addition, analytical evaluations are provided in [10] and a
focus on the link quality is posed in [14].
Focusing on Mode 4 of LTE-V2X, the standard protocol is always shown to guarantee the same
performance at larger distance when the basic 6 Mb/s MCS is adopted by IEEE 802.11p, especially
under low channel load conditions [7–9]. Adopting a different MCS in IEEE 802.11p, however, it is
shown in [7] that the outcome might be different, especially when the generated data traffic increases.
Results using Mode 3 are given in [8] using a not specified algorithm, in [10] under ideal conditions,
in [11] adopting an algorithm that blocks some calls in order to guarantee a configurable reliability
at a given distance, and in [15] with a simple round-robin based scheduling. Even if the mentioned
algorithms all appear as simplified solutions, yet the more complete knowledge of the network status
is shown to allow outperforming IEEE 802.11p in all cases.
In summary, all of the mentioned works basically agree that C-V2X can provide a longer range
than IEEE 802.11p. However, most of them adopt settings that might not be optimal for IEEE 802.11p.
In addition, results are not able to distinguish the contribution from the PHY and MAC layers and it is
difficult to infer what happens if the PHY layer is re-designed in IEEE 802.11p. With these limitations
in mind, in the following, we provide results from large scale simulations performed in different
scenarios and with different MCSs; in addition, we consider an ideally modified IEEE 802.11p with a
PHY layer providing the same performance as that of C-V2X.
Future Internet 2019, 11, 122 7 of 20

3. Simulation Tool and Settings

In the following, we use IEEE 802.11p to denote vehicular Wi-Fi and LTE-V2X to indicate sidelink
C-V2X, based on Release 14. In this section, some details about the simulation platform used for
numerical results, the main settings, and the output metrics are provided.

3.1. LTEV2Vsim
LTEV2Vsim is a discrete-event simulation tool, developed in MATLAB (version R2016b or later,
by MathWorks, Natick, MA, USA), aiming at modelling vehicular networks, with a focus on the
cooperative awareness service. Regarding the position and movements of vehicles, they can be either
produced following mobility models or reading the information from external traffic traces. Designed
to simulate LTE-V2X, it also includes IEEE 802.11p.
The block diagram of LTEV2Vsim is shown in Figure 1 and consists of the following parts:

Scenario description
Initialization x = f(t, ...)
y = f(t, ...)
Config files
Trace files Mobility model

C-V2X 802.11p
First radio resources
assignment
Position
Action
Simulation update
cycle
(time)
New packet Backoff
generation Transmission end conclusion

Quality assessment
Simulation
cycle
(time)
Radio resources
reassignment

Performance evaluation

Output
Output files

Figure 1. LTEV2Vsim block diagram.

• Initialization. At the beginning, configuration parameters contained in configuration files or

in the command line are read. If the position of vehicles is obtained from a traffic trace, an
interpolation is possibly performed to have position granularity not larger than the beacon period.
• C-V2X resource assignment and 802.11p management of the next event. If LTE-V2X is
simulated, the simulation step is the beacon period. At the beginning of each step, the allocation
of radio resources is performed with the adopted algorithm. In case of IEEE 802.11p, there is no
time step; the instants of the next events are continuously sorted and elaborated one after the
other. The possible events are: (1) a new packet generation, triggering the sensing and backoff
Future Internet 2019, 11, 122 8 of 20

procedures; (2) the end of a transmission; and (3) the conclusion of a backoff procedure and the
start of a transmission.
• Position update. Every beacon period, the position of all vehicles is updated, adopting the
model or traffic traces. In case of LTE-V2X, an estimated position can also be derived with some
inaccuracy to reproduce the effective knowledge at the eNodeB (e.g., assuming the use of methods
like uplink time difference of arrival (UTDOA)).
• Quality assessment. At the end of each time step in LTE-V2X and of each transmission in
IEEE 802.11p, the output metrics are updated. In particular, the correctness of each transmission
is assessed based on the average signal to noise and interference ratio (SINR) and a comparison
with a given threshold that depends on the technology and MCS. In addition to the error rate,
possibly other statistics are collected, including those related to the delay.
• End of simulation and output generation. The simulation is cyclically repeated until the
simulated time exceeds the set duration. At that point, the various metrics are elaborated into
final outputs and the simulation ends.

3.2. Main Settings

The main settings are reported in Table 5. The investigated application is the cooperative
awareness service, where each vehicle transmits in broadcast periodic messages informing of its
status and movements. Such service is at the basis of most other applications, especially those
related to safety and coordinated manoeuvres. The messages, which are called cooperative awareness
messages (CAMs) by ETSI and are a subclass of the basic safety messages (BSMs) defined by SAE, are
sent with a 100 ms period, which is the lowest value considered by ETSI. Messages of 300 bytes are
assumed, which is the largest size suggested by 3GPP in [29].
The channel bandwidth of IEEE 802.11p, i.e., 10 MHz, is assumed. All nodes transmit with
the same transmission power of 23 dBm, adopting antennas with 3 dB gain. The propagation is
modelled following the WINNER+ model, scenario B1, with a path loss exponent of 4 (except for small
distances) and with correlated shadowing, as suggested by 3GPP in [29]. The log-normal shadowing
has a standard deviation of 3 dB in line-of-sight (LOS) and 4 dB in non-line-of-sight (NLOS), and
is correlated with decorrelation distance 10 m in urban and 25 m in highway scenarios. Per each
transmission and receiver, the average SINR is calculated, taking into account the interference from all
nodes transmitting at the same time. If the bandwidth is not the same, the SINR calculation includes
the presence of IBE [30].

