Radio Resource Allocation Algorithm for Device to Device based on

LTE-V2X Communications

Ahlem Masmoudi, Souhir Feki, Kais Mnif and Faouzi Zarai

NTS’COM Research Unit, National School of Electronics and Telecommunications of Sfax, University of Sfax, Tunisia

Keywords: V2X Communication, D2D Communication, Scheduling, Resource Allocation.

Abstract: Communication is important to regain transportation by accommodating real-time, easily reliable, and

actionable information flows to allow safety, mobility and environmental applications. Device-to-Device

(D2D) communication has become an emerging technology for wireless network engineers to optimize the

network performance. It is considered as an enabler for Vehicle-to-everything (V2X) applications, with

stringent reliability and latency requirements due to its ability to improve traffic efficiency, safety and

comfort. In this paper, we investigate the radio resource management (RRM) for D2D-based V2X

communications including both vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) communication.

A new algorithm called Efficient Resources Allocation for V2X Communications (ERAVC) is proposed in

order to maximize the sum rate of V2I users while guaranteeing the reliability requirement of V2Vusers.

Simulation results indicate promising performance of the proposed ERAVC scheme compared with another

existing method.

1 INTRODUCTION

Vehicle-to-everything (V2X), which includes

vehicle-to-vehicle (V2V), vehicle-to-pedestrian

(V2P), and vehicle-to-infrastructure/Network

(V2I/N) communications, improves road safety,

traffic efficiency, and the availability of infotainment

services. The third Generation Partnership Project

(3GPP) has finally deployed Long Term Evolution

(LTE)-based V2X in Release 14 standard to provide

solutions for V2X communications (3GPP TR 36.885

v.14.0.0, 2016). This latter have been greatly studied

in recent years due to its potential to improve

intelligent transport systems, effective driving

assistance and traffic safety. This technology enables

vehicles to communicate with vehicles, pedestrians,

networks and infrastructures to perceive potential

dangerous situations by collecting the information

about the environment and exchange it in real time.

These applications have small message size and strict

requirements on reliability and latency (3GPP TR

22.185 v.14.3.0, 2017).

The main problem is that the current adhoc

solutions for V2X communications over the Institute

of Electrical and Electronics Engineers (IEEE)

802.11p standard are optimized for Wireless Local

Area Network (WLAN) environment with very low

mobility. So, this does not fulfil the requirement of

V2X communications with high mobility. The

device-to-device (D2D) communication based

cellular networks is a promising enabler for V2X

communications. 3GPP Release 12 and Release 13

provide D2D to proximity services (ProSe) where

devices can directly communicate with each other

over PC5 interface (Sidelink) without passing via a

network infrastructure (3GPP TS 23.303 v.13.3.0,

2016). The D2D communications have been proposed

to meet the requirements of diverse vehicular links

with the benefits of proximity gain, reuse gain, and

hop gain (Fodor et al., 2012).

The set of resources assigned to D2D

communications is taken from the uplink resources

due to their lower peak-to-average power ratio

(PAPR) and because the uplink sub-frames are

usually less occupied than the downlink (Laurent and

Jérôme 2017). There are two resource allocation

modes the underlay mode and the overlay mode. In

the underlay mode, the cellular and vehicular users

(V-UEs) can share the same resources, so they can

achieve a best spectrum efficiency but, a strong

interference could be generated among these users. In

contrast, dedicated resources are allowed for V-UEs

in the overlay mode. The advantage of this mode is

Masmoudi, A., Feki, S., Mnif, K. and Zarai, F.

Radio Resource Allocation Algorithm for Device to Device based on LTE-V2X Communications.

DOI: 10.5220/0006857302650271

In Proceedings of the 15th International Joint Conference on e-Business and Telecommunications (ICETE 2018) - Volume 1: DCNET, ICE-B, OPTICS, SIGMAP and WINSYS, pages 265-271

ISBN: 978-989-758-319-3

265

that the eNodeB (eNB) does not need to handle

interference among the cellular and V-UEs. In fact,

radio resource management (RRM) plays a crucial

role in the performance of V2X systems, and it faces

many new challenges.

