A Data Collection Scheme with Multi-Agent Based Approach for VSNS

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 03 | Mar -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 2468
A Data collection scheme with multi-agent based approach for VSNs
Shilji k c1
1P G scholar, Dept. of Information and technology, Govt. engineering college Barton Hill Trivandrum, Kerala, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - VSNs (vehicular sensor networks) is a newly
progressive networking system of VANET (vehicular ad-Hoc
networks). And it’s also considered as a subgroup of VANET.
This paper focusses upon improving the energy efficiency and
data collection in urbannetworksusingagentmechanism. The
important components like sensor nodes (vehicles) and
roadside units together comprise a Vehicular Sensor Network
(VSN). Vehicles are equipped with sensors for the collection
and dissemination of data over thenetwork. Theroadnetwork
is divided into small groups of nodes called clusters. It is a
challenging task to manage the networks in a dynamic
environment due to the various characteristics like high
mobility of vehicles, high relativevelocity, differentbehaviorof
drivers and occurrence of unpredictable link failures. To solve
these issues an agent based communication is one of the
potential solution approach. But singlemobileagentsareused
in small road segments and have energy constraint problems
as well as unnecessary communication delay. Thismechanism
focusses on a suitable and an efficient aspect to collect sensor
readings from VSNs. It also shows the pros and cons of both
cluster based and agent based solution approach and also
discovers a new and improved scheme. It deploys multiple
mobile agents for collecting sensor readings from the given
urban areas. Mobile agents serve as an emerging potential
solution to efficiently collect sensor readings in VSNs. It
achieves remarkably higher network coverageonurbanareas
and reduces energy and delay constraint issues based on the
following assumptions. Finally to overcome the short comings
of an agent based scheme an improved multi agent based
three layer network architecture approach is also discussed.
Key Words: Mobile agents, Road segment interest (RSI),
Vehicular sensor networks (VSN)
1. INTRODUCTION
Vehicular Ad Hoc Networks (VANETs) [5] are discovered
from the principles of mobile ad hoc networks (MANETs).
Similarly VSN (Vehicular sensor networks) is an emerging
networking system from the base concepts of VANET’s. It
creates a wireless network for data exchange in the domain
of vehicles within the road network. These are the key
components of intelligenttransportationsystems(ITS).VSNs
support a wide range of applications from simple one hop
information collection and dissemination to multi-hop data
collection and dissemination of messages over vast
distances. VSN (Vehicular sensor networks) is an extension
of VANET, in which it consists of sensor nodes as vehicles
and road side units (RSUs) [6]. VSN provides a sensing
platform in which vehicles and RSUs communicates by
collecting and disseminating theinformationthroughoutthe
network. Most of the aspects in Vehicular ad hoc networks
and VANETs are same, but the details of both differ. Vehicles
interact with the road side units and it can be characterized
in an accurate way. Similarly the vehicles mobility range is
restricted in an organized fashion. Vehicular Sensor
Networks have been spotlighted for many applications such
as the driving safety, driving efficiency, and data services in
the road networks [7]. The sensors in the vehicles play a
vital role in monitoring the road surfaces to detect obstacles
and any road hazards. The monitoring is performed by
collecting the data from various motion sensors (e.g.,
accelerometer), and obstacle detection sensors (e.g.,
ultrasonic and laser sensors) etc. Vehicles are embedded
with sensors for the driving safety, particularly for
monitoring the roadsurface andidentifyingtheroadhazards
are critically important, and those sensing pieces of
information can be shared among the vehicles to avoid
possible dangerous circumstances and situations.
Similarly VSN provide a wide range of services in which
vehicles are equipped with some short-range and medium-
range wireless sensors. In VSN there exists three types of
communication: Vehicle-to-Vehicle andVehicle-to-roadside
units, and inter road side communication .The road side
units might be a sensor oriented base station. Supposeatthe
mid-night in some rural area, a vehicle has a very important
data packet (i.e. detection of an accident) which should be
forwarded to the following vehicles immediately. The
probability of low density of vehicles in the rural areas at
mid-night is very high. Consequently, in this situation the
packet has to be broadcasted, and arrival of the following
vehicles in the accident area is unavoidable. To overcome
these issues there are two types of sensor nodes that are
considered. Some are embedded in the vehicles known as
vehicular sensor nodes (VN) and others are deployed in
predetermined distances besides the highway road, known
as Road Side units. We can have some Base Stations (BS)
such as Police Traffic Station. The base stations considered
may be either stationary or mobile. The sensor nodescollect
the real data such as vehicle’ velocity and position are
forward towards road side units. On the other hand, for

sending a query from RSU’s, sensor nodes receive it and
rebroadcast the packet to the vehicles in its coverage area.