3.2.1. Minimum SINR Thresholds

At the PHY layer, a threshold model on SINR is adopted, as done in several similar works
(e.g., [31–33]). Indeed, especially for LTE-V2X, reference curves relating the SINR to the packet error
rate (PER) are not available and the few presented (e.g., in [5,34]): (1) do not cover all possible
combinations of MCS and packet size; and (2) all have an almost threshold behaviour.
In order to set the thresholds, in IEEE 802.11p, we used the transmission power, noise power
density, and minimum sensitivity level indicated in [35]. In IEEE 802.11p, the threshold only depends
on the MCS, whereas the packet length has no impact. Regarding LTE-V2X, we adopt the procedure
detailed in ([36], Appendix A), which starts from the Shannon capacity and considers a parametric
loss as suggested by 3GPP in [37]; this approach lets us set the threshold for any possible combination
of MCS and packet size.
Future Internet 2019, 11, 122 9 of 20

The values obtained with the detailed approaches are listed in Tables 2 and 3. Both for IEEE
802.11p and LTE-V2X, our assumptions might be restrictive and real devices could perform better.
However, the performance of the devices is strongly affected by the specific hardware. In any case, it
can be noted that, if we focus as a specific example to MCS 3 of IEEE 802.11p (QPSK, coding rate 1/2)
and MCS 6 of LTE-V2X (QPSK, coding rate 0.48), we have approximately 7 dB difference, which is
somehow in the middle between what is indicated in [5,6]. In addition, in order to overcome this
aspect and verify the performance of the two systems if they performed the same at the PHY layer, we
also simulate what we call IEEE 802.11p with LTE PHY layer, as detailed in the following subsection.

Table 2. IEEE 802.11p modulation and coding schemes and corresponding values.

MCS # Modulation, Gross Rate Minimum Range [m] Duration

Coding Rate (Mb/s) SINR [dB] (only PL) [µs]
1 BPSK, 1/2 3.0 10.0 223 848
2 BPSK, 3/4 4.5 11.0 210 584
3 QPSK, 1/2 6.0 13.0 188 448
4 QPSK, 3/4 9.0 15.0 167 312
5 16QAM, 1/2 12.0 18.0 141 248
6 16QAM, 3/4 18.0 22.0 112 176
7 64QAM, 2/3 24.0 26.0 89 144
8 64QAM, 3/4 27.0 27.0 84 136

Table 3. LTE-V2X modulation and coding schemes and corresponding values.

MCS # Modulation, RB Min. Range [m] Subch. Packets

Coding Rate Pairs SINR [dB] (only PL) Size per TTI
0 QPSK, 0.13 86 −2.83 418 5 -
1 QPSK, 0.17 66 −1.38 411 5 -
2 QPSK, 0.21 54 −0.22 404 5 -
3 QPSK, 0.27 41 1.49 392 5 1
4 QPSK, 0.33 34 2.76 382 6 1
5 QPSK, 0.41 28 4.40 365 5 1
6 QPSK, 0.48 23 5.79 354 5 2
7 QPSK, 0.57 20 7.30 336 6 2
8 QPSK, 0.65 18 8.60 320 5 2
9 QPSK, 0.73 16 9.88 306 6 2
10 QPSK, 0.82 14 11.16 294 6 2
11 16QAM, 0.41 14 11.16 294 6 2
12 16QAM, 0.46 12 12.83 278 5 3
13 16QAM, 0.52 11 14.46 258 5 3
14 16QAM, 0.59 10 16.39 237 6 4
15 16QAM, 0.67 9 18.73 213 6 4
16 16QAM, 0.72 8 20.05 203 5 5
17 16QAM, 0.75 8 21.10 191 5 5
18 16QAM, 0.84 7 23.54 172 5 5
19 16QAM, 0.92 7 25.93 150 5 5
20 16QAM, 1.00 6 28.10 137 5 5
Future Internet 2019, 11, 122 10 of 20

Table 4. Related work dealing with Vehicular Wi-Fi vs. Sidelink Cellular-V2X. n.s. means not specified.

Reference Vehicular Wi-Fi Sidelink Cellular-V2X Scenarios Main results

Molina et al. [7] IEEE 802.11p QPSK-1/2 LTE-V2X, Mode 4, pk = 0, Highway, 120 v/km LTE-V2X mostly provides better performance, but IEEE
and 16QAM-3/4 QPSK-1/2 802.11p with 16QAM-3/4 might be preferable with
higher channel load
Min et al. [8] IEEE 802.11p QPSK-1/2 Mode 3 & LTE-V2X, Mode 4 Manhattan, dense, & LTE-V2X outperforms IEEE 802.11p, Mode 3 slightly
(both n.s.), QPSK-1/2 Highway, ∼70 v/km better than Mode 4
Nguyen et al. [9] IEEE 802.11p QPSK-1/2 LTE-V2X, Mode 4 (pk n.s.), Manhattan, n.s. density, & LTE-V2X outperforms IEEE 802.11p; improvement is
(smaller power than LTE) QPSK-1/2 Highway, max ∼120 v/km large in highway, small in urban
Bazzi et al. [10] IEEE 802.11p, various MCS LTE-V2X, Mode 3 with Highway, variable density LTE-V2X mostly outperforms IEEE 802.11p, except for
blocking, various MCS high density if blocking is considered due to HD
Cecchini et al. [11] IEEE 802.11p QPSK-1/2 LTE-V2X, Mode 3 with Highway, ∼120 v/km LTE-V2X Mode 3 outperforms IEEE 802.11p; the
blocking & Pre-standard pre-standard LTE-V2X Mode 4 and IEEE 802.11p
Mode 4, QPSK-0.21&0.33 perform similarly
Vukadinovic et al. [12] IEEE 802.11p QPSK-1/2 LTE-V2X, ideal Mode 3 & Platoon in Highway, max LTE-V2X Mode 3 is collision-free; Mode 4 outperforms
modified Mode 4, pk = 0, ∼80 v/km IEEE 802.11p only at large distances with high density
QPSK-1/2
Thota et al. [13] IEEE 802.11p QPSK-1/2 LTE-V2X, non-standard Not detailed Highway With standard and non-standard application layer
Mode 4, QPSK-1/2 codes, LTE-V2X outperforms IEEE 802.11p
Anwar et al. [14] IEEE 802.11p, various MCS LTE-V2X, any Mode, Single link LTE-V2X outperforms IEEE 802.11p with similar MCS
various MCS
Kühlmorgen et al. [15] ITS-G5, QPSK-1/2 Round-robin Mode 3, Realistic rural area, low LTE-V2X outperforms ITS-G5
16QAM-0.36 density
Future Internet 2019, 11, 122 11 of 20

