An Experimental Comparison of Dynamic Routing Protocols in Mobile

Networks

Glazunov Vadim

, Gusikhin Oleg

, Kurochkin Leonid

, Kurochkin Mihail

and Popov Sergey

Telematics department, Saint-Petersburg State Polytechnic University,

Polytechnicheskaya st. 29, Saint-Petersburg, 195251, Russia

Research and Advanced Engineering, Ford Motor Company, Dearborn, Michigan, 48121, U.S.A.

Keywords:

Routing Protocol, Mesh-networks, Vehicle, Wireless Networks, Simulation, B.A.T.M.A.N., OLSR,

Real-world Simulation, LTE, Linux, 802.11s, Testbed.

Abstract:

The paper deals with the research of dynamic attributes of routing protocols in the vehicle’s movement sim-

ulation in urban environment. Simulation environment allows to modify the volume of the transmitted data,

transmission speed, failures’ character and intensity. The research was performed for OLSR and B.A.T.M.A.N.

protocols. The obtained results permit to formulate optimal conditions for using these protocols on the vehi-

cle’s board.

1 INTRODUCTION

Application of wireless protocols for data transmis-

sion in vehicles considerably widens information in-

teraction of all trafﬁc parties, from vehicle manufac-

tures to passengers (Zaborovski et al., 2013). Nowa-

days they help to solve tasks of trafﬁc control, to mon-

itor the vehicle conditions, to prevent accidents, to

provide aid in emergency cases, to offer entertain-

ment activities to passengers, and to plan transporta-

tion process (Glazunov et al., 2013b). This results in

a considerable change in the volume and frequency of

transmitted data.

Currently, it is possible to set a multi-protocol

unit (MPU) aboard the vehicle. It supports an ar-

bitral number of data transfer protocols and ensures

high reliability of message transfer between the ve-

hicle and the access point. A stationary transmitter

of LTE, DSRC network or mobile one — for mesh-

network (Chakraborty and Nandi, 2013) is an access

point. To transmit data the vehicle makes a stable

connection with the nearest access point within 100

to 800 meters. Constant replacement of one access

point by another one occurs during the motion and in

some time intervals the connection can fail.

Besides, structural features of buildings distort

the standard picture of signal propagation in the air,

which also has an impact on connection stability. In

this situation the task of reliability research of data

transfer for each type through different technologies

of wireless connection is of special priority (Glazunov

et al., 2013a).

This paper deals with two technologies: LTE and

Mesh as they are most common among mobile clients.

The objects of the research are B.A.T.M.A.N. (Pereira

et al., 2012) and OLSR (Laouiti et al., 2008) dynamic

routing protocols since they are directly oriented at

use in dynamic mobile networks.

2 RELATED WORK

There have been several studies about the perfor-

mance of different routing protocols in wireless mesh

networks (Murray et al., 2010), (Abolhasan et al.,

2009) and (Kulla et al., 2012). In these papers au-

thors compare the performance in terms of transport

layer protocols in Ad-Hoc networks. They analyze

the following terms: overhead, latency, data transfer

speed.

In the paper (Murray et al., 2010) and (Abol-

hasan et al., 2009) authors compare OLSR and

B.A.T.M.A.N. routing protocols by average ﬁle and

ICMP requests set transfer time within experimental

Wi-Fi networks. They are using only 802.11a tech-

nology in experiments.

In the paper (Kulla et al., 2012) authors test and

verify the quality of message delivery and network

loading by OLSR, B.A.T.M.A.N. and DSR protocols

in buildings.

775

Glazunov V., Gusikhin O., Kurochkin L., Kurochkin M. and Popov S..

An Experimental Comparison of Dynamic Routing Protocols in Mobile Networks.