Types of vehicular communications are shown in fig -1.
Fig -1: Wireless Communication System
In VSNs due to the highly varying network topology data
collection in dynamic networks getting challenging task. To
efficiently collect sensor readings in Vehicular sensor
networks mobile agentsarethepotential solution[8][2]. The
term mobile agent refers from the agent technology system,
it is a software component which migrates independently
throughout the network. It can interact with other agents in
the network and learn the required information and then
perform its own tasks [1].Sensor nodes are used to provide
variety of services in VANETs [16]. Hence each vehicle as
well as the road side sensors monitor the traffic situation,
such as density, average speed, etc. The basic idea behind
mobile agents is to move code across a network from one
node to another.
To address the problems of VSNs such as network
management, service provisioning, information collection
and distribution which is intrinsically difficult, a solution is
proposed in which vehicles integrates with mobile agents
and norm-aware agents [2].The basic idea behind mobile
agents is to move code across a network from one node to
another. Similarly the behavior of any interacting set of
agents is regulated by norms can be considered a norm-
governed system, where a norm is a rule that prescribes
organizational conceptslike (institutional)obligation,power
permission, sanction,andothermorecomplexrelations. This
network implementation in VANETs extends into VSNs
under dynamic topology in which the sensor nodes are used
to provide variety of services in VANETs. This study is
focusing on to solve one of the issue of VSN’s that is, efficient
data collection on urban areas and to provide a solution
approach.
The rest of this paper is organized as follows.First,section
2 describes the literature review. This is followed by
explanation about different VSN data collection schemes.
Section 3 addresses VSN’s three layer network
CommunicationArchitecture.Finally,section4concludesthe
paper.
2. LITERATURE SURVEY
There are two aspects which are involved for the data
collection in VSNs- routing related and data-related [10].
The routing-related aspect is the routing of raw sensor
readings, while the data-related aspect is for the coding and
compression of aggregated data, and it focuses to reducethe
data transmission overhead.
The paper focusses on the study about how to effectively
gather raw sensor readings from the urbanarea in whichthe
vehicular density is very much higher. All the proposed
methods that address the issues and challenges of data
collection in VSNs. These works are mainly fall into two
models depending on whether raw sensor data is collected
on demand: push-based or pull-based [11]. In the push-
based model, every vehicle periodically broadcasts the
gathered sensor readings to its one-hop neighbors and the
collected information is disseminated to all other regions or
collected by a data sink. Generally this model ispreferredfor
traffic safety applications.
Due to the dissemination of collected data over the entire
network will cause higher communication overhead. To
reduce this a common technique adopted in the pull-based
model is to divide the entire vehicles in the given area into a
set of groups called clusters in which one vehicle(thecluster
head) collects sensor data from all other vehicles in that
cluster and exchanges the information with the next cluster
by communicating via cluster heads. In [12], the authors
propose the Clustered Gathering Protocol (CGP) in which
vehicles collect data within a specified region by combining
cluster-based data dissemination mechanism with
geographical and opportunistic approaches. In the scheme,
vehicles are grouped into a number of clusters and for each
cluster, the cluster head is elected by using a back-off based
algorithm. Cluster heads manage the collection and
dissemination of data’s over the network as well as it
forwards to the corresponding RSUs therefore reducing the
required data transmission bandwidth.However,itcreatesa
single point-of-failure due to the unique choice of cluster
head collecting the information within each cluster and the
efficiency of data gathering depends on the stability of the
clusters.