3.2.2. IEEE 802.11p with LTE Physical Layer

In order to focus on the performance of the MAC layer, we also simulate a version of IEEE 802.11p
where the PHY layer is assumed with the same data rate and reliability as LTE-V2X. Please remark
that this assumption, corresponding to a redefinition of the PHY of the Wi-Fi technology, also reflects
what is targeted by IEEE 802.11bd (see Section 2.1.1).
More specifically, the MAC procedure remains the same as in standard IEEE 802.11p, including
the sensing and backoff procedure, the duration of the inter-frame spaces and the MAC overhead.
Unchanged is also the first part of the message, where the training sequence and the control part are
transmitted (which is also compliant with the retro-compatibility addressed by IEEE 802.11bd). Like in
standard IEEE 802.11p, all transmissions always occupy the full bandwidth.
Differently from the standard version, the following two metrics are taken from Table 3, i.e.,
from the numbers of LTE-V2X: (1) the data rate adopted to calculate the duration of the payload
transmission; and (2) the SINR threshold. In general, these assumptions mean shorter transmissions
and/or higher reliability than standard IEEE 802.11p. Clearly, the number of possible MCSs is the
same as with LTE-V2X.

Table 5. Main simulation parameters and settings.

Application
Beacon periodicity 10 Hz
Beacon size 300 B
Common PHY layer
Channel bandwidth 10 MHz
Transmission power 23 dBm
Antenna gain (both tx and rx) 3 dB
Noise figure 9 dB
Propagation model WINNER+, Scenario B1
Shadowing variance LOS 3 dB, NLOS 4 dB
Shadowing decorrelation distance Highway 25 m, Urban 10 m
Related to IEEE 802.11p
Duration of the initial interframe space 58 µs
Random backoff [0 ÷ 15] × 13 µs
Carrier sensing sensitivity −85 dBm
Related to LTE-V2X
Mode 3
Allocation interval 2s
UTDOA positioning: 95th-percentile of error 100 m
Mode 4
Probability to maintain the allocation (pk ) 0 or 0.8
Sensing threshold to assume the channel busy −110 dBm

3.2.3. LTE-V2X Resource Allocation Algorithms

In LTE-V2X, the 10 MHz channel corresponds to 50 resource block pairs (RBPs) per subframe.
An adjacent allocation of the SCIs (i.e., control information adjacent to related data) is assumed, with a
number of subchannels that is optimized based on the number of RBPs per packet and reported in the
second column from the right of Table 3. The adopted resource allocation algorithms are as follows.
Mode 3 algorithm. When the allocation is controlled by the network, the algorithm detailed
in [38] is adopted. The network allocates the resources trying to maximize the distance between nodes
using the same subchannels. More specifically, handling the vehicles in random order, it allocates to
each of them the subframe used farther than its position (assumed known with a Gaussian error of
less than 100 m in 95% of the cases) and, within that subframe, it reserves the subchannels used farther
than its position. The allocation process is repeated every 2 s.
Future Internet 2019, 11, 122 12 of 20

Mode 4 algorithm. The standard Mode 4 described in [39,40] is adopted. It has been shown
in [22,23,41] that its performance is significantly affected by the value of pk , which allows a trade-off
between reliability and latency (see Section 2.2). For this reason, both the minimum pk = 0 and the
maximum pk = 0.8 are considered. In both cases, the initial threshold on the measured power to define
the channel as busy is set to −110 dBm as in [7].

3.2.4. Useful CAMs and Significant Neighbours

Normally, when the cooperative awareness service is addressed, the performance of IEEE 802.11p
and/or LTE-V2X is calculated for all possible nodes at some distance or within some given range.
This might include also vehicles that are on parallel roads with respect to the transmitter or are moving
away in the opposite direction, which are cases where the correct/wrong reception of the packet has
indeed no particular relevance.
To cope with this situation, here we only consider for our metrics those nodes that are possibly
interested in the sent CAM, selected with the procedure detailed in [42]. Such procedure exploits the
moving direction of each vehicle to verify if their trajectories can somehow intersect or are completely
disjoint. We call the identified nodes significant neighbours. In the example shown in [42], referring to
both technologies in an urban scenario, the reliability is significantly higher if we limit the attention to
the significant neighbours and the improvement is larger for IEEE 802.11p.

3.2.5. Scenarios
Results are provided in three scenarios:

• Cologne: This scenario is a 1.85×1.85 km2 portion, at 7:30, of the urban trace detailed in [43],
which is moderately dense; on average, there are 930 vehicles and 7.8 significant neighbours
within 100 m;
• Bologna: This scenario is a 1.6×1.3 km2 urban area, denoted in [44] as congested; queues are
present at some intersections; on average, there are 670 vehicles and 13.1 significant neighbours
within 100 m, although in some cases the significant neighbours within 100 m exceed 40;
• Highway: The scenario, detailed in [44], corresponds to approximately 16 km of an almost
congested 3 + 3 lanes highway; on average, there are 2016 vehicles and 25.4 significant neighbours
within 200 m.

In Figure 2, the cumulative distribution function (cdf) of the vehicles within a given range
(indicated as all neighbours) and the significant neighbours within the same range are shown.