In this paper, we propose a resource allocation

algorithm named ERAVC which aims at maximizing

the sum rate of the V2I users (V2I-UEs) and

guaranteeing reliability requirement of V2V users

(V2V-UEs). The main focus is how V2V-UEs share

resources with V2I-UEs.

This paper is organized as follows. Section 2

provides related works to the proposed resource

allocation algorithm. In Section 3, we will present the

system model of resources shared among vehicles.

Section 4 introduces the proposed scheme algorithm.

In section 5, we will discuss the results and evaluation

of our proposed algorithm. Section 6 will conclude

this paper.

2 RELATED WORKS

To address the RRM challenges, a number of recent

works have proposed focusing on resource allocation

based-D2D for V2X communications. Several works

were discussed the resource allocation for V2V

services where resource are shared only among V-

UEs. Other works were considered resource

allocation for both V2V and V2I services where

resources are shared among V2V-UEs and V2I-UEs.

(Xiguang et al., 2016) designed a two-location

resource allocation algorithms (Centralized and

Distributed Scheduling) V2V broadcast services. The

main objective is to improve resource utilization

efficiency, transmission accuracy and time delay. In

the centralized scheduler, resources are allocated to

V-UEs which have less relative distance than the

distance of resource reuse. In the distributed

scheduler, authors divided the highway into several

areas and resource pool into several groups, where

users in each area select resources from a specific

group. Simulation results, show that the distributed

scheduler performs slightly better than the centralized

one.

(Shiyu et al., 2016) proposed a radio resource

allocation based on (resource block) RB sharing to

maximize the number of concurrent V2V

transmissions instead of sum rate, where multiple V-

UEs can access to one RB. The main objective is to

allow non-orthogonal access for V-UEs, where the

number of V-UEs to share the same RB is not limited.

Firstly, they transform the reliability requirement into

constraint of spectral radios matrix to limit the

interference. Then, they utilize the theory of spectral

radius estimation to improve the spectrum efficiency

greatly.

(Ashraf et al., 2017) designed a novel Quality of

Service (QoS) and proximity-aware resource

allocation for V2V communication to minimize the

total power transmission considering the queuing

latency and reliability. They achieve that by

exploiting the spatial-temporal aspects of V-UEs in

terms of their traffic demands and physical proximity.

First, a novel clustering mechanism is proposed to

group V-UEs in zones based on their physical

proximity. Then, RB are assigned to each zone based

on their QoS requirements and traffic demands.

(Jihyung et al., 2018) proposed a resource

allocation scheme based on vehicle direction,

position, speed, and density for V2V communication.

This scheme includes two resource allocation

strategies according to vehicle location, the freeway

case and the urban case. Specific resources pools are

assigned for each geometric area. For the urban case,

high vehicle density occurs in the intersection region,

so a special resource was allocated in this region

based on traffic density. For the freeway case,

resources are allocated based on vehicle direction and

position. Each zone of the freeway has a specific

resources pool and when a vehicle enters a zone, it

must allocate resources of this zone.

(Abanto-Leon et al., 2017) described a graph-

based resource allocation algorithm for broadcast

V2V communications in order to maximize the sum-

rate capacity of the system. The area is grouped into

several Broadcast Communication clusters where

vehicles should transmit in orthogonal way. Whereas

vehicles in different communications clusters can

share the same RBs. So, a solution based on bipartite

graph was introduced aims to assign every V-UE with

a RBs that maximize sum rate.

(Liang et al., 2017) designed a spectrum sharing

resources for both V2V and V2I links to guarantee the

reliability for each V2V link while maximizing the

ergodic capacity of the V2I connections. The

resources sharing can take place between V2V and

V2I users. So, they pair each V2V user with the

corresponding V2I user that satisfy the minimum

capacity requirement.