DOI: 10.5220/0005062107750782

In Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics (IVC&ITS-2014), pages 775-782

ISBN: 978-989-758-040-6

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

In this paper OLSR and B.A.T.M.A.N were se-

lected based on above mentioned studies. How-

ever our paper considers the changed task that re-

places the homogeneous network by heterogeneous

Wi-Fi(802.11bg) + Mesh(802.11s) networks. Multi-

protocol node will be alternative means of data trans-

mission over wireless local area networks. In dy-

namic routing protocols instead of transport protocols

the user’s high-level protocols are taken for the most

adequate estimation of the user’s actual scenarios.

3 ANALYSIS OF RESEARCH

PARAMETERS

Data Volume Analysis

In modern intelligent transport systems (Zaborovskiy

et al., 2013) the following scenarios of data transfer

(see Table 1) are realized between mobile clients (Vil-

han and Hudec, 2013). Limited volume of data and

frequency renewal are selected for each scenario.

Table 1: Limited data volume.

Data Transfer Scenario Limited

Volume,

KBytes

Event fre-

quency

Messaging of weather conditions, of pave-

ment conditions, current changes in trafﬁc

network

10–100 event-

trigger

Gathering and displaying of location data

of other vehicles

10–10

constant

Autosave of the vehicle’s motion tracks

with video record of key motion points

during motion or by request

10–10

constant

Accidents messaging 10–100 event-

trigger

Distant upload of new software 10

–10

occasional

Audio and video consultation on the vehi-

cle technical conditions

–10

occasional

Telecommunication environment for pas-

sengers (media content)

–10

constant

Messaging on technical checkup or urgent

repair

100–10

event-

trigger

Distant control of the vehicle’s units 10–10

event-

trigger

The ﬁles’ volume ranges from 10KB to 100MB.

For example, the amount of data for:

• information messages, such as eCall, geoinforma-

tion, service messages, are less than 10K;

• text documents, e-mails and hyperlink ﬁles range

from 10 to 100K;

• ﬁles of media-contentgraphic data or software up-

date range from 1M to 10M;

• audio and video ﬁles, user software and program

data updates range from 10M to 100M;

Therefore, 10KB, 100KB, 1MB, 10MB, 100MB

are typical data volumes, which will be used in exper-

iments.

Channel Analysis

Factual speed of data transfer via communication

channels essentially differs from maximum declared

for a technology in question. In practice it is limited

by:

• service providers;

• network hardware settings;

• software settings of server infrastructure;

Besides, the transfer speed at a speciﬁc point de-

pends on the channel loading that has a random char-

acter and depends on clients’ activity. Loading varia-

tion of the LTE channel causes the user’s uncertainty

about current speed of access to ﬁles and services. To

reduce the time of information exchange it is neces-

sary to have alternative ways of ﬁle delivery and ser-

vice access via a chain of retranslators.

The following technologiesof access through con-

nection channels are chosen for this research:

• LTE provides connection with the cloud environ-

ment and ﬁle access services.

• Mesh provides alternative opportunity of interac-

tion between mobile clients and connection with

the cloud environment via retranslators.

Analysis of Routing Protocols

OLSR and B.A.T.M.A.N. protocols are mainly used

in mobile networks with high dynamics network

topology changes. For example these protocols cur-

rently used in urban mesh-networks with Wi-Fi tech-

nology. These networks consist of more than 1000

nodes and now is used in Greece, Italy, Germany

and other European countries. In these networks mo-

bile users used as a mobile routers. At the moment,

there are applications for smartphones for the An-

droid operating system supporting B.A.T.M.A.N.pro-

tocol (Jayasooriya et al., 2013). For example the

mobile smartphone application is described in the

article (Jayasooriya et al., 2013) presents functions

for accessing web-content (multimedia) and reduce

download global data networks. Another important

example of use of protocols OLSR and B.A.T.M.A.N.

is information networks in vehicles (VANET). Such

networks are now being actively investigated and

modeled in order to further the development of infras-

tructure and dynamic stationary transmitters in mo-

bile networks, the development of intelligent trans-

port systems (ITS).

