This paper was presented as part of the Mobility Management in the Networks of the Future World (MobiWorld) Workshop at

VeMAC: A Novel Multichannel MAC Protocol for

Vehicular Ad Hoc Networks
Hassan Aboubakr Omar and Weihua Zhuang

Li Li

Department of Electrical and Computer Engineering

University of Waterloo, Canada

Communications Research Center

Ottawa, Ontario, Canada, K2H 8S2

AbstractThis paper introduces VeMAC, a novel multichannel

TDMA MAC protocol designed specifically for a vehicular ad
hoc network. The network has one control channel and multiple
service channels. On the control channel nodes acquire time slots
in a distributed way, while on the service channels nodes are
assigned time slots in a centralized manner. VeMAC decreases the
probability of transmission collisions caused by node mobility by
assigning disjoint sets of time slots to vehicles moving in opposite
directions and to road side units. Analysis and simulation results
are presented to demonstrate the efficiency of VeMAC and
compare it to ADHOC MAC, an existing MAC protocol based
on TDMA. It is shown that, for the same number of contending
nodes and available time slots, nodes can acquire time slots on the
control channel much faster in VeMAC than in ADHOC MAC,
when the number of available time slots is sufficiently larger than
the number of contending nodes.

I. I NTRODUCTION
A vehicular ad hoc network (VANET) consists of a set of
vehicles equipped with on board units for wireless communications and a set of stationary units along the road known
as road side units (RSUs). The objective of a VANET is to
provide reliable vehicle to vehicle (V2V) and vehicle to RSU
(V2R) communications. Based on these two kinds of communications, a VANET can support a wide variety of applications
in safety, entertainment, and vehicle traffic optimization [1].
Motivated by the importance of vehicular communications, the
U.S Federal Communication Commission (FCC) has allocated
75MHz in the 5.9GHz band for Dedicated Short Range
Communications (DSRC) to be exclusively used by V2V and
V2R communications. The DSRC spectrum is divided into
seven 10MHz channels: six service channels for safety and
non-safety related applications, and one control channel for
transmission of control information and high priority short
safety applications.
The special characteristics of VANETs, such as the highly dynamic network topology and diverse quality of service (QoS)
requirements of potential applications, result in significant
challenges in the design of an efficient medium access control
protocol (MAC) protocol. Various MAC protocols have been
proposed for VANETs based either on IEEE 802.11 [2] or on
channelization such as time division multiple access (TDMA)
[3], space division multiple access [4], and code division multiple access [5]. The IEEE 802.11p is a recently proposed MAC
standard for VANETs. The protocol is based on the legacy
This work was supported by a research grant from the Natural Science and
Engineering Research Council (NSERC) of Canada.

978-1-4244-9920-5/11/$26.00 2011 IEEE

IEEE 802.11 standard which is widely implemented. However,

the protocol suffers from problems such as Orphan frames [6]
and the unreliable broadcast service which is subject to the
hidden terminal problem. ADHOC MAC is based on TDMA
and is proposed for inter-vehicle communication networks [3].
Unlike the IEEE 802.11p, ADHOC MAC can support reliable
broadcast service without the hidden terminal problem, and
can cover the whole network using a significantly smaller
number of relaying nodes than that using a flooding procedure.
Moreover, in ADHOC MAC, each node is guaranteed to access
the channel at least once in each frame, which is suitable for
non delay-tolerant applications. However, simulation results
show that, due to node mobility, the throughput reduction
can reach 30% for an average vehicle speed of 50km/h [7].
Another limitation of ADHOC MAC is that it is a single
channel protocol, not suitable for the seven DSRC channels.
This paper proposes VeMAC, a multichannel MAC protocol
based on TDMA and designed specifically for a VANET. The
VeMAC assigns disjoint sets of time slots to vehicles moving
in opposite directions and to RSUs, and hence can decrease
the rate of merging collision [7] in ADHOC MAC caused by
node mobility. In addition, for the same number of contending
nodes and available time slots, nodes can acquire slots on the
control channel much faster in VeMAC than in ADHOC MAC,
when the number of available time slots is sufficiently larger
than the number of contending nodes.
II. S YSTEM M ODEL
The VANET under consideration consists of a set of RSUs
and a set of vehicles moving in opposite directions on two-way
vehicle traffic roads. A vehicle is said to be moving in a left
(right) direction if it is currently heading to any direction from
north/south to west (east), as shown in Fig.1. Based on this
definition, if two vehicles are moving in opposite directions
on a two-way road, it is guaranteed that one vehicle is moving
in a left direction while the other vehicle is moving in a right
one. Time is partitioned to frames. A frame consists of a
fixed number S of constant-duration time slots. Each frame
is partitioned into three sets of time slots: L, R, and F, as
shown in Fig.2. The F set is associated with RSUs, while the
L and R sets are associated with nodes moving in left and
right directions respectively. Every node (i.e. vehicle or RSU)
is equipped with a GPS receiver. Each vehicle can determine
its direction using GPS, and synchronization among nodes can
be performed using the 1PPS signal provided by any GPS