Another problem is that the collisions caused by intra-
cluster communications, since the cluster head collects the
required data’s from all cluster members .To address this
problem, the TDMA Access Control Algorithm (TACA) was
proposed in [13] to regulate the medium access of cluster
members, where the given road segment of a cluster is
partitioned into small grids, and for the transmission each
grid is assigned with a time slot. However, when vehicle

density in the cluster is low the algorithm results in
unnecessary latency. That is each vehicle in the cluster must
need to wait for its allocated time slot to respond its data to
the cluster head. In [14] and [15], the authors proposed to
adopt SDMA or TDMA to control medium access of vehicles,
where the cluster heads are responsibleforthespaceortime
slot allocation and their synchronization. However, due to
the mobility of the vehicles cause the instability of cluster
heads similarly the coordination of communication is costly
in many cases.
To solve these communication issues a new scheme is
proposed an agent based data collection scheme [1]. In
which this approach is based on deploying a single agent to
collect the data within the specified region. In this scheme a
remote data sink initiates the data packet called an agent,
which specifies the coordinates of the road segment. It
migrates among the vehicles and collects the requesteddata
item. It combines an agent migration algorithm with the
termination algorithm to achieve the data collection as well
as reducing the communication overhead. Where all the
vehicles are in single hop communication region. However
the problem of this scheme shows that the transmission
power of each vehicle is very much higher due to the single
hop communication, hence it reduces the lifetime of each
vehicles. Similarly due tothe lossylinksthemaintenance will
cause communication delay.
3. DISCUSSIONS AND SOLUTIONS
Based on our study on various data collection schemes of
VSN we found that an agent based data collection
mechanism is a better option for urban areas. But it can
cover small segment of network area with a single hop
communication. To achieve urban area data coverage ratio
one of the solution method is todeploysmultipleagentsover
the given network area instead of single agent mechanism.
Hence MAS (mobile agent systems) based vehicular
information vehicular sensor network in urban area isbeing
proposed to solve the issues related to single agent scheme.
Here, all vehicles are considered to be part of a VSN. Each
vehicle as well as the road side sensors monitors the traffic
situation, such as density, average speed, temperature,
pressure etc. Based on these assumptions, in this section, a
multi agent-based approach is considered, and also it
incorporates with a clustering scheme to solve the existing
network problems in VSN is discussed.
3.1 Advantages of using agents
 Flexibility: Agents provides flexible services for the
information collection and dissemination over the
vehicular sensor network.
 Adaptability: Agents adapt to variednetwork conditions
such as mobility of vehicles, critical traffic situations,
weather conditions, road changes, etc.
 Maintainability: Agents provides less maintainable
issues over the communication link loss and critical
events.
 Survivability: Wireless networks are very much
vulnerable to failures. Similarly, in VSNs,itmaybethata
node fails to comply with the system specifications,
either by design, by accident or from necessity. Agents
are Dealing with such non-compliant behavior, can also
be addressed.
Considering all the propertiesoftheagentsystems,a multi
agent based data collection scheme by extending the agent
technology in VANETs into VSNs is required to implement.
Therefore, our solution is to propose a three-layer network
architecture comprising: 1) mobile agents; 2) static agents.
The mobile agent layer focusses in the implementation
decisions made by static agents. The mobile agents operates
on the domain of network whereas the static agents for
applications. Thus an idea is proposed to converge both
agent scheme and clustering scheme for the efficient data
collection and dissemination over the network. In the
following sub-section, we elaborate further on this novel
network architecture. Three layer architecture shown in
following Fig -2.
Fig – 2: Three layer Network Architecture
The system comprises the vehicles which are
communicating themselves on the road which are within a
given cluster (VSN). A cluster is formedbasedonradio range
coverage of vehicles and the road-side base station.Herethe
given urban area is divided into a set of clusters in which
each cluster comprises of cluster head called base stations
and the clusters may communicate by using these base
stations.
The function of each cluster head is to manage and
coordinate the communications among vehicles as well as
broadcasts the collected information’s to the corresponding
RSU’s. Cluster heads are required toreducethetransmission

bandwidth similarly it can reduce the transmissionpower of
vehicles.
The sink node refers to a transport station which initiates
all the network communications and manages the
performances. It consists a vehicle agency for the
synchronization of data’s.