The research considers two dynamic routing pro-

tocols: OLSR and B.A.T.M.A.N. They are devel-

oped for mobile networks and permit to ﬁnd routes

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

776

of data transfer after the network reconﬁguration.

Realizations of two classes of protocols are consid-

ered: B.A.T.M.A.N. is a protocol based on calcula-

tion of distance vector while OLSR is based on chan-

nel condition analysis. There are differences between

them: OLSR builds a complete route to the recipi-

ent whereas B.A.T.M.A.N. calculates a distance vec-

tor and chooses next transition on this basis (Klein

et al., 2012). It allows to measure time difference

when ﬁles of different size are transferred at different

speeds with or without interference.

Data transfer between the server and the client is

performed by means of layer protocols application:

ftp and http (Zaborovskyet al., 2009). These are stan-

dard protocols available in many web-applications

and are used to receive hypertext, media, and graphic

data.

Interference Analysis

Local technologiesin the connection of mobile clients

regulate two modes of connection function:

Stable connection — the vehicle is immovableand

is connected to a mobile local network, e.g. via Wi-Fi.

Unstable connection — the vehicle is moving

from one access point to another. In this case, inter-

vals of connection failure occurs. Time of connec-

tion failure can be simulated on the basis of recipro-

cal positions of immovable access points and mobile

retranslators, e.g., via mesh.

To study the features of the second mode of lo-

cal technologies functions, a simulation testbed, that

provides time registration of the chosen volume data

interchange, was assembled.

4 THE EXPERIMENT

Stand Hardware and Software Conﬁguration

The testbed is a hardware-software complex imple-

menting interaction of multi-protocol unit (MPU)

prototypes with a load simulator for LTE, Mesh, and

Wi-Fi. Structural chart of testbed’s hardware are

shown in Fig. 1.

To implement the scenario the testbed is set as fol-

lows: it consists of two MPU blocks, a load simu-

lator, and a Wi-Fi access point. The load simulator

provides mobile network clients with access to ﬁles.

Each MPU block provides operation with two inter-

faces — Mesh and Wi-Fi or Mesh and LTE which al-

lows to transfer data in different directions using dif-

ferent technologies. MPU-1 acts as a client that con-

nects to the load simulator. MPU-2 is designed for the

retranslator that connects MPU-1 and the load simu-

lator. The logic chart of the testbed connection for the

scenario realization is shown in Fig. 2

Figure 1: Structural chart of hardware conﬁguration.

Figure 2: The logical chart of the testbed connection of the

system and client software update scenario.

The applied technologies for MPU-1 connection,

realizing the client’s functions, and the load simula-

tor, acting as a server, and MPU-2, functioning as the

retranslator between MPU-1 and the load simulator,

are shown on the chart.

Software architecture shown in Fig. 3 is imple-

mented to realize scenarios of mobile clients mo-

tion, data transfer stream management, interference

simulation, routing protocols choice, communication

AnExperimentalComparisonofDynamicRoutingProtocolsinMobileNetworks

777

channels feature setting.

Figure 3: Software architecture.

The testbed’s software consists of ﬁve modules.

Network interface module. It sets up IP-addresses,

VPN connection, internal and external interfaces of

points.

Routing module. It conﬁgures parameters of dy-

namic routing protocols, adds ﬂoating static routes,

start services (daemons) of dynamic routing proto-

cols.

Network trafﬁc generation module. It starts ﬁle

services (daemons), in this research being FTP, HTTP

and MQTT (Hunkeler et al., 2008) protocols. It starts

FTP, HTTP and MQTT for clients to download ﬁles

of chosen size, with ﬁles previously generated in back

end.

Firewall module. It makes interference on chan-

nel level at certain points of time that prevent distri-

bution of routes and data in a chosen communication

channel. This module is used by generator module to

create a mobile chart of the point movement.

Statistics modules. It records ﬁle download time

with consideration of reconnection time after connec-

tion failure. Besides, the module turns on data aggre-

gation facilities to calculate numerical characteristics

for graphic dependencies construction.