413

III. V E MAC P ROTOCOL

A. VeMAC Preliminaries
In the VeMAC protocol, each packet1 transmitted on the
cch is divided into four main fields: header, announcement
of services (AnS), acceptance of services (AcS), and high
priority short applications, as shown in Fig.3. Information in

Fig. 3: Format of each packet transmitted on the cch.

Fig. 1: Right and left directions.

Fig. 2: Partitioning of each frame into L, R and F sets.

receiver. The rising edge of this 1PPS is aligned with the start
of every GPS second with accuracy within 100ns even for
inexpensive GPS receivers. Consequently, this accurate 1PPS
signal can be used as a common time reference among all
nodes. Each second contains an integer number of frames as
shown in Fig.2. Hence, at any instant, each node can determine
whether the current time slot belongs to the L, R, or F set.
The VANET has one control channel (cch), and M service
channels (schs), denoted by sch1 , sch2 , . . . , schM . The cch is
mainly used for transmission of two kinds of information: high
priority short applications (such as periodic or event driven
safety messages), and control information required for the
nodes to determine which time slots they should access on
the cch and schs. The M schs are used for transmission of
safety or non-safety related application messages. Each node
has two transceivers: transceiver1 is always tuned to the cch,
while transceiver2 can be tuned to any of the M schs. For a
certain node x, the sch to which transceiver2 is currently tuned
is denoted by sch(x). It is assumed that the transmission power
levels on the cch and schs are fixed and known to all nodes.
All channels are symmetric, in the sense that node x is in the
communication range of node y if and only if node y is in
the communication range of node x. For a certain node x, the
following two sets are defined:
Ncch (x): the set of one hop neighbours of node x on the
cch, from/to which node x can receive/transmit packets
on the cch;
Nm (x): the set of expected one hop neighbours of node
x on schm , m = 1, . . . , M .
The set Nm (x) is constructed by node x, m = 1, . . . , M ,
as follows. When node x receives a packet on the cch from
another node y indicating that sch(y) = m, based on the
position of node y which is included in the header of the
packet, node x can estimate its distance to node y. Based on
this estimated distance and on the fixed transmission power
on schm which is known to node x, if node x decides that
node y is in its communication range on schm , it adds node
y to Nm (x); otherwise, node x does not update Nm (x).

the header, AnS, and AcS fields are necessary for a node to
decide which time slots it should access on the cch and schs.
On the cch, nodes acquire time slots in a distributed way and
each node must acquire exactly one time slot in a frame. A
provider is a node which announces on the cch for a service
offered on a specific sch, while a user is a node which receives
the announcement for a service and decides to make use of
this service. It is the responsibility of the provider to assign
time slots to all users and announce this slot assignment on
the sch on a specific time slot called providers main slot. For
the purpose of time slot assignment on the cch and schs, in the
header of each packet transmitted on the cch, the transmitting
node x should include i) sch(x) and the time slots used by
node x on sch(x), ii) Ncch (x) and the time slot used by each
node y Ncch (x), iii) Nm (x), where m = sch(x), and the
time slots used by each node y Nm (x), and iv) the position
and the current direction of node x (right, left, or RSU).
Subsections B and C in the following explain in details how
the nodes acquire time slots on the cch and schs respectively.
Two types of collision on time slots can happen [7]: access
collision among nodes trying to acquire time slots, and merging collision among nodes already acquiring time slots. Access
collision happens when two or more nodes within two hops
of each other attempt to access the same available time slot.
On the other hand, merging collision happens when two or
more nodes accessing the same time slot become members
of the same two-hop set (THS)2 due to node activation or
node mobility. In VANETs, merging collision is likely to occur
among vehicles moving in opposite directions or between a
vehicle and a stationary RSU. For example, in Fig.4, if vehicle
A moves to THS2, and if A is using the same time slot as D,
then collision will occur at C. Upon detection of a merging
collision on the cch, the colliding nodes should release their
time slots and acquire new ones, which may generate more
access collisions.
B. Accessing Slots on the Control Channel
Suppose node x is just powered on and needs to acquire
a time slot on the cch. Node x starts listening to the cch
for one complete frame. At the end of this frame, node x
1 The term packet is used instead of frame to refer to Layer 2 Protocol
Data Unit, in order to avoid confusion with frame which is a collection of
time slots.
2 A two-hop set is a set of nodes in which each node can reach any other
node in two hops at most.