1 1
Colonia Bologna
@100m, Cologne
0.9 @100m Bologna @100m, LOS
@100m
NLOS
Cumulative distribution function
Cumulative distribution function

0.8 0.8
Cologne
0.7 @100m,
NLOS
Bologna
0.6 Highway
0.6 @100m, LOS
@200m, LOS
0.5

0.4 0.4

0.3
Highway
0.2 @200m 0.2
Significant
0.1
All
0 0
0 20 40 60 80 100 0 10 20 30 40 50
Number of neighbours Number of significant neighbours
(a) Significant vs. all neighbours. (b) Significant neighbours in LOS vs. NLOS.
Figure 2. Distribution of the number of neighbours in the three scenarios.
Future Internet 2019, 11, 122 13 of 20

3.3. Outputs
The following output metrics will be used.

• Packet reception ratio (PRR): the average ratio between the number of significant neighbours
correctly decoding a beacon and the total number of significant neighbours;
• Update delay (UD): given a destination and source couple, it is the time difference between the
instant a message is correctly received and the one the last of the previous messages was correctly
received. The UD quantifies how long a node does not receive any update from one neighbour
and implicitly gives information about the correlation among errors. The UD is calculated for all
significant neighbours within 100 m in the urban scenarios and 200 m in the highway scenario.

4. Results
Results are shown focusing on the Highway in Figure 3, on Bologna in Figure 4, and on Cologne
in Figure 5. Each of these figures is divided into four subfigures and all results are shown varying the
adopted MCS. The upper two subfigures show the maximum distance that allows a PRR higher than
0.9. The lower two subfigures show the 99.99 percentile of the update delay. On the left, results refer to
IEEE 802.11p, legacy or with LTE PHY layer (11p w/LTE-PHY in the figures), while on the right, results
correspond to LTE-V2X. From Figures 3–5, we can derive the following considerations:

• Varying the MCS in IEEE 802.11p: Observing the impact of the MCS on PRR in IEEE 802.11p,
it can be noted that the optimal value strictly depends on the considered scenario. The optimal
MCS is in fact subject to a clear trade-off: a higher MCS allows for shortening the transmission
duration and to decrease the congestion of the channel, but, at the same time, it reduces the
reliability of the communication. For example, considering the standard version, the best MCS
moves from 4 in Highway to 6 in Bologna, where there is a more congested situation, and to 3 in
Cologne, where the propagation plays a more relevant role than the congestion. In addition, the
performance in terms of UD is not monotonic and the best MCS is not coincident with the one
that maximizes the PRR. In the case of UD, what is relevant is not only the error rate, but also the
correlation among errors, which increases when the channel is more congested. Anyway, except
for high MCSs, the variability of the UD is limited.
• Varying the MCS in LTE-V2X: In LTE, the trade-off between congestions and reliability is also
influenced by the way packets are allocated in the frequency domain. For example, moving from
MCS 3 to MCS 5, the number of available resources is always 1 per TTI (see Table 3), thus the
reduced reliability does not reflect a less congested channel. Due to this reason, in all scenarios
and settings, the best MCS in terms of PRR is either MCS 3 (the lowest with one packet per TTI),
MCS 6 (the lowest with two packets per TTI), or MCS 12 (the lowest with three packets per
TTI). In particular, MCS 3 provides the best performance in Cologne and MCS 6 in Highway
and Bologna (except for Highway, Mode 4 with pk = 0, where MCS 12 is slightly preferable).
In addition, in this case, with few exceptions, the variability of UD is not remarkable and the best
MCS in terms of UD does not exactly coincide with the one obtained looking at the PRR. It can
also be noted that, in most cases, the best MCS in terms of UD is higher than the one in terms of
PRR, since, in LTE-V2X, a higher MCS means more orthogonal possible allocations for the packets
and thus less probability to choose the same resource for long intervals.
• Standard IEEE 802.11p vs. LTE-V2X Mode 3: Comparing the PRR of the standard IEEE 802.11p
and that of LTE-V2X with Mode 3, the difference is always remarkable and, considering the
best cases, the latter provides from 1.4x (in Bologna) to 2.9x (in Highway) times more distance.
In terms of UD, LTE-V2X is again preferable in Highway, where neighbours are considered within
200 m, whereas IEEE 802.11p provides better results in the urban scenarios, when neighbours are
considered within 100 m. The lower UD achieved in some cases by IEEE 802.11p does not appear
anyway sufficient to balance the results in terms of PRR.
Future Internet 2019, 11, 122 14 of 20

• Standard IEEE 802.11p vs. LTE-V2X Mode 4: When Mode 4 of LTE-V2X is adopted,
the comparison with IEEE 802.11p leads to less definite results. Indeed, the conclusion changes
depending on pk : if pk = 0 is assumed, the PRR of IEEE 802.11p is higher than that of LTE-V2X
in Bologna, similar in Highway, and lower in Cologne, whereas the UD is lower in Bologna and
Cologne, and similar in Highway; if pk = 0.8 is set, then the PRR of LTE-V2X is higher, but at the
cost of a UD that is also higher.
• IEEE 802.11p with LTE-V2X PHY vs. LTE-V2X: If the LTE-V2X PHY is assumed for IEEE 802.11p,
the gap to LTE-V2X with Mode 4 becomes small. In this case, the PRR is comparable with that
of LTE-V2X with pk = 0.8 and the UD is always lower. If Mode 3 is considered, LTE-V2X still
remains preferable in terms of PRR, with a UD that is similar in Highway and Cologne and higher
in Bologna.

300 300
1 packet per TTI
IEEE 802.11p 2 packets per TTI
IEEE 802.11p w/ LTE-PHY
250 250 3 packets per TTI
Max distance with PRR>0.9 [m]

Max distance with PRR>0.9 [m]

4 packets per TTI
200 200
5 packets per TTI

150 150

100 100

LTE-V2X Mode 3
50 50 LTE-V2X Mode 4, p k =0
LTE-V2X Mode 4, p k =0.8

0 0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
MCS MCS
(a) PRR, IEEE 802.11p. (b) PRR, LTE-V2X.
10 10
IEEE 802.11p LTE-V2X Mode 3
9 9
IEEE 802.11p w/ LTE-PHY LTE-V2X Mode 4, p k =0
Update delay for target 10-4 [s]

Update delay for target 10-4 [s]