Core software includes:

• Linux Debian 7.4/Kernel 3.2.0-4.

• HTTP Server: Lighttpd 1.4.31

• FTP Server: Vsftpd 2.3.5

• MQTT 3.1 broker/client: Mosquitto 0.15-2

• OLSR protocol implementation: olsrd 0.6.2

• B.A.T.M.A.N. protocol implementation: batmand

debian-0.3.2-12 (compatibility version 5)

• FTP/HTTP Client: wget 1.13.4.

• Network ﬁlter: iptables 1.4.14.

Experiment Procedure

The developed procedure includes the following

stages:

1. Interface set-up (IP-address, VPN connection,

mesh-network).

2. Firewall set-up.

3. Set of dynamic routing protocol.

4. Selection of mesh channel function mode.

5. Set of application layer protocols.

6. Selection of message ﬁle size.

7. Selection of server response speed.

8. Selection of amount of iterations (10000).

Time of ﬁle transfer is measured.

The Experiment

Experimental research was carried out in order to de-

termine ﬁle transfer time with application of OLSR

and B.A.T.M.A.N routing protocols under conditions

of interference corresponding to the vehicle’s move-

ment in the urban.

Let us simulate a situation when the vehicle that

has a LTE dedicated channel periodically connects to

a mesh-network that has a high-speed gate in Internet.

The vehicle downloads ﬁles of user software updates

and program data.

To simulate the process of ﬁle download from the

server to the client we chose FTP, HTTP or MQTT

protocols as the most available for this application.

Two alternative channels, LTE and Mesh, for data re-

ceiving were used in the experiment.

Mesh-network channel as more speedy was cho-

sen as a priority direction of ﬁle download.

In-out parameters of experiments are given in Ta-

ble 2.

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

778

Table 2: Experiments initial data.

Parameter Value

Routing protocol OLSR, B.A.T.M.A.N

File size 10KB, 100KB, 1MB,

10MB, 100MB

Server download speed

limit

128Kbit/s, 1024Kbit/s,

11Mbit/s(Unlimited)

File download protocol FTP, HTTP, MQTT

Fixed Parameters

• OLSR parameters: Hello Interval — 1.0;

HelloValidityTime — 3.0; LinkQualityFishEye

— 1.

• B.A.T.M.A.N. parameters: OGM Interval — 1.0.

• Ftp, http download parametrs (wget): Network

timeout — 2; Waitretry — 1; Tries — 0 (unlim-

ited); Retry-Connrefused — yes.

• ftp server parameters (vsftpd): anon

max rate —

6750000, 128000, 16000.

• http server parameters (lighttpd): server.kbytes-

per-second — 6750, 128, 16.

During the experiment ﬁnite time of ﬁle transfer

for FTP, HTTP and MQTT protocols is measured.

Mesh-network interference chart during the vehicle

movement on the highway is shown in Fig. 4. MQTT

used as a publish-subscribe based “light weight” mes-

saging protocol for use on top of the TCP/IP proto-

col. MQTT as ftp and http uses as a transport layer

protocol — TCP. It corresponds to the vehicle move-

ment between cross-roads at speed of 60 kmh. The

simulation is introduced by interchange of time inter-

vals with stable and unstable connection. One cycle

is equal to 60 sec which corresponds to one kilometer

of the highway.

The ﬁrst interval of a short-time connection failure

of 1 sec illustrates the situation of reconnection to a

new access point. Next 7 seconds the vehicle is mov-

ing within the range of new access point functioning,

then the signal from all neighborswill fails, which can

be compared with driving in a tunnel or driving in a

territory without 802.11s technology coverage. After

that a zone with stable connection with several vari-

eties of network access appears which lasts 15 sec-

onds (250 meters). Further on, a short-time connec-

tion failure occurs for a shorter period of time (3 sec).

Finally, a complete simulation cycle is repeated.

This operation model corresponds to the opera-

tional mode of dynamic routing protocol and data

transfer under conditions of increased instability of

network topology.