414

Fig. 4: merging collision caused by node mobility.

can determine Ncch (x) and the time slot used by each node
i Ncch (x). In addition, since each i Ncch (x) announces
Ncch (i) and the time slot used by each j Ncch (i), node
x can determine the time slot used by each of its two hop
neighbours, j Ncch (i), j
/ Ncch (x), i Ncch (x). Hence,
by listening to one complete frame, node x can determine
the set of time slots used by all nodes within its two-hop
neighbourhood, denoted by Ucch (x). This set represents the
time slots that node x cannot use on the cch, in order to avoid
any hidden terminal problem.
Given Ucch (x), node x determines the set of accessible time
slots Vcch (x) (to be discussed) and then attempts to acquire a
time slot by randomly accessing any time slot in Vcch (x), say
time slot k. If no other node in the two-hop neighbourhood of
node x attempts to acquire time slot k, then no access collision
happens. In this case, the attempt of node x is successful and
all nodes i Ncch (x) add node x to the sets Ncch (i) and
record that node x is using time slot k. On the other hand, if
at least one node within the two-hop neighbourhood of node
x accesses time slot k, then all the transmissions fail and time
slot k is not acquired by any of the contending nodes. In this
case, node x will discover that its attempt was unsuccessful
as soon as it receives a packet from any node i Ncch (x)
indicating that node x
/ Ncch (i). Node x then re-accesses
one of the time slots in Vcch (x), and so on until all nodes
i Ncch (x) indicate that node x Ncch (i) and announce
the time slot accessed by node x. Note that node x needs at
most S 1 time slots to discover whether or not its attempt
to acquire a time slot was successful.
Consider that node x is moving in one of the right directions.
Initially, node x limits the set Vcch (x) to the available time
slots associated with the right directions, i.e. Vcch (x) =
Ucch (x)R. If after a certain number of frames, say frames,
node x cannot acquire a time slot, then node x augments
Vcch (x) by adding the time slots associated with the opposite
direction, i.e. Vcch (x) = Ucch (x) (R L). If, after more
frames, node x still cannot acquire a time slot, node x will
start to access any available time slot, i.e. Vcch (x) = Ucch (x).
The same procedure applies for a vehicle moving in a left
direction by replacing R with L. Similarly, if node x is an
RSU, for the first frames Vcch (x) = Ucch (x) F, and then
Vcch (x) = Ucch (x). The parameter is referred to as split
up parameter, and the choice of the value is critical since
it directly affects the rates of access collision and merging
collision. For example, when = 0, the rate of merging
collision is maximized since vehicles moving in opposite