8 8 LTE-V2X Mode 4, p k =0.8

7 7

6 6

5 5

4 4

3 3

2 2

1 1

0 0
0 5 10 15 20 0 5 10 15 20
MCS MCS
(c) UD @200m with target 10−4 , IEEE 802.11p. (d) UD @200m with target 10−4 , LTE-V2X.
Figure 3. Packet reception ratio and update delay in Highway.
Future Internet 2019, 11, 122 15 of 20

100 100
IEEE 802.11p
90 90
IEEE 802.11p w/ LTE-PHY

80 80
Max distance with PRR>0.9 [m]

Max distance with PRR>0.9 [m]

70 70

60 60

50 50

40 40

30 30

20 20 LTE-V2X Mode 3
LTE-V2X Mode 4, p k =0
10 10 LTE-V2X Mode 4, p k =0.8

0 0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
MCS MCS
(a) PRR, IEEE 802.11p. (b) PRR, LTE-V2X.
5 5
IEEE 802.11p LTE-V2X Mode 3
4.5 4.5
IEEE 802.11p w/ LTE-PHY LTE-V2X Mode 4, p k =0
Update delay for target 10-4 [s]

Update delay for target 10-4 [s]

4 4 LTE-V2X Mode 4, p k =0.8

3.5 3.5

3 3

2.5 2.5

2 2

1.5 1.5

1 1

0.5 0.5

0 0
0 5 10 15 20 0 5 10 15 20
MCS MCS
(c) UD @100m with target 10−4 , IEEE 802.11p. (d) UD @100m with target 10−4 , LTE-V2X.
Figure 4. Packet reception ratio and update delay in Bologna.

100 100
IEEE 802.11p
90 IEEE 802.11p w/ LTE-PHY 90

80 80
Max distance with PRR>0.9 [m]

Max distance with PRR>0.9 [m]

70 70

60 60

50 50

40 40

30 30

20 20 LTE-V2X Mode 3
LTE-V2X Mode 4, p k =0
10 10 LTE-V2X Mode 4, p k =0.8

0 0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
MCS MCS
(a) PRR, IEEE 802.11p. (b) PRR, LTE-V2X.

Figure 5. Cont.
Future Internet 2019, 11, 122 16 of 20

5 5
IEEE 802.11p LTE-V2X Mode 3
4.5 4.5
IEEE 802.11p w/ LTE-PHY LTE-V2X Mode 4, p k =0
Update delay for target 10-4 [s]

Update delay for target 10-4 [s]

4 4 LTE-V2X Mode 4, p k =0.8

3.5 3.5

3 3

2.5 2.5

2 2

1.5 1.5

1 1

0.5 0.5

0 0
0 5 10 15 20 0 5 10 15 20
MCS MCS
(c) UD @100m with target 10−4 , IEEE 802.11p. (d) UD @100m with target 10−4 , LTE-V2X.
Figure 5. Packet reception ratio and update delay in Cologne.

Given the best MCS per each technology in terms of PRR, inferred from Figures 3–5, a comparison
in terms of PRR, varying the transmitter–receiver distance, and in terms of UD, varying the target, is
provided in Figures 6 and 7, respectively.

1 1 1
LTE-V2X Mode 4 LTE-V2X Mode 4
LTE-V2X Mode 3 pk = 0.8, MCS 6 pk = 0, MCS 3
0.9 MCS 6 0.9 0.9
LTE-V2X Mode 3
0.8 0.8 MCS 6 0.8 LTE-V2X Mode 4
LTE-V2X Mode 4 LTE-V2X Mode 4 pk = 0.8, MCS 3
0.7 LTE-V2X Mode 4 pk = 0.8, MCS 6 0.7 pk = 0, MCS 6 0.7
pk = 0, MCS 12 IEEE 802.11p
IEEE 802.11p LTE-V2X Mode 3
0.6 IEEE 802.11p 0.6 0.6 w/LTE-PHY, MCS 3
w/LTE-PHY, MCS 12 MCS 3
w/LTE-PHY, MCS 8
PRR

PRR

PRR

0.5 0.5 0.5

IEEE 802.11p
0.4 0.4 0.4 MCS 3
IEEE 802.11p IEEE 802.11p
MCS 4 MCS 6
0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0 0 0
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 0 50 100 150 200 250 300
Distance [m] Distance [m] Distance [m]

(e) Highway. (f) Bologna. (g) Cologne.

Figure 6. Packet reception ratio vs. distance.

12 9 8

8 7
10
7 LTE-V2X Mode 4
pk = 0.8, MCS 6 6 LTE-V2X Mode 4
IEEE 802.11p pk = 0.8, MCS 3
8 6
MCS 4
5
5
UD [s]

UD [s]
UD [s]

LTE-V2X Mode 4
6 LTE-V2X Mode 4 4
pk = 0.8, MCS 6
4 pk = 0, MCS 6
LTE-V2X Mode 4
3 pk = 0, MCS 3
IEEE 802.11p LTE-V2X Mode 4
4 3 LTE-V2X Mode 3
w/LTE-PHY, MCS 8 pk = 0, MCS 12
MCS 6 IEEE 802.11p
2
2 LTE-V2X Mode 3 w/LTE-PHY, MCS 3
2 MCS 3
IEEE 802.11p
LTE-V2X Mode 3 1 1
MCS 6
MCS 6 IEEE 802.11p IEEE 802.11p
w/LTE-PHY, MCS 12 MCS 3
0 0 0
10 -5 10 -4 10 -3 10 -2 10 -5 10 -4 10 -3 10 -2 10 -5 10 -4 10 -3 10 -2
Target Target Target

(a) Highway @200m. (b) Bologna @100m. (c) Cologne @100m.

Figure 7. Update delay vs. target.