(Liang et al., 2018) proposed a graph based

resources allocations for V2V and V2I

communication. This scheme aims at maximizing the

sum V2I communication while guaranteeing the

reliability requirement of V2V communications.

Firstly, V2V users are assigned into different clusters

based on their mutual interference. Then, all V2V

users in the same cluster are allowed to share the same

WINSYS 2018 - International Conference on Wireless Networks and Mobile Systems

266

RB with one of the V2I users, while V2V users in

different clusters cannot share RB.

These proposed algorithms do not consider the

QoS requirement in totality such as delay, queuing

length, buffer state, etc. In fact, most of these

algorithms are interested only in maximizing sum rate

without giving priority among users. In this work, we

present a novel algorithm that gives priority to V-UE

based on the buffer size, throughput, and packet delay

in the time domain (TD) and maximize sum rate in

the frequency domain (FD).

3 SYSTEM MODEL

We consider a cellular vehicular environment with M

V2V and K V2I communications users. Orthogonal

Frequency Division Multiple Access (OFDMA) can

support multiple access for both V2X and cellular

communications. The total uplink bandwidth is

divided into F RBs for each scheduling unit (each

RB).

Figure 1: System Model.

As shown in figure 1, traffic efficiency messages

and infotainment applications require generally

frequent access to the remote servers or Internet for

media streaming. Hence, it is ideally supported by

high capacity V2I communication where V2I-UEs

are considered as cellular users. Whereas, safety-

critical information (e.g. cooperative awareness

messages (CAMs) (ETSI EN Std 302 637-2, 2013),

implies broadcast safety related messages among

vehicles whether in event triggered or periodic way.

Hence, it is supported by the D2D links, which

impose strict latency and reliability requirements.

In this paper, we consider the uplink direction

where orthogonal RBs are allocated to V2I-UEs to

communicate with eNB and orthogonal RBs are used

to communicate among V2V-UEs (Wanlu, 2016).

However, the V2V-UEs share the uplink resources

with V2I-UEs e.g. a RB can be shared by V2I-UE and

V2V-UE. Therefore, interference exists among V2I-

UEs and V2V-UEs. In order to allocate resources

efficiently among V2I-UEs and V2V-UEs and reduce

the complexity, a resource allocation will be proposed

in section 4.2.

4 THE PROPOSED ALGORITHM

In this section, we introduce our proposed scheduling

and resource allocation algorithm named “Efficient

Resource Allocation for V2X Communication

(ERVAC)”. The main objective is to provide the QoS

guarantees for each V-UEs, by maximizing the sum

rate of all V2I-UEs and guaranteeing the reliability

and latency requirement of V2V-UEs. For these

reasons, our proposed scheduler is able to

differentiate between these two V-UEs classes in the

time domain (TD) and the frequency domain (FD).

The overall resource allocation algorithm is

shown in figure 2. First, ERAVC classifies V-UEs in

two classes the V2V-UEs class and the V2I-UEs

class. After that, vehicles in each class are prioritized

according to their QoS requirements in the TD

scheduler (section 4.1). Then, in the FD scheduler,

ERAVC allocates resources firstly for V2I-UEs and

then for V2V-UEs (section 4.2).

4.1 TD Scheduler

V-UEs are picked from different QoS classes to be

sorted during TD scheduler which is responsible for

differentiating V-UEs according to their QoS

requirements. TD scheduler selects limited number of

V-UEs for scheduling during the next Transmission

Time Interval (TTI), and determines the priority of V-

UEs to be scheduled. As each service class has its

specific requirements, each service class has its own

metric. Then, it is necessary to study the performance

metrics of both V2I-UEs and V2V-UEs.

4.1.1 V2V-UEs Metric

The proposed metric indicates that the V2V-UE

packet with the longest delay time in the buffer

achieves the higher priority compared to a short

packet delay time. The weighted function of the m

V2V-UE is calculated as follows:



















(1)

Radio Resource Allocation Algorithm for Device to Device based on LTE-V2X Communications

267

Figure 2: Overview resource allocation algorithm.