From the proposed network architecture, the information
collection and network management model can be
categorized into four parts: vehicle tovehicle,vehicleto base
station, cluster to cluster and cluster to sink node [2]. There
are several agents are used for the data communication
between vehicles within the clusters. Following shows the
different types of communications and thesetofagentsused
for.
3.2. Vehicle to Vehicle Communication
Each of the participating vehicles withintheclusterwould
be equipped with a set of agents such as Vehicle Manager
Agent (VMA), Alarm Agent (AA), and Service Discovery
Agent (SDA). These agents are responsible for collectingand
disseminating traffic information. The vehicleagencywithin
the sink node also consists of a traffic knowledge base(TKB)
which is used for communication among agents.
VMA: It is a static agent deployed at every vehicle and
synchronizes the interaction of all the agents within the
vehicles.
AA: It is a mobile agent moves across the vehicles in the
network by creating its clones, and it is used for the critical
information dissemination. It also informs the VMA and
updates the TKB.
SDA: It is a mobile agent which travels in the entire network
for the services requested by the vehicle user and search for
the services such as road maps,traffic densitymaps,Internet
services, and location aware services (commerce,
entertainment, parking, fuel stations, etc.). The agent also
updates the TKB with services discovered.
TKB: It consists of information of critical events such as
accidents, traffic density and the services available in the
vehicle, services accessed, recently accessed road maps and
it updates regularly.
3.3. Vehicle to Base station
A fixed infrastructure comprised of a number of base
stations strategically placed in close proximity to the
highways is necessary to facilitate the upload/download of
data from/to the vehicles. The information from the sensors
is the input to the base station. Information could be traffic
density, vehicle types, adverse road conditions, etc. Each
cluster consists a base station. The base station agency
comprises of following components. Base Station Manager
Agent (BSMA), Service Agent (SA), Advertisement Agent
(ADA) and Cluster Knowledge Base (CKB).
BSMA: It is a static norm-governed agent deployed at each
base station which maintainsandsynchronizesall theagents
that are associated with base station.Itregularlyupdates the
CKB with the visited vehicles information and services by
interacting with VMA of each vehicle.
SA: It is a static agent responsible for collecting the
information and services from the service providers of the
cluster and regularly updates the CKB.
ADA: It is a mobile agent which travels in the network and
informs the visited VMAs about the collected information’s
for the requested vehicle user.
CKB: It regularly updates with information such as critical
events within visited vehicle information, traffic density
maps, road conditions, location aware services,
advertisements, etc.
3.4. Cluster to Cluster
Clustering provides a method to disseminate the traffic
information as well as provide varieties of services with less
communication overhead. The whole networkisdividedinto
groups of vehicles called clusters, in which a cluster head is
selected for the overall communication management. These
clusters are self-management groups thatcontinuallyadapts
the changing network topology and new cluster
configurations.Communicationbetweentheclusterswilltake
place with the help of BSMAslocated in base stationsthatare
fixed on the road side or within the cluster heads.
3.5. Cluster to Sink Node
Sink node consists of complete information of the
transportationinfrastructurewhichiscollectedfromvarious
cluster BSMAs. BSMAs of the clusters communicate with the
Sink node manager. Sink node manager periodically
constructs an overall picture of the road ways in terms of
traffic, critical events, road conditions, etc., and constructs a
road map and distributes to the BSMAs.
A multi-agent approach [9] [17] is presented in this paper
is to deal with the shortcomings of single agent
communication in VSN.As well as multi agent approach is
efficient if the network area is extended into an urban road
network. All the agentpropertiesareinheritsbymultiagents
and the system responds well in urban networks. In an
urban road networks the density of vehicle is very much
higher, it will cause communication link loss and data
interferences between the vehicles. An agent technology is
an efficient solution to observe and learn the desired
environment andcollecttherequiredinformation,thereafter
optimize communication among the vehicles, hence find the
best routes to reduce the network traffic. Various routing
protocols are available for the information routing. The
routing protocols are classified into various categories such
as cluster based routing protocols, topology based routing
protocols, position based routing protocols, geo cast and
broadcast routing protocols.