If we compare the performance of LTE-V2X Mode 3 with the other cases, a higher PRR is always
observable and the UD is only slightly worse than that of IEEE 802.11p in Bologna. Even if the used
algorithm is only based on the positions of the nodes (not taking channel variability and NLOS into
account), still the wider knowledge of the network allows for achieving better performance.
Focusing on a the standard IEEE 802.11p compared to LTE-V2X Mode 4, although a lower range
is overall observable, it can be noted that its performance until a certain distance is comparable in
Future Internet 2019, 11, 122 17 of 20

terms of PRR (especially if pk = 0) and better in UD (especially if pk = 0.8). The UD is worse only in
the Highway scenario, where it is calculated within 200 m and thus affected by a low PRR.
If IEEE 802.11p with LTE PHY layer is considered, then similar performance in terms of PRR and
better in terms of UD can be noted. This suggests that, when out-of-coverage scenarios are addressed,
most of the improvement shown in the literature with LTE-V2X compared to IEEE 802.11p is due to
the PHY layer.

5. Discussion
Taking into account the conclusions from the literature (see Section 2.3) and the novel results
shown in Section 4, it is overall agreed that the improved PHY layer of LTE-V2X allows for achieving
the desired delivery rate at a longer distance in most scenarios, compared to the standard IEEE 802.11p.
However, whereas this appears clear using Mode 3 (i.e., when the resources are assigned by the
network), doubts arise assuming Mode 4 (i.e., when the selection is autonomous). Indeed, up to some
distance and subject to an optimal choice of the MCS, the standard IEEE 802.11p might behave similar
to LTE-V2X in terms of PRR and better in terms of UD. These considerations are valid also at higher
distances when an improved PHY is used in IEEE 802.11p.
In general, LTE-V2X is anyway expected to provide a longer range than the standard IEEE 802.11p
and overall significant performance improvement in most scenarios, especially if Mode 3 is assumed.
However, it should be noted that tests with real hardware are still at an initial stage. In addition,
aspects that are of simple management or have been deeply investigated in IEEE 802.11p still need
insights in LTE-V2X. For example, the congestion control mechanisms designed for IEEE 802.11p
might not be optimal for LTE-V2X, as early works suggest [45,46]. As another example, dealing with
messages of different sizes is trivial in IEEE 802.11p; in LTE-V2X, instead, given the time-frequency
organization of the resources and the SPS procedure, variable size messages need specific solutions,
as discussed for example in [7,36].

6. Conclusions
This paper provides, in its first part, an overview of the two main wireless access technologies
for V2X communications and of their roadmap and evolutions: IEEE 802.11p is ready for deployment
and is gaining momentum; concurrently, C-V2X is running to be ready as well, having the advantage
that its infrastructure is already deployed. For both technologies, we have briefly discussed the work
ongoing for a next generation, namely IEEE 802.11bd and 5G NR, respectively.
Then, we have presented a performance comparison in realistic scenarios to evaluate which
conditions and parameters mainly affect the performance of the two technologies. To summarize:
(1) LTE-V2X has an improved PHY layer allowing wider range than IEEE 802.11p, an aspect that is
well known and considered for IEEE 802.11bd; (2) LTE-V2X Mode 3 outperforms IEEE 802.11p and
LTE-V2X Mode 4, since resources are reserved by the network, which has a wider knowledge of the
node positions and allocations; (3) LTE-V2V Mode 4 outperforms IEEE 802.11p in most cases, but it
suffers from higher update delay (i.e., higher probability of several consecutive errors). Thinking of
the evolution of IEEE 802.11p, we have also evaluated its performance with an improved PHY layer,
demonstrating that, in such case, the performance in terms of packet delivery rate becomes similar to
LTE-V2X Mode 4, still granting a quite lower update delay.

Author Contributions: Conceptualization, A.B., G.C., B.M.M. and A.Z.; Data curation, A.B. and G.C.; Formal
analysis, A.B.; Funding acquisition, A.B., B.M.M. and A.Z.; Investigation, A.B., G.C. and M.M.; Methodology,
A.B. and A.Z.; Project administration, A.B.; Software, A.B., G.C. and M.M.; Supervision, A.B.; Validation, A.B.;
Visualization, A.B. and G.C.; Writing – original draft, A.B., G.C. and M.M.; Writing, review and editing, A.B.,
B.M.M. and A.Z.
Funding: This research received no external funding
Conflicts of Interest: The authors declare no conflict of interest.
Future Internet 2019, 11, 122 18 of 20