Where 





is the largest delay time of the packet in

the buffer of the m

V2V-UE, where, 





of a packet

is the difference between the actual time and the

packet arrival time at the buffer, and 



represents

the delay tolerance of the m

V2V-UE.

4.1.2 V2I-UEs Metric

The weighted function of the k

V2I-UE is calculated

in (2) in order to include a certain fairness level and

to minimize packets waiting:



















 1









∆





(2)





 1 



 (3)





 1 



1









 1









(4)

Where:

 



represents the throughput of the k

V2I-UE

calculated in (3) using the Shannon theorem,

 



  1is the average throughput for the k

V2I-UE at time (t – 1) as shown in (4),

 τ is a constant value for averaging the k

V2I-UE

data rate,

 



 :is the queue buffer size of the k

V2I-UE,

 ∆



 : is the sum of all V2I-UEs queue buffer.















: is the PF metric (Kais et al., 2014).







∆





: aims to prioritize V2I-UE that has several

waiting packets.

4.2 FD Scheduler

The FD scheduler is responsible for allocating RBs to

each V2I-UEs and V2V-UEs classes, which are

selected by the TD scheduler for data transmission. In

this section, we introduce a resource allocation

scheme that fulfils the QoS requirements for each

class. We consider resource allocation based D2D

link for V2V-UE and cellular link for V2I-UE, where

one V2V link can share spectrum with one V2I link.

To do that, the set of RBs are allocated in two steps

as presented in Table 1.

The first step of the FD scheduler is responsible

for allocating RBs only for V2I-UEs. Whereas, the

second step is responsible for allocating RBs already

assigned to V2I-UEs from the first step to V2V-UEs.

In the latter allocation, RBs are shared with V2V-UEs

only if they don’t interfere with V2I-UEs.

In the first step, RBs are allocated in order to

maximize the sum rate for V2I-UEs using Shannon

capacity as follows:





1



















(5)

Where 





is the SINR of the k

V2I-UEs over the f

RB.

After allocating each RB to the corresponding

V2I-UE in the first step, we will follow the RB

already allocated to V2I-UEs, for each V2V-UEs, in

the second step. If the condition (6) is verified, this

user can share this RB with V2I-UEs. So, the V2V-

UE will share RBs with the corresponding V2I-UE if

the SINR of the k

V2V-UE (

,

) is higher than the

SINR threshold (





) as shown in (6):



,





,











,





,







(6)

Where 

,

and 

,

denote the transmit power of the

V2V-UE transmitter and the k

V2I-UE

transmitter over f

RB, respectively,



and 

,

are

the desired channel power gain of the m

V2V-UE

and the interference channel power gain from the k

V2I-UE to the m

V2V-UE, respectively, and 



the noise power.

WINSYS 2018 - International Conference on Wireless Networks and Mobile Systems

268

Table 1: ERAVC Algorithm.

Algorithm 1: ERAVC

K ← total number of V2I-UEs

M ← total number of V2V-UEs

F ← total number of RBs







: % SINR threshold of V2V-UEs

// TD Scheduling

Sort V2V queue buffer according to (1)

Sort V2I queue buffer according to (2)

// FD Scheduling

Step 1: RBs allocation for V2I-UEs

Fork:= 1 to K do

Forf:=1 to F do

Calculate C(k,f); % capacity of the

V2I-UE on the f

End for

// selecting the RB that maximize sum rate of the

V2I-

UE according to (5)

Find (k,f) = max C(k,f), with (k,n) ∈K*F

Assign the

RB to the k

V2I-UE

End for

Step 2: RBs allocation for V2V-UEs

For f:= 1 to F do

For m:=1 to M do

//Find the RB already assigned for V2I-UE to be share

with V2V-UE.