directions and RSUs are accessing the same set of time slots.
However, when merging collision happens, the probability of
an access collision is minimized since each colliding node
x can choose to access any time slot in Ucch (x). On the
other extreme, when = , the rate of merging collision
is minimized since vehicles moving in opposite directions
and RSUs are accessing disjoint sets of time slots, and hence
merging collision only happens among vehicles moving in the
same direction or when a vehicle changes direction. However,
when merging collision happens, for example among vehicles
moving in a right direction, the probability of access collision
is maximized since the choice of each colliding vehicle x is
limited to time slots in Ucch (x) R. How to determine a
suitable value for to balance between the rates of access
and merging collisions should be further studied.
C. Accessing Slots on the Service Channels
As mentioned, the assignment of time slots to nodes on the
schs is performed by the providers in a centralized way. For
the slot assignment without a hidden terminal problem, each
node x should determine Um (x) defined as the set of time
slots used on schm by all nodes which are expected to be
within the two-hop neighbourhood of node x on schm . This
set represents the time slots that node x cannot use on schm ,
and will be used by the provider to assign time slots to nodes
without causing any hidden terminal problem. Each node x
constructs Um (x), m = 1, . . . , M , as follows. When node x
receives a packet on the cch from another node y indicating
that sch(y) = m, if y Nm (x), node x adds to Um (x) the
time slots used by each node j Nm (y); otherwise, node x
does not update Um (x).
When a provider (R) has a service to offer on an sch, it
announces the following information in the AnS field of the
next packet transmitted on the cch: priority of the service,
address(es) of the intended user(s), providers main slot, and
the sch on which the service will be offered. Based on the
information announced by provider R on the cch, each node
x Ncch (R) determines whether or not to make use of the
announced service. If node x decides to use the service by
provider R on schm , it transmits the following information
in the AcS field of the next packet transmitted on the cch:
Um (x), address of provider R, and the number of time slots
that node x needs. Once node x indicates its acceptance of the
service, it tunes transceiver2 to schm and waits for the time
slot assignment transmitted on the providers main slot.
After the provider announces the service, it listens to the cch
for the duration of one complete frame to determine the set
I of users which are interested in the service. For each node
x I, the provider has Um (x), the movement direction of
node x, and the number of time slots required by node x.
Accordingly, the provider assigns time slots for each x I
and announces this slot assignment in the main slot.
Because of node mobility, Um (x) may change for some
node(s) x I. Hence, the slot assignment needs to be
recalculated based on the new Um (x). For this reason, each

415

(a) Collision remains for one frame.

(b) Collision remains for four frames

Fig. 5: Collision resolving on schm

user x keeps on transmitting the updated Um (x) in the AcS
field of each packet transmitted on the cch. The provider is
therefore aware of the updates in Um (x), x I, and can
recalculate the slot assignment either periodically or when
there is a change in Um (x) for any node x I. However,
when a collision happens, some delay may exist until the
collision is resolved and new time slots are assigned by the
provider to the colliding node(s). This delay depends mainly
on the time slot assignment on the cch and schm . For example,
consider two extreme cases for the THS configuration shown
in Fig.5. Initially, nodes B, C, and D use the service by
provider R on schm , while node A
/ Ncch (R) uses the same
time slot as node D on schm . When node A moves to THS2,
transmissions from nodes A and D on schm collides at node
C and provider R. Assuming no collision happens on the cch,
for the slot assignment in Fig.5a, collision is resolved in the
next frame; while for the slot assignment in Fig.5b, collision
remains for four frames until node D is assigned a new slot by
the provider. The effect of this delay on VeMAC performance
needs further investigation.
D. Broadcast Service
This subsection shows that the efficient broadcast service
presented in [3] for ADHOC MAC can be directly supported
by VeMAC on the cch and schs, and hence no layer 3 broadcast
protocol is required. Suppose node x transmits a broadcast
packet on the cch. For each node i Ncch (x), define Zi as
the set of one-hop neighbours of node i which did not receive
the packet broadcast by node x. Node i does not relay the
packet if one of the following holds:

Zi = ;
j Ncch (i)\Zi such that Zi Ncch (j) and
|Ncch (j)| > |Ncch (i)|;
j Ncch (i)\Zi such that Zi Ncch (j), |Ncch (j)| =
|Ncch (i)|, and ID(j) > ID(i); where ID(i) denotes the
address of node i.

When node i receives a broadcast packet from node x, it listens

to the cch for S successive time slots. At the end of this
duration, node i can determine the sets Ncch (j), j Ncch (i),
and Zi = {j Ncch (i) : x
/ Ncch (j)}. Accordingly, node i
relays the packet if none of the previous three conditions is
satisfied. By using this broadcasting procedure, it is shown
in [3] and [8] that, in most cases, the minimum set of
relaying nodes needed to cover the whole network is selected.
The previous broadcasting procedure can be applied to schm
by replacing Ncch () with Nm (). However, on schm , the
broadcast service is less reliable since Nm () is the set of
expected (not the actual) one-hop neighbors on schm .
IV. P ERFORMANCE E VALUATION
Consider K contending nodes, each of which needs to
acquire a time slot on the cch. We want to determine the
average number of nodes which acquire time slots within n
frames, the probability that a specific node acquires a time
slot within n frames, and the probability that all the nodes
acquire a time slot within n frames. To simplify the analysis,
the following assumptions are made: a) all the contending
nodes belong to the same set of THSs, with the same Ucch
and Vcch , e.g. nodes C and R in Fig.5; b) the set of THSs to
which the contending nodes belong does not change; c) the
set Vcch is not augmented when a node fails to acquire a
time slot after frames, i.e. = 0; d) at the end of each
frame, each node is aware of all acquired time slots during the
frame, and updates Ucch and Vcch accordingly, i.e. all nodes
are within the communication range of each other; e) at the
end of each frame, all contending nodes are informed whether
or not their attempts to access a time slot during this frame
were successful. Based on this information, each colliding
node randomly chooses an available time slot from the updated
Vcch , and attempts to access this slot during the coming frame.
Let N be the number of initially available time slots, and Xn
be the total number of nodes which acquired time slots within