References
1. Production Statistics from the International Organization of Motor Vehicle Manufacturers (OICA). Available
online: http://www.oica.net/production-statistics/ (accessed on 1 May 2019).
2. Jadhav, A.; Sonpimple, A. Connected car market by technology (2G, 3G, and 4G/LTE), connectivity solutions
(integrated, embedded, and tethered), service (driver assistance, safety, entertainment, well-being, vehicle
management, and mobility management), and end market (OEM and aftermarket): Global opportunity
analysis and industry forecast, 2018–2025. Allied Market Research, November 2018. Available online: https:
//www.premiummarketinsights.com/reports-amr/connected-car-market (accessed on 10 May 2019).
3. Autotalks and NXP Semiconductors. IEEE 802.11p Ahead of LTE-V2V for Safety Applications, 2017.
Available online: https://www.nxp.com/docs/en/white-paper/ROADLINK-TECH-WP.pdf (accessed on
1 May 2019).
4. 5GAA. An Assessment of LTE-V2X (PC5) and 802.11p Direct Communications Technologies for Improved Road
Safety in the EU, 2017. Available online: http://5gaa.org/news/an-assessment-of-lte-v2x-pc5-and-802-11p-
direct-communications-technologies-for-improved-road-safety-in-the-eu/ (accessed on 1 May 2019).
5. 5GAA. V2X Functional and Performance Test Report; Test Procedures and Results, 2018. Available online:
http://5gaa.org/wp-content/uploads/2018/11/P-180106-V2X-Functional-and-Performance-Test-
Report_Final_051118.pdf (accessed on 10 May 2019).
6. NXP USA, I. Annex A of the Comments of NXP USA, Inc., in the Letter before the Federal Communications
Commission, Washington, WA, USA, 18 March 2019. Available online: https://ecfsapi.fcc.gov/file/
10318493028780/Wi-Fi%20Alliance%206%20GHz%20Reply%20Comments%203.18.2019.pdf (accessed on
1 May 2019).
7. Molina-Masegosa, R.; Gozalvez, J. LTE-V for Sidelink 5G V2X Vehicular Communications: A New 5G
Technology for Short-Range Vehicle-to-Everything Communications. IEEE Veh. Technol. Mag. 2017, 12, 30–39.
[CrossRef]
8. Min, W.; Winbjork, M.; Zhang, Z.; Blasco, R.; Do, H.; Sorrentino, S.; Belleschi, M.; Zang, Y. Comparison of
LTE and DSRC-Based Connectivity for Intelligent Transportation Systems. In Proceedings of the IEEE 85th
Vehicular Technology Conference (VTC Spring), Sydney, Australia, 4–7 June 2017.
9. Nguyen, T.V.; Shailesh, P.; Sudhir, B.; Kapil, G.; Jiang, L.; Wu, Z.; Malladi, D.; Li, J. A Comparison of cellular
vehicle-to-everything and dedicated short range communication. In Proceedings of the IEEE Vehicular
Networking Conference (VNC), Torino, Italy, 27–29 November 2017.
10. Bazzi, A.; Masini, B.M.; Zanella, A.; Thibault, I. On the Performance of IEEE 802.11p and LTE-V2V for the
Cooperative Awareness of Connected Vehicles. IEEE Trans. Veh. Technol. 2017, 66, 10419–10432. [CrossRef]
11. Cecchini, G.; Bazzi, A.; Masini, B.M.; Zanella, A. Performance comparison between IEEE 802.11p and
LTE-V2V in-coverage and out-of-coverage for cooperative awareness. In Proceedings of the IEEE Vehicular
Networking Conference (VNC), Torino, Italy, 27–29 November 2017; pp. 109–114.
12. Vukadinovic, V.; Bakowski, K.; Marsch, P.; Garcia, I.D.; Xu, H.; Sybis, M.; Sroka, P.; Wesolowski, K.; Lister, D.;
Thibault, I. 3GPP C-V2X and IEEE 802.11p for Vehicle-to-Vehicle communications in highway platooning
scenarios. Ad Hoc Netw. 2018, 74, 17–29. [CrossRef]
13. Thota, J.; Abdullah, N.F.; Doufexi, A.; Armour, S. Performance of Car to Car Safety Broadcast Using Cellular
V2V and IEEE 802.11P. In Proceedings of the IEEE Vehicular Technology Conference (VTC Spring), Porto,
Portugal, 3–6 June 2018, pp. 1–5.
14. Anwar, W.; Kulkarni, K.; Augustin, T.R.; Franchi, N.; Fettweis, G. PHY Abstraction Techniques for IEEE
802.11p and LTE-V2V: Applications and Analysis. In Proceedings of the 2018 IEEE Globecom Workshops
(GC Wkshps), Abu Dhabi, United Arab Emirates, 9–13 December 2018; pp. 1–7.
15. Kühlmorgen, S.; Schmager, P.; Festag, A.; Fettweis, G. Simulation-based Evaluation of ETSI ITS-G5 and
Cellular-VCS in a Real-World Road Traffic Scenario. In Proceedings of the IEEE Vehicular Technology
Conference (VTC Fall), Chicago, IL, USA, 27–30 August 2018.
16. Campolo, C.; Molinaro, A. Multichannel communications in vehicular Ad Hoc networks: A survey.
IEEE Commun. Mag. 2013, 51, 158–169. [CrossRef]
17. Festag, A. Standards for vehicular communication—from IEEE 802.11p to 5G. e i Elektrotechnik und
Informationstechnik 2015, 132, 409–416. [CrossRef]
Future Internet 2019, 11, 122 19 of 20