If

,







then

Assign the f

RB to the m

V2V-UE

//V2V-UE share the

RB with V2I-UE.

End If

End for

5 SIMULATION RESULTS

In this section, we present the simulation result to

validate our proposed resource allocation algorithm

ERAVC. We consider a simulation model composed

of a single cell of the radius equal to 1.5 km, one eNB

carrier frequency of 2 GHz, a system bandwidth of

5MHz. The number of V2V-UEs and V2I-UEs

varying between 50 and 500. We consider the

simulation of freeway case as detailed by (3GPP TR

36.885 v.14.0.0, 2016). The path loss model for V2V-

UEs links is calculated according to WINNER model

(WINNER II, 2007). Whereas the path loss for V2I-

UEs links is computed as follows:

  128.137.6log



 (7)

Where d is the distance of the communication link (in

km). The simulation and configuration parameters are

presented in Table 2.

Table 2: Simulation parameters.

Parameter Value

Cell Radius 1.5 km

Number of eNB 1

Bandwidth 5 MHz

Number of RBs 25 RBs

OFDM symbols per slot 7

Shadow fading standard deviation 9 dB

Carrier frequency 2 GHz

Maximum V2V-UE transmit power 23 dBm

Maximum V2I-UE transmit power 23 dBm

SINR threshold of V2V-UEs 5 dB

V2V and V2I-UEs speed Random (5, 150) km/h

TTI 1 ms

Number of V2V-UEs / V2I-UEs 50 -500

Simulation length 5000 slot

Scheduling/Allocation resource slot

Average Packet size for V2V-UEs 50-300 Bytes

Average Packet size for V2I-UEs 1200 bytes

Noise power 



-114 dbm

Path loss model

LOS in

Winner + B1

(WINNER II, 2007)

Figure 3: Average throughput for V2I-UEs.

Figure 4: Packet Drop Rate for all V2X-UEs.

Radio Resource Allocation Algorithm for Device to Device based on LTE-V2X Communications

269

Figure 5: Average Queuing Delay for V2V-UEs.

To evaluate our proposed ERAVC algorithm and

demonstrate its effectiveness, we compare it with the

“Position-based resource allocation” (Jihyung, 2018).

It can be noted that the “Position-based resource

allocation” algorithm gives the worst performance in

terms of throughput, delay, and Packet Drop Rate

(PDR). This is achieved by RBs sharing among V2I-

UEs and V2V-UEs.

Figure 3 demonstrates the average throughput of

the system for V2I-UEs. The proposed algorithm

(ERAVC) gives the highest rate as it selects the users

having the maximum reported SNIR value in order to

improve network efficiency and cell performance.

Therefore, ERAVC utilizes efficiently the radio

resources since it selects packets of users with the best

channel conditions. Whenever the network is more

congested, ERAVC algorithm becomes more

efficient, since it gives priority to V2I-UEs that have

several waiting packets, and reaches the best rate.

Figure 4 presents the PDR of all V2X-UEs where

our ERAVC algorithm offers the best rate compared

with the “Position-based Resource Allocation”. For

this latter, the average PDR performance becomes

worse as it does not take in consideration the buffer

size as in our ERAVC scheme. Indeed, each zone has

its specific RBs, which are selected based on vehicles

position and direction, which may result resource

starvation.

Figure 5 illustrates the average queuing delay

versus the number of V2V-UEs. It can be seen that

our ERAVC algorithm achieves a queuing delay

reduction for V2V-UEs as compared to the “Position-

based Resource Allocation”. This is achieved by the

TD scheduler that favours packets with the longest

delay time in the buffer and reduces the number of

miss deadline packets.