416

n frames. Under the assumptions, Xn is a stationary discretetime Markov chain with the following transition probabilities.
If K N ,
W (ji,Ki,N i)
, 0 i K 1,

(N i)Ki

ijK
pij =

1,
i=j=K

0,
elsewhere

Let P be the one-step transition probability matrix, and P n

the n-step transition probability matrix. Given that initially
all nodes are contending for time slots, i.e. X0 = 0 with
probability 1, the unconditional probability distribution of Xn
is represented by the first row of P n . That is,
n
p(Xn = i) = P1,i+1
, i = 0, . . . , K.

where W (l, v, s) is the number of ways by which l nodes can

acquire a time slot given that there are v contending nodes
each randomly choosing a time slot among s available time
slots. A node acquires a time slot if no other nodes choose to
access the same slot. The Markov chain is illustrated in Fig.6.

The probability that all nodes acquire a time slot within n

frames is
n
Fnall = p(Xn = K) = P1,K+1
.
The average number of nodes which acquire a time slot within
n frames is
K
X
n
n =
iP1,i+1
.
i=0

The probability that a specific node, say node x, acquires a

time slot within n frames is
Fn =

K
X

p(E|Xn = i)p(Xn = i) =

i=0

Fig. 6: Markov chain for Xn when K N .

K
K1
X
Ci1
i=1

CiK

n
P1,i+1
=

n
K

where E is the event that node x acquires a time slot within n

C K1

If K > N ,

= Ki since all nodes have

frames and p(E|Xn = i) = Ci1
K
i
equal chances of acquiring a time slot.

W (ji,Ki,N i)
, 0 i N 1,

(N i)Ki

ij N 1
pij =

1,
i
= j, N i K

0,
elsewhere.
The Markov chain is illustrated in Fig.7.

To calculate

Fig.8 shows Fnall for different values of N and K. It is

observed that, with a probability around 0.9, all nodes acquire
a time slot within two frames for the cases (N = 15, K = 5)
and (N = 20, K = 7), four frames for the cases (N =
15, K = 10) and (N = 20, K = 12), and eight frames for the
cases(N = 15, K = 15) and (N = 20, K = 20).

Fig. 7: Markov chain for Xn when K > N .

W (l, v, s), considering v different balls randomly distributed
in s different boxes with equal probabilities, W (l, v, s) is the
number of ways of having l boxes each containing exactly one
ball. This special occupancy problem is solved in a recursive
way as follows [9].
If v s,


vl

Clv Asl (s l)

Pvl
W
(i,
v

l,
s

l)
, 0l<v
i=1
W (l, v, s) =

A ,
l=v

l
0,
l>v
where Asl =
If v > s,

s!
(sl)!

and Cls =

Asl
l! .

vl
v s

C
A
(s l)

l
l


Psl
W (l, v, s) =
W
(i,
v

l,
s

l)
, 0l<s
i=1

0,
l s.

Fig. 8: Probability that all nodes acquire a time slot within n

frames.
Simulations have been conducted using MATLAB to verify
the analysis. In the simulations, assumption e) is removed.
This assumption is not very realistic since in VeMAC, node x
detects collision as soon as it receives a packet from any node
i Ncch (x) indicating that node x
/ Ncch (i). Consequently,
it is not realistic that all nodes detect collision together at the
end of each frame. Some nodes may detect collision earlier