Figure 4: Chart of noise into mesh-network.

Table 3 shows experimental values of ﬁle down-

load time in mesh-network for different values of

server response speed. These values can be consid-

ered as the best ones as they were received under

ideal conditions without interference and failures in

communication channel operation.

Table 3: Experimental values of ﬁle download time in

mesh-network without noise.

File size,

KBytes

Download Speed Limit, Kbit/s

11000 1024 128

10 0.17 0.17 0.39

100 0.23 0.67 3.70

0.91 8.20 43.49

6.63 80.26 432.28

64.37 801.09 4320.94

During the investigation 3 experiment series for

each speed were performed, with each series amount-

ing to 400 measurements. Figures 5, 6 and 7 show

graphic presentation of results of these experiments

ﬁndings.

2000

4000

6000

8000

10000

12000

10 100 1000 10000 100000

time, sec (less is better)

file size, KBytes

B.A.T.M.A.N.

OLSR

experimental values without noise

Figure 5: Dependency of ﬁle transfer time on size for

B.A.T.M.A.N and OLSR protocols at 128 Kbit/s speed (ﬁrst

series).

Results of the ﬁrst series demonstrate dynamics

of data transfer at low speed (128 Kbit/s). Such

speed is typical of bulk data transfer that includes

information about trafﬁc network and other clients.

The chart presents dependency of data transfer time

AnExperimentalComparisonofDynamicRoutingProtocolsinMobileNetworks

779

via http-protocol on ﬁle size with the application of

B.A.T.M.A.N and OLSR dynamic routing protocols

and ideal transfer version via mesh-network without

interference produced by trafﬁc route. The experi-

ment proved better performance of OLSR dynamic

routing protocol in comparison with B.A.T.M.A.N. in

bulk data transfer (over 10 MB); gain in transfer speed

amounts up to 35%. It is connected with the fact

that the amount of interference during the ﬁle down-

load rises proportionally to the ﬁle size and trans-

fer speed. In this case, therefore, it is necessary to

switch as soon as possible to alternative transfer chan-

nel without waiting for main transfer channel recov-

ery as standstill time will substantially inﬂuence the

data transfer volume. B.A.T.M.A.N. protocol, on the

contrary, waits for channel recovery via main channel

without deleting the route from routing table.

It was also found out in the experiment that http-

protocol proves to be a better application layer pro-

tocol at low speed data transfer (128 Kbit/s) for big

ﬁles (over 10 MB). It manifested lower transfer time

by 10 to 30% because http-protocol uses fewer exec-

utive instructions that should be performed before the

ﬁle download. In case of connection failure (inter-

ference) and ﬁle download the protocol should repeat

these instructions.

200

400

600

800

1000

1200

1400

1600

10 100 1000 10000 100000

time, sec (less is better)

file size, KBytes

B.A.T.M.A.N.

OLSR

experimental values without noise

Figure 6: Dependency of ﬁle transfer time on size for

B.A.T.M.A.N and OLSR protocols at 1024 Kbit/s speed

(second series).

Results of the second series demonstrate dynam-

ics of data transfer at mean speed (1024 Kbit/s). Such

speed is typical of bulk data transfer including web-

trafﬁc, online-conferences. The chart presents depen-

dency of data transfer time via http-protocol on ﬁle

size with the application of B.A.T.M.A.N and OLSR

dynamic routing protocols and ideal transfer version

via mesh-network without interference produced by

trafﬁc route. The experiment proved better perfor-

mance of OLSR dynamic routing protocol in com-

parison with B.A.T.M.A.N. in bulk data transfer (over

10 MB); gain in speed transfer makes up to 20%. It

is connected with the fact that amount of interference

during ﬁle download is rising proportionally to the ﬁle

size and transfer speed.

100

120

140

160

10 100 1000 10000 100000

time, sec (less is better)

file size, KBytes

B.A.T.M.A.N.

