Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more ➡
Download
Standard view
Full view
of .
Add note
Save to My Library
Sync to mobile
Look up keyword
Like this
0Activity
×
0 of .
Results for:
No results containing your search query
P. 1
Timer-based distributed dissemination protocols for VANETs and their interaction with MAC layer

Timer-based distributed dissemination protocols for VANETs and their interaction with MAC layer

Ratings: (0)|Views: 57|Likes:
Published by Mario De Felice
A key paradigm enabled by Vehicular Ad hoc NETworks is the support of location aware push-mode info-tainment services, besides the safety services that motivate the deployment of the technology in the first place. Message and advertisements are spread via multi-hop communications from the originating Road Side Units connected to the Internet. Support of such services relies on robust and efficient dissemination protocols. We define a timer-based vehicular backbone network protocol, where each vehicle can take forwarding decisions only based on the information read in the message header, its current state and local measurements. We analyze its performance taking into account the effect of the IEEE 802.11p MAC layer. A comparison with other literature dissemination protocols is carried out in a highway setting.
A key paradigm enabled by Vehicular Ad hoc NETworks is the support of location aware push-mode info-tainment services, besides the safety services that motivate the deployment of the technology in the first place. Message and advertisements are spread via multi-hop communications from the originating Road Side Units connected to the Internet. Support of such services relies on robust and efficient dissemination protocols. We define a timer-based vehicular backbone network protocol, where each vehicle can take forwarding decisions only based on the information read in the message header, its current state and local measurements. We analyze its performance taking into account the effect of the IEEE 802.11p MAC layer. A comparison with other literature dissemination protocols is carried out in a highway setting.

More info:

categoriesTypes, Research
Published by: Mario De Felice on Aug 08, 2013
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See More
See less

08/08/2013

pdf

text

original

 
Timer-based distributed dissemination protocols forVANETs and their interaction with MAC layer
Pierpaolo Salvo
, Mario De Felice
, Andrea Baiocchi
, Francesca Cuomo
and Izhak Rubin
§
DIET at University of Roma “Sapienza”, Rome, Italy – Email: (salvo, defelice, baiocchi, cuomo)@diet.uniroma1.it
§
Electrical Engineering Department, UCLA, Los Angeles, CA, USA – Email: rubin@ee.ucla.edu
 Abstract
—A key paradigm enabled by Vehicular Ad hoc NET-works is the support of location aware push-mode info-tainmentservices, besides the safety services that motivate the deploymentof the technology in the first place. Message and advertisementsare spread via multi-hop communications from the originatingRoad Side Units connected to the Internet. Support of suchservices relies on robust and efficient dissemination protocols.We define a timer-based vehicular backbone network protocol,where each vehicle can take forwarding decisions only based onthe information read in the message header, its current stateand local measurements. We analyze its performance taking intoaccount the effect of the IEEE 802.11p MAC layer. A comparisonwith other literature dissemination protocols is carried out in ahighway setting.
 Index Terms
—Vehicular Ad Hoc Networks; data dissemina-tion; IEEE 802.11p MAC; message forwarding
I. I
NTRODUCTION
Vehicular Ad-Hoc Networks (VANET) are a topic of grow-ing interest both as a research field and a promising technologyto enhance safety and improve the driving experience byproviding infotainment services (e.g., electronic toll collec-tion, urban sensing, parking directions, advertisements). Anemerging standard in this field is the DSRC (Dedicated Short-Range Communications) [1] that allows the communicationsamong vehicles (V2V) through ad-hoc mode and between aRoad Side Unit (RSU) and a vehicle (V2I). In these networks,information dissemination is crucial. The simplest solution,message flooding, is known to incur the so called broadcaststorm problem in dense scenarios [2]. This problem can bemitigated by designing a smart dissemination protocol.Several proposals have appeared in the literature to providemessage spreading through a VANET. One of the most knownis the Distance Defer Transmission (DDT) [3], which assignsmessage forwarding to the node farthest away from the senderamong those that have received the message. This is achievedby setting a timer in each node
A
whose value is decreasingwith the distance between sender and
A
. The other key featureof the protocol is that a node receiving a second copy of amessage gives up forwarding it and resets its timer. The RoadOriented Dissemination (ROD) protocol [4] disseminates dataseparately in each direction, and optimizes data disseminationat the intersection. The nodes are assumed to get their GPSpositions, like in DDT. RSU service area extension has beeninvestigated in other papers, like in Urban Multi-hop Broadcast(UMB) on [5] that also selects for re-transmission the mostdistant node with respect to the transmitter and uses physicalrepeaters at intersections to retransmit a message and toovercomethe problem of large buildings obstructing a messagepath. The aim of UMB is to avoid MAC collisions causedby hidden nodes, to make efficient use of the channel, andgetting reliable transmissions, thus spreading messages in alldirections at the intersection. In [6] a so called Triangle Ruleis introduced to improve dissemination in all the directionsat intersection, without further use of infrastructure, like inUMB. The problem of finding the best relay node is alsofaced in [7], with a signal burst process to select the bestrelay node to disseminate the packet. An interesting work ispresented in [8], where the Authors provide a protocol thatrelays packets using vehicle trajectories and travel statistics tomake the packet arrive at a precise point along the interestedvehicle trajectory, exactly when it passes by. It is an attractiveapproach, even if it needs many parameters to be properlytuned. In [9] a clustering algorithm is proposed, where vehiclemobility patterns are used to set vehicle clusters with highstability. This solution needs beacon exchange among vehicle,which is quite a big overhead.Our aim is to define a fully distributed dissemination pro-tocol that does not need beaconing and is suitable to supporthigh rate message flows without collapsing. This is obtained byrealizing, in a distributed way, a backbone network of vehiclesrelaying messages. The relay election is carried out by meansof local timers at the vehicles. We show through our modelsand simulations, that our algorithm meets those goals. Weaccount for the MAC delay and finite capacity. In particular,we highlight that the timer quantization effect due to MACmessage delivery delays can bring to dissemination abruptlystopping if the protocol is not properly designed.The rest of the paper is structured as follows: Section IIdescribes the proposed dissemination protocol. In Section IIIwe provide a theoretical analysis to find out the trade-ofbetween information spreading distance and end to end delay,in Section IV we evaluate the timer dimensioning with respectto the MAC delay. Simulation results of the full stack in ahighway scenario are reported in Section V. Finally in SectionVI we draw our conclusions.II. T
IMER
-
BASED DISSEMINATION PROTOCOL PRINCIPLE
The protocol architecture of nodes participating in theVANET comprises PHY and MAC layers, based on the IEEE802.11p standard, and a Forwarding Layer (FL) above, that isin charge of the dissemination logic. We refer to infotainment
 