18. Fallah, Y.P.; Huang, C.; Sengupta, R.; Krishnan, H. Analysis of Information Dissemination in Vehicular
Ad-Hoc Networks With Application to Cooperative Vehicle Safety Systems. IEEE Trans. Veh. Technol. 2011,
60, 233–247. [CrossRef]
19. IEEE P802.11-TASK GROUP BD (NGV) MEETING UPDATE. Available online: http://www.ieee802.org/11/
Reports/tgbd_update.htm (accessed on 1 May 2019).
20. Naik, G.; Choudhury, B.; Park, J. IEEE 802.11bd & 5G NR V2X: Evolution of Radio Access Technologies for
V2X Communications. CoRR 2019, abs/1903.08391. Available online: https://arxiv.org/abs/1903.08391
(accessed on 1 May 2019).
21. Dos Reis Fontes, R.; Campolo, C.; Esteve, R.C.; Molinaro, A. From Theory to Experimental Evaluation:
Resource Management in Software-Defined Vehicular Networks. IEEE Access 2017, 5, 3069–3076. [CrossRef]
22. Bazzi, A.; Cecchini, G.; Zanella, A.; Masini, B.M. Study of the Impact of PHY and MAC Parameters in 3GPP
C-V2V Mode 4. IEEE Access 2018, 6, 71685–71698. [CrossRef]
23. Molina-Masegosa, R.; Gozalvez, J.; Sepulcre, M. Configuration of the C-V2X Mode 4 Sidelink PC5 Interface
for Vehicular Communications. In Proceedings of the 14th Conference on Mobile Ad-hoc and Sensor
Networks (MSN 2018), Shenyang, China, 6–8 December 2018.
24. Moller, A.; Nuckelt, J.; Rose, D.M.; Kurner, T. Physical Layer Performance Comparison of LTE and IEEE
802.11p for Vehicular Communication in an Urban NLOS Scenario. In Proceedings of the Vehicular
Technology Conference (VTC Fall), Vancouver, BC, Canada, 14–17 September 2014; pp. 1–5.
25. Mir, Z.H.; Filali, F. On the Performance Comparison between IEEE 802.11p and LTE-Based Vehicular
Networks. In Proceedings of the 2014 IEEE 79th Vehicular Technology Conference (VTC Spring), Seoul,
Korea, 18–21 May 2014; pp. 1–5.
26. Mir, Z.H.; Filali, F. LTE and IEEE 802.11p for vehicular networking: a performance evaluation. EURASIP J.
Wirel. Commun. Netw. 2014, 8, 1–15.
27. Vinel, A. 3GPP LTE Versus IEEE 802.11p/WAVE: Which Technology is Able to Support Cooperative Vehicular
Safety Applications? IEEE Wirel. Commun. Lett. 2012, 1, 125–128. [CrossRef]
28. Chen, S.; Hu, J.; Shi, Y.; Zhao, L. LTE-V: A TD-LTE-Based V2X Solution for Future Vehicular Network.
IEEE Internet Things J. 2016, 3, 997–1005. [CrossRef]
29. Technical Specification Group Radio Access Network; Study on LTE-Based V2X Services. 3GPP TR 36.885
V14.0.0; 2016. Available online: https://www.3gpp.org/dynareport/36885.htm (accessed on 10 May 2019).
30. Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA);
User Equipment (UE) Radio Transmission and Reception. 3GPP TS 36.101 V14.8.0; 2018. Available online:
https://www.3gpp.org/dynareport/36101.htm (accessed on May 2019).
31. Hung, S.C.; Zhang, X.; Festag, A.; Chen, K.C.; Fettweis, G. An Efficient Radio Resource Re-Allocation Scheme
for Delay Guaranteed Vehicle-to-Vehicle Network. In Proceedings of the IEEE 84th Vehicular Technology
Conference (VTC-Fall), Montreal, QC, Canada, 18–21 September 2016; pp. 1–6.
32. Martin-Vega, F.J.; Soret, B.; Aguayo-Torres, M.C.; Kovacs, I.Z.; Gomez, G. Geolocation-Based Access for
Vehicular Communications: Analysis and Optimization via Stochastic Geometry. IEEE Trans. Veh. Technol.
2018, 67, 3069–3084. [CrossRef]
33. Fritzsche, R.; Festag, A. Location-Based Scheduling for Cellular V2V Systems in Highway Scenarios.
In Proceedings of the 2018 IEEE 87th Vehicular Technology Conference (VTC Spring), Porto, Portugal, 3–6
June 2018; pp. 1–5.
34. Huawei. DMRS Enhancement of V2V. 3GPP TSG-RAN WG1 Meeting n.84, R1-160284; 2016. Available
online: https://portal.3gpp.org/ngppapp/CreateTdoc.aspx?mode=view&contributionId=681811 (accessed
on 1 May 2019).
35. IEEE Std 802.11-2016 - IEEE Standard for Information Technology—Telecommunications and Information
Exchange between Systems Local and Metropolitan Area Networks—Specific Requirements Part 11: Wireless
LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, 2016. Available online:
https://standards.ieee.org/standard/802_11-2016.html (accessed on 1 May 2019).
36. Bazzi, A.; Zanella, A.; Masini, B.M. Optimizing the Resource Allocation of Periodic Messages with Different
Sizes in LTE-V2V. IEEE Access 2019, 7, 43820–43830. [CrossRef]
37. Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA);
Radio Frequency (RF) System Scenarios. 3GPP TR 36.942 V13.0.0; 2016. Available online: https://www.
3gpp.org/dynareport/36942.htm (accessed on 1 May 2019).
Future Internet 2019, 11, 122 20 of 20

38. Cecchini, G.; Bazzi, A.; Menarini, M.; Masini, B.M.; Zanella, A. Maximum Reuse Distance Scheduling
for Cellular-V2X Sidelink Mode 3. In Proceedings of the 2018 IEEE Globecom Workshops (GC Wkshps),
Abu Dhabi, United Arab Emirates, 9–13 December 2018; pp. 1–6.
39. Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA);
Physical Layer Procedures. 3GPP TS 36.213 V14.7.0; 2018. Available online: https://www.3gpp.org/
dynareport/36213.htm (accessed on 1 May 2019).
40. Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA);
Medium Access Control (MAC) Protocol Specification. 3GPP TS 36.321 V14.7.0; 2018. Available online:
https://www.3gpp.org/dynareport/36321.htm (accessed on May 2019).
41. Toghi, B.; Saifuddin, M.; Mahjoub, H.N.; Mughal, M.; Fallah, Y.P.; Rao, J.; Das, S. Multiple Access in Cellular
V2X: Performance Analysis in Highly Congested Vehicular Networks. In Proceedings of the 2018 IEEE
Vehicular Networking Conference (VNC), Taipei, Taiwan, 5–7 December 2018; pp. 1–8.
42. Bazzi, A.; Cecchini, G.; Masini, B.M.; Zanella, A. Should I Really Care of That CAM? In Proceedings of
the IEEE 29th Annual International Symposium on Personal, Indoor and Mobile Radio Communications
(PIMRC), Bologna, Italy, 9–12 September 2018; pp. 1–6.
43. Uppoor, S.; Trullols-Cruces, O.; Fiore, M.; Barcelo-Ordinas, J.M. Generation and Analysis of a Large-Scale
Urban Vehicular Mobility Dataset. IEEE Trans. Mob. Comput. 2014, 13, 1061–1075. [CrossRef]
44. Bazzi, A.; Masini, B.M.; Zanella, A.; Calisti, A. Visible Light Communications as a Complementary
Technology for the Internet of Vehicles. Comput. Commun. 2016, 93, 39–51. [CrossRef]
45. Toghi, B.; Saifuddin, M.; Fallah, Y.P.; Mughal, M. Analysis of Distributed Congestion Control in Cellular
Vehicle-to-everything Networks. arXiv 2019, arXiv:1904.00071
46. Mansouri, A.; Martinez, V.; Härri, J. A First Investigation of Congestion Control for LTE-V2X Mode 4.
In Proceedings of the IEEE Wireless On-demand Network systems and Services Conference (WONS),
Wengen, Switzerland, 22–24 January 2019.

c 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).