OLSR

experimental values without noise

Figure 7: Dependency of ﬁle transfer time on size for

B.A.T.M.A.N and OLSR protocols at 11 Mbit/s speed (third

series).

The results of the third series demonstrate dynam-

ics of data transfer at maximum speed (ca. 11 Mbit/s).

Such speed is typical of bulk data transfer including

multimedia and program update. The chart presents

dependency of data transfer time via http-protocol

on ﬁle size with the application of B.A.T.M.A.N and

OLSR dynamic routing protocols and ideal transfer

version via mesh-network without interference pro-

duced by trafﬁc route. The experiment proved better

performance of B.A.T.M.A.N. dynamic routing pro-

tocol in comparison with OLSR in bulk data trans-

fer (over 10 MB); gain in speed transfer makes up

to 50%. It is connected with the fact that the proto-

col is oriented at recovery connections (frequent fail-

ure with same-route recovery). Hence, the route in

question is not deleted from routing table (even if the

neighboring one is lost) but resumes being active. It

means that during interference time the data transfer

route is not rebuild and it does not require additional

time to re-calculate the route when the connection re-

establishes as it occurs with OLSR protocol.

The results of three series of experiments demon-

strate that small data packets independent on a cho-

sen dynamic routing protocol take shorter time for

delivery even at low speed communication channels

or server limitations on data transfer. It allows to or-

ganize eCall emergency messaging or gathering ve-

hicle’s statistics applying the current network infras-

tructure.

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

780

5 CONCLUSION

The research resulted in assessment of different dy-

namic routing protocol applicability in conditions of

the vehicle movement in urban environment for dif-

ferent data volumes. OLSR dynamic routing protocol

proved to be the best at low speed (128 Kbit/s) for

ﬁle transfer of over 10MB, with the difference of ﬁle

download time making up to 35%. HTTP-protocol is

a recommended application layer protocol to transfer

data at low speed with transfer time shorter by 10 to

30%. OLSR protocol also proves the best at middle

speed (1024 Kbit/s) and for ﬁles over 10 MB; the gain

in ﬁle download speed is up to 20%.

B.A.T.M.A.N dynamic routing protocol is prefer-

able at high speed data transfer (11 Mbit/s); the gain

in ﬁle download speed amounts to 50%. It can be

concluded according to the experimental results that

B.A.T.M.A.N is less preferable for frequently chang-

ing network topology. It is more oriented at con-

nections recovery (frequent failure with same route

recovery) and is worth applying at high speed data

transfer.

According to the study it can be concluded that

the protocol B.A.T.M.A.N. performs better in situa-

tions where the cars move in organized groups. These

conditions, frequent disconnects do not immediately

rebuild the network topology and data transmission

quickly restored when the neighbors are again within

sight of the transmitter. OLSR shows a good perfor-

mance when used in an urban environment, in a situa-

tion where the available set of routes for data transfer,

and it is unlikely that the driver returns to the network

recently vacated. Thus, it is necessary, depending on

the driving conditions, switch between the dynamic

routing protocols, or the simultaneous use both dy-

namic routing protocols. Alternate use of two rout-

ing protocols will increase the amount of trafﬁc the

service transmitted via the communication channels

(increase the load on the mesh-network).

ACKNOWLEDGEMENTS

This research was supported by a grant from the Ford

Motor Company. This paper funded by RFBR grant

13-07-12106.

REFERENCES

Abolhasan, M., Hagelstein, B., and Wang (2009). Real-

world performance of current proactive multi-hop

mesh protocols. Communications, 2009. APCC 2009.

15th Asia-Paciﬁc Conference on, pages 44–47.

Chakraborty, S. and Nandi, S. (2013). Ieee 802.11s mesh

backbone for vehicular communication: Fairness and

throughput. Vehicular Technology, IEEE Transactions

on, 62(5):2193–2203.

Glazunov, V., Kurochkin, L., Kurochkin, M., and Popov, S.