417

Fig. 9: Average number of nodes acquiring a time slot within

n frames.
or later based on which slots they attempted to access during
this frame. The average number of nodes which acquire a time
slot within n frames is calculated for different K and N , and
sim
is less
denoted by sim
n . The 98% confidence interval of n
than 0.33 node for all n, K, and N . As shown in Fig.9, the
obtained from simulations without assumption
results of sim
n
e) are very close to n obtained from analysis for different K
and N .
Finally, VeMAC is compared with ADHOC MAC in terms
via simulations. For VeMAC assumptions a) to d)
of sim
n
hold, while for ADHOC MAC all nodes are assumed to be
within the communication range of each other (assumption
d)). Simulation results obtained for ADHOC MAC matches
the results originally presented in [3] for all K and N , with
the only difference that we neglect the first frame in which
each node listens to the channel to determine which slot the
node is allowed to access. This is necessary for the comparison
to be fair, as in VeMAC all nodes are initially aware of the
sets Ucch and Vcch . Fig.10 shows sim
for ADHOC MAC
n
and VeMAC with different values of K and N . It is clear
that, when N > K, nodes acquire a time slot much faster
in VeMAC than in ADHOC MAC. For example, for (N =
100, K = 50) and (N = 200, K = 100), all nodes acquire
a time slot within three frames in VeMAC, which increases
to six frames in ADHOC MAC. This result indicates that,
for the same N and K, when N > K VeMAC can decrease
the rate of access collision, as compared to ADHOC MAC.
When K increases, the gap between the performances of the
two protocols decreases, and ADHOC MAC performs slightly
better when K = N .

Fig. 10: Average number of nodes acquiring a time slot within

n frames.
any hidden terminal problem. Each node is provided with
full knowledge of the channel access of its one and two
hop neighbours, which is useful information for some layer
3 protocols. Compared to ADHOC MAC, VeMAC can make
use of the seven DSRC channels, support the same broadcast
service on the control channel and service channels, and
decrease the rates of merging and/or access collision based
on the value of the split up parameter . In the future, we
plan to determine the value of in order to balance between
merging and access collisions, to determine how the provider
should assign time slots to users on a service channel, and to
investigate the performance of VeMAC via simulations with
realistic mobility models.

V. C ONCLUSIONS AND F UTURE W ORK

This paper proposes VeMAC, a multichannel MAC protocol
based on TDMA for VANETs. Each node is ensured to access
the control channel once per frame, and hence nodes have
equal opportunities to announce for services provided on the
service channels and to transmit their high priority application
messages. The ways that nodes acquire time slots on the
control channel and service channels are designed to avoid

418

R EFERENCES
[1] Manifesto of the Car-to-Car Communication Consortium, www.car-tocar.org, 2007.
[2] 802.11p-2010 IEEE Standard for Information technology
Telecommunications and information exchange between systems
Local and metropolitan area networksSpecific requirements Part 11:
Wireless LAN Medium Access Control (MAC) and Physical Layer
(PHY) Specifications Amendment 6: Wireless Access in Vehicular
Environments, 2010.
[3] F. Borgonovo, A. Capone, M. Cesana, and L. Fratta, ADHOC MAC:
new MAC architecture for ad hoc networks providing efficient and
reliable point-to-point and broadcast services, Wireless Networks, vol.
10, no. 4, pp. 359-366, 2004.
[4] J. B. Jeremy and E. Azim, A reliable link-layer protocol for robust and
scalable intervehicle communications,IEEE Trans. Intelligent Transportation Systems, vol. 8, no. 1, pp. 4-13, 2007.
[5] H. Nakata, T. Inoue, M. Itami, and K. Itoh, A study of inter vehicle
communication scheme allocating PN codes to the location on the
road,Proc. IEEE Intelligent Transportation Systems, vol. 2, pp. 15271532, 2003.
[6] C. Nakjung, C. Sungjoon, S. Yongho, K. Taekyoung, and C. Yanghee, A
solicitation-based IEEE 802.11p MAC protocol for roadside to vehicular
networks,Proc. Mobile Networking for Vehicular Environments, pp. 9196, 2007.
[7] F. Borgonovo, L. Campelli, M. Cesana, and L. Fratta, Impact of user
mobility on the broadcast service efficiency of the ADHOC MAC
protocol,Proc. IEEE VTC, vol. 4, pp. 2310-2314, 2005.
[8] F. Borgonovo, L. Campelli, M. Cesana, and L. Coletti, MAC for ad-hoc
inter-vehicle network: services and performance,Proc. IEEE VTC, vol.
5, pp. 2789-2793, 2003.
[9] J- H. Wen and J- W. Wang, A recursive solution to an occupancy
problem resulting from TDM radio communication application,Applied
Mathematics and Computation, vol. 101, no. 1, pp. 1-3, 1999.