push mode services, where a continuous flow of advertise-ments comes out of the RSU directed to all vehicles in a givenarea around the RSU (typically several square km). Since thetarget area is larger than the area directly covered by the RSU,multi-hopping is used to spread the messages.We refer to the Vehicular Backbone Network (VBN) dis-semination protocol. With VBN the messages sent out by theRSU are forwarded by the nodes that are closest to
nominalrelaying positions
, spaced out by
D
. The distance
D
is chosento as to provide a SINR adequate for the intended messageflow rate (see [10] for a full analysis of the centralized VBN).In the following we assume nodes know the value of 
D
, thatcan be optimized and sent to all nodes on a much longer timeschedule than message flow delivery (e.g., hourly).Here we present a new fully distributed version of VBN,based on local timers at the FL. The distributed version of theVBN, namely Timer-based Backbone Network disseminationprotocol (TBN) is stated with reference to a linear roadscenario, e.g., a highway. If the RSU location is taken asthe origin, nominal relaying positions are at
1
(
x
k
,y
k
) =(
kD,
0)
,k
Z
.The key idea is that each vehicle receiving a message startsa timer, the smaller the closer the vehicle to a nominal relayingposition. The vehicle closest to the nominal relay position willforward the message and by doing so will cause all othervehicles in the same forwarding interval to cancel the messagebefore they forward it as well (inhibition rule).Let us assume that each message carries a sequence number
s
and sender node coordinates
(
x
,y
)
. Node
A
maintains astate composed of three elements.1) the index of the closest nominal relay position,
k
A
=argmin
k
Z
{
d
A,k
=
 