(2013a). Instrumental environment of multi-protocol

cloud-oriented vehicular mesh network. In Ferrier, J.-

L., Gusikhin, O. Y., Madani, K., and Sasiadek, J. Z.,

editors, ICINCO (1), pages 568–574. SciTePress.

Glazunov, V., Kurochkin, L., Kurochkin, M., Popov, S.,

and Timofeev, D. (2013b). Road trafﬁc efﬁciency

and safety improvements trends. In Ferrier, J.-L.,

Gusikhin, O. Y., Madani, K., and Sasiadek, J. Z., edi-

tors, ICINCO (2), pages 439–446. SciTePress.

Hunkeler, U., Truong, H. L., and Stanford-Clark, A. (2008).

Mqtt-s; a publish/subscribe protocol for wireless sen-

sor networks. In Communication Systems Software

and Middleware and Workshops, 2008. COMSWARE

2008. 3rd International Conference on, pages 791–

798.

Jayasooriya, I., Nallaperuma, T., Herath, U., Ranasinghe,

S., Liyanage, M., Tennekoon, R., and Rupasinghe, L.

(2013). Decentralized peer to peer web caching for

mobile ad hoc networks (icache). In Computer Sci-

ence Education (ICCSE), 2013 8th International Con-

ference on, pages 218–223.

Klein, A., Braun, L., and Oehlmann, F. (2012). Perfor-

mance study of the better approach to mobile adhoc

networking (b.a.t.m.a.n.) protocol in the context of

asymmetric links. In World of Wireless, Mobile and

Multimedia Networks (WoWMoM), 2012 IEEE Inter-

national Symposium on a, pages 1–7.

Kulla, E., Hiyama, M., Ikeda, M., and Barolli, L. (2012).

Performance comparison of olsr and batman routing

protocols by a manet testbed in stairs environment.

Comput. Math. Appl., 63(2):339–349.

Laouiti, A., Muhlethaler, P., Sayah, F., and Toor, Y. (2008).

Quantitative evaluation of the cost of routing protocol

olsr in a vehicle ad hoc network (vanet). In Vehicu-

lar Technology Conference, 2008. VTC Spring 2008.

IEEE, pages 2986–2990.

Murray, D., Dixon, M., and Koziniec, T. (2010). An ex-

perimental comparison of routing protocols in multi

hop ad hoc networks. In Telecommunication Net-

works and Applications Conference (ATNAC), 2010

Australasian, pages 159–164.

Pereira, A., Pharris, P., Oursler, J., and Lauf, A. (2012).

Distributed aerial reconnaissance ad-hoc networking

protocol. In MILITARY COMMUNICATIONS CON-

FERENCE, 2012 - MILCOM 2012, pages 1–6.

Vilhan, P. and Hudec, L. (2013). Building public key in-

frastructure for manet with help of b.a.t.m.a.n. ad-

vanced. In Modelling Symposium (EMS), 2013 Eu-

ropean, pages 566–571.

Zaborovski, V. S., Chuvatov, M., Gusikhin, O. Y., Makkiya,

A., and Hatton, D. (2013). Heterogeneous multipro-

tocol vehicle controls systems in cloud computing en-

vironment. In Ferrier, J.-L., Gusikhin, O. Y., Madani,

K., and Sasiadek, J. Z., editors, ICINCO (1), pages

555–561. SciTePress.

AnExperimentalComparisonofDynamicRoutingProtocolsinMobileNetworks

781

Zaborovskiy, V., Lukashin, A., Popov, S., and Vostrov, A.

(2013). Adage mobile services for its infrastructure.

In ITS Telecommunications (ITST), 2013 13th Inter-

national Conference on, pages 127–132.

Zaborovsky, V., Gorodetsky, A., and Muljukha, V. (2009).

Internet performance: Tcp in stochastic network envi-

ronment. 1st International Conference on Evolving

Internet, INTERNET 2009, pages 21–26. cited By

(since 1996)0.

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

782