6 CONCLUSION

In this paper, we have investigated the resource

allocation for D2D-based V2X networks. In this type

of networks, both V2V-UE and V2I-UE

communication links coexist and each V2V-UE can

share spectrum with one V2I-UE. Indeed, we

proposed a new algorithm called “Efficient Resources

Allocation for V2X Communications (ERAVC)” that

aims at maximizing the sum rate of V2I-UEs while

guaranteeing the reliability and latency requirement

of V2V-UEs. Compared to an existing algorithm

“Position-based resource allocation”, our proposed

algorithm can achieve best performance in term of

throughput, delay and PDR since it takes into

consideration the buffer size, throughput, and packet

delay.

REFERENCES

3GPP TR 36.885 v.14.0.0, 2016. Study on LTE-based V2X

Services (Release 14).

3GPP TR 22.185 v.14.3.0, 2017. Service Requirements for

V2X Services.

3GPP TS 23.303 v.13.3.0, 2016. Universal mobile

telecommunications system (UMTS); LTE; proximity-

based services (ProSe); Stage 2 (release 13).

Abanto-Leon L.F., Koppelaar, A. and Heemstra S.M.,

2017. Graph-based resource allocation with conflict

avoidance for V2V broadcast communications, IEEE

International Symposium on Personal, Indoor and

Mobile Radio Communications (PIMRC), Canada.

Ashraf, M. I., Liu, C-F., Bennis, M., and Walid S., 2017.

Towards low-latency and ultra-reliable vehicle-to-

vehicle communication, IEEE European Conference

Networks and Communications (EuCNC),. p. 1-5.

ETSI EN Std 302 637-2 v. 1.3.0., 2013. Intelligent transport

systems; Vehicular communications; Basic set of

applications; Part 2: Specification of cooperative

awareness basic service.

Fodor, G., Dahlman, E., Mildh, G., Parkvall, S., Reider, N.,

Miklos, G., and Turanyi, Z., 2012. Design aspects of

network assisted device-to-device communications,

IEEE Communications Magazine, vol. 50, no. 3, pp.

170–177.

Jihyung, K., LEE, J., MOON, S., and Intae H., 2018. A

Position-based Resource Allocation Scheme for V2V

Communication. Wireless Personal Communications.

vol. 98, no 1, p. 1569-1586.

Kais, M., Ahlem, M., and Lotfi, K., 2014. Adaptive

efficient downlink packet scheduling algorithm in LTE-

advanced system, IEEE International Symposium on

Networks, Computers and Communications.

Laurent, G. and Jérôme, H., 2017. A Case Study for Safety-

Critical V2X Communications, IEEE vehicular

technology magazine.

WINSYS 2018 - International Conference on Wireless Networks and Mobile Systems

270

Liang, L., Lu, G. Y., et Xu, W., 2017. Resource allocation

for D2D-enabled vehicular communications, IEEE

Transactions on Communications. vol. 65, no 7, p.

3186-3197.

Liang, L., XIE, S., Li, G. Y., Zhi D., and Xingxing

Y., 2018. Graph-Based Radio Resource Management

for Vehicular Networks, arXiv preprint arXiv:

1801.02679.

Shiyu, Z. H., Yanzhao, X., X., and X. T., 2016. Resource

allocation in D2D-based V2V communication for

maximizing the number of concurrent transmissions,

IEEE International Symposium on Personal, Indoor,

and Mobile Radio Communications (PIMRC). p. 1-6.

Wanlu, S., 2016. On Medium Access Control for Vehicular

Communication Over Device-to-device Links: Radio

Resource Management and Network Synchronization.

Chalmers University of Technology.

WINNER II Channel Models, 2007. IST-4-027756

WINNER II D1.1.2 V1.2. [Online]. Available:

http://projects.celtic-

initiative.org/winner+/WINNER2Deliverables/D1.1.2

v1.1.pdf.

Xiguang, Z., Shang Y., Li X., and Jiayi F., 2016. Research

on overlay D2D resource scheduling algorithms for

V2V broadcast service, IEEE Vehicular Technology

Conference (VTC-Fall). p. 1-5.

Radio Resource Allocation Algorithm for Device to Device based on LTE-V2X Communications

271