(
x
A
x
k
)
2
+ (
y
A
y
k
)
2
}
,where
(
x
A
,y
A
)
are the coordinates of 
A
.2) a list
D
A
of integers, representing the sequence numbersof the messages that the node has processed (received,possibly forwarded, payload passed on to upper layers)3) a set
A
=
{
s
1
,...,s
m
}
of integers, representingthe sequence numbers of the
m
messages currentlyhandled by node
A
(
 pending messages
); if no message iscurrently under processing by node
A
then
A
is void.4) a set of timer values, one for each pending message:
T  
=
{
τ 
1
,...,τ 
m
}
.Let a message with header values
s,
(
x
,y
)
be receivedby a node
A
and let
k
denote the index of the nominal relayposition closest to the sender node S. If 
s
∈ D
A
, the messageis discarded and no further action is taken. If 
s
D
A
but
s
∈ P 
A
, then let
s
=
s
j
. Station
A
checks whether
k
A
=
k
.In that case, the received message is a duplicate sent by a nodein a different relaying interval than
A
, so the original messageis kept with its timer running. If instead
k
A
=
k
, then somestation has forwarded that message before station
A
; hencestation
A
has to drop message
s
j
and give up forwarding it(
inhibition rule
). Finally, if 
s
D
A
and
s
A
, this is a
1
Ordinate value of 0 is taken as the center of the road, namely the lineseparating opposite direction lanes.
new message. Its payload is passed onto to the upper layer.If 
|
k
A
k
|
= 1
, i.e., S and
A
belong to adjacent relayingintervals, then
m
is increased by one and a new entry isadded to the sets
A
and
T  
A
with values
s
m
+1
=
s
and
τ 
m
+1
=
max
2
d
A,k
A
/D
, where
max
is the maximum delaythe FL imposes to a message before forwarding. The receivedmessage is schedule to be forwarded when the timer of value
τ 
m
+1
expires.III. A
NALYSIS OF DISSEMINATION PROTOCOLS
We analyze the performance of message disseminationwith VBN, under an ideal MAC (no delay, no bandwidthlimitation). We denote the overall covered distance from theRSU via multi-hop with
, while
is the overall delay,built up as the sum of the forwarding layer timer values.Let us assume vehicles are scattered along a linear road spanaccording to a Poisson process with average spatial density
λ
.Let us assume that the nominal relay distance
D
be suchthat
D
R/
2
, where
R
is the radio coverage distance.Then, wherever the forwarder is located inside an intervalof length
D
, it can cover the entire next interval of length
D
. A message sent from the
(
k
1)
-th interval to the
k
-thinterval will be successfully forwarded provided that there isat least one station in the
k
-th interval. The probability of successful forwarding is
1
e
a
, where
a
=
λD
. Therefore,the probability of 
=
m
successful hops is
Q
m
=
(
=
m
) = (1
e
a
)
m
e
a
for
m
0
. It can be easily found that
E[
] =
m
0
mQ
m
=
e
a
1
.The covered distance is
=
D/
2 + (
1)
D
+
+
R
,where
is the displacement of the last relay mode with respectto the left extreme of the
-th interval. Since there is no biasand the timer function has a symmetric law with respect tothe center of the interval, then
E[
] =
D/
2
. So,
E[
] =E[
]
D
+
R
= (
e
a
1)
D
+
R
.As for the timer, provided there is at least one station, theprobability that the timer is larger than
t
is the probability of the event that no vehicle is found in the central part of theinterval within a distance
d
= (
D/
2)(
t/T 
max
)
of the centerof the interval. We use the variable
[
]
to denote the numberof vehicles in the interval
and let
(
α
)
[
αD/
2
,αD/
2]
.Then, we obtain:
(
τ > t
) =
k
=1
(
[
(
t/T 
)] = 0
,
[
(1)] =
k
)
(
[
(1)]
>
0)=11
e
a
k
=1
e
at/T 
[
a
(1
t/T 
)]
k
k
!
e
a
(1
t/T 
)
=
e
at/T 
e
a
1
e
a
, t
[0
,
]
We can let
τ 
=
V T 
max
, where
is a random variable in
[0
,
1]
with pdf 
(
v
) =
ae
av
1
e
a
, v
[0
,
1]
(1)The mean value of 
is
E[
] = [
e
a
1
a
]
/
[
a
(
e
a
1)]
.The overall delay is
=
k
=1
τ 
k
=
max
k
=1
k
. Since
 
and the
k
’s are independent, the mean delay is
E[
] =
max
E[
]E[
] =
max
e
a
1
a
1
(2)If we define the normalized quantities
ˆ
= (
R
)
/R
and
ˆ
=
W/T 
max
, then
a
= log(1 + E[ˆ
]
R/D
)
and
E[ˆ
] =E[ˆ
]
R/D
log(1 + E[ˆ
]
R/D
)
1
(3)This is the trade-off between average dissemination delayand average dissemination coverage. As the latter increases,also the former increases. With VBN the trade-off is quasi-linear for a given ratio between the interval length
D
and thecoverage distance
R
. The analysis above holds for
R/D
2
.In the following Sections we account for the effect of MAClayer, first as for the time quantization implied by non nulldelivery times through the MAC interface, even for light load.Then, by means of simulation we assess the effect of differenttraffic loads as the message flow rate varies.IV. E
FFECT OF
MAC
DELAY ON TIMER DIMENSIONING
At light loads, MAC frame delivery times in the IEEE802.11p are in the order of 1 ms, depending on frame payloadand air bit rate. In case two vehicles are located so that theirrespective timers differ by an amount less than the MACdelivery delay, say
ϑ
, the inhibition rule will not work. Notethat this issue affects
any
timer-based dissemination protocol.In the following we analyze this effect on TBN, again withPoisson vehicle spatial distribution.Conditional on the number
k
of vehicle in a nominalrelay interval of length
D
, vehicle positions are i.i.d. uniformrandom variables over
(0
,D
)
. Since the location of each of the
k
vehicles is uniform over
(0
,D
)
, it will be in a segmentof length
Δ
with probability
= Δ
/D
. Let us divide theinterval into
n
equal size segments, namely
Δ =
D/n
, sothat
= 1
/n
. The timer
τ 
is a map between the intervaland time, i.e.,
τ 
:
i
{
1
,...,n
}
(0
,
max
)
. Weconsider a generic nominal relay interval, say
(
D/
2
,D/
2]
,with nominal position
0
. Then, the timer is quantized to
n
values, the
i
-th of which is
τ 
(
i
)
and it is associated to the set
(
i
D
2
n
,
(
i
1)
D
2
n
]
((
i
1)
D
2
n
,i
D
2
n
]
,
i
= 1
,...,n
.We are then in a similar situation as with a random back-off mechanism where
k
competing nodes select at randomone out of 
n
slots to attempt transmission. In our case, eachvehicle “selects” at random segment
i
and schedules therelaying of the packet with a timer value
τ 
(
i
)
. Thanks to thePoisson distribution properties, the “selection” of each vehicleis independent of all other vehicles. We assume
τ 
(
i
) =
, sothat a node that sets the timer value
τ 
(
i
)
has time enough topass the message on to the MAC layer and have it deliveredto the other nodes in the road interval
before
any other vehiclethat has selected a timer value
τ 
(
 j
)
,j > i
can release its copyof the message to the MAC layer. In other terms,
ϑ
is no lessthan the MAC “resolution” time.Let us denote
(
i
) =
(
node selects τ 
(
i
))
and
G
(
i
) =
(
node selects τ 
(
 j
)
,j
i
)
,
i
= 1
,...,n
. Let
k
be theprobability that there is no forwarding superposition, i.e., asingle node out of the
k
forwards the packets and all othersbecome aware of the forwarding action on time so as to canceltheir own scheduled forwarding. Then
k
=
n
i
=1
kf 
(
i
)
G
(
i
+ 1)
k
1
(4)In our case
(
i
) = 1
/n
and
G
(
i
) = (
n
i
+ 1)
/n
,
i
= 1
,...,n
, so
k
= (
k/n
)
n
1
j
=0
(
 j/n
)
k
1
. To recover theunconditional probability of no forwarding superposition,
,we weight
k
with the probability that there are
k
vehicle inthe considered interval of length
D
, conditional on there beingat least one vehicle in the interval, i.e.,
a
k
e
a
/k
!1
e
a
for
k
1
,with
a
=
λD
. Therefore
=
k
=1
e
a
1
e
a
a
k
k
!
k
=
n
i
=1
(
i
)
ae
a
1
e
a
k
=0
[
aG
(
i
+ 1)]
k
k
!=
ae
a
1
e
an
i
=1
(
i
)
e
aG
(
i
+1)
(5)In our special case:
=
ae
a
1
e
an
i
=1
1
ne
a
(
n
i
)
/n
=
a/ne
a/n
1
(6)The probability of superposition
¯
= 1
tends to 0 as
n
increases since
1
/
[1 +
a/
(2
n
) +
o
(1
/n
)]
as
n
.For
a
= 10
, we obtain
¯
0
.
075
for
n
65
. Note that themaximum value of the timer is
max
=
. In the numericalexample above, it is
max
97
.
5
ms
for
ϑ
= 1
.
5
ms
.Concurrent forwarding of two or more vehicles in the samenominal interval due to MAC induced timer quantization canhave a catastrophic effect on message dissemination if thereis an inhibition rule. In that case, vehicles in the
(
k
+ 1)
-th nominal interval receive two or more copies of the samemessage from the nodes in the
k
-th interval and are forced tocancel message forwarding, thus stopping dissemination. Thisimpairment can be worked around by means of TBN rules,thanks to the variables
k
A
and
k
(see Section II).The average value of the winning timer is given by
E[
τ 
] =
ae
a
1
e
an
i
=1
iϑf 
(
i
)
e
aG
(
i
+1)
=
ϑa/ne
a
1
n
1
j
=0
(
n
 j
)
e
a/n
j
=
ϑa/ne
a
1(
e
a
1)
e
a/n
n
(
e
a/n
1)(
e
a/n
1)
2
1
a
1
e
a
1
the last relationship holding asymptotically as
n
→ ∞
. Hence,for large values of 
n
, the expected delay induced by therelaying timer of the FL is of the order of 
max
/a
. Thisanalysis provides equations to dimension the value of 
max
,

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->