In this scheme for the routing of information within the
network the cluster based routing protocol is considered. In

cluster based routing protocols, the entire network is
divided into a number of small partitions called clusters.
Clustering improves the performance of the network, hence
can be utilized in numerousvehicularapplications[4].Inthis
approach instead of communicating vehicles directly to the
BSMA a cluster head is elected for the communication
between the clusters. In which the clusterheadcontinuously
collects the data over the network and broadcasts to the
corresponding BSMA’s. The limitations of cluster baseddata
collection mechanism are that the single point failures and
intra-cluster problems are addressed by using multi agents.
All nodes always monitor their signal strength to the cluster
head. If the signal strength decreasesbetweenthenodes and
cluster head means the current cluster head has no longer
been able to attach with all nodes. So theclusterheadtriesto
handover all the member nodestoa neighboringcluster. The
development of optimized routing system for the agent
technology has been a promising solution.
Vehicular communications are formed in mobile
environment of urban and rural areas and works well in
both. Considering the large network with much number of
nodes participating in the communication cause overhead
due to their high mobility, Researches still exist about the
study of feasibility of routing protocols [3].
4. CONCLUSION AND FUTURE WORK
On the basis of these studies it can be concluded that VSN’s
are facing several limitations .In which my study is focusing
on to solve the energy efficiency problem facing by vehicles
in their single hop communication, and the higher
transmission power of vehicles.Thiswill reducethevehicle’s
or the network’s life time. In this paper both cluster based
and agent based data collection schemes are discussed from
which it can be concluded that due to the instability of
cluster heads the agent based data gathering scheme
performs well in a small segment of area instead of cluster
based scheme. But the main drawback of this scheme is that
less energy efficiency and higher transmission power.
Findings of this study indicated that this problem can be
solved by extending the agent based concept into a multi-
agent system with a clustering mechanism that is able to
monitor large urban regions through cooperation among
several agents.
This paper has given a solution approach to the existing
agent collection scheme. So the current work emphasizes on
deploying multiple agent to collectdata withinanurbanarea
network by a three-layer network architecture approach.
Similarly it reduces the transmission power of the vehicles
as well as improving the energy efficiency by multi-hop
communication within the multiple agents. It achieves
energy efficiency and higher data collection under dense
traffics by deploying multiple agents. Similarly the multi
agent based clustering scheme overcomes the limitations of
single point failure and intra-cluster problems. It can be
concluded that multi-agent systems have a greatpotential to
influence the design of future VSN's and their applications.
Multi-agent systems should be providing more adaptability,
and flexibility, for the realization of services and its
applications within next generation VSN's environments.
The future work is the implementation of multi agent
based three layer networking system with a clustering
scheme.
ACKNOWLEDGEMENT
I would like to express my sincere gratitude towards my
mentor Ms. Sangeetha U, Assistant Professor in Department
of Information Technology at GEC Trivandrum, for her kind
co-operation and encouragement which helped me to
understand the vast topic of Vehicular Sensor Networks. I
would like to thank my family and friends as well, for their
continuous support and encouragement.
REFERENCES
[1] Hui Huang, Lavy Libman, Glenn Geers. “An Agent
Based Data Collection Scheme for Vehicular Sensor
Networks,” IEEE International Conference, (2015).
[2] S.S. Manvi, M.S. Kakkasageri, Jeremy Pitt, Alex Rathmell,
“Multi-Agent Systemsasa PlatformforVANETs,”pp,2-7.
( 2010).
[3] Amira Harrabi, Walid Chainbi, Khaled Ghedira.“AMulti-
Agent Approach For Routing On Vehicular Ad-Hoc
Networks,” The 4th International Conference on
Ambient Systems, Networks and Technologies, (ANT
2013).
[4] Rasmeet S Bali , Neeraj Kumar , Joel J.P.C. Rodrigues b.
“Clustering in vehicular ad hoc networks: Taxonomy,
challenges and solutions,” Vehicular Communications
@science Direct, may (2014).
[5] Moath Al-Doori, Ali H Al-Bayatti, “A comprehensive
survey on vehicular Ad Hoc network,” Article In Journal
Of Network And Computer Applications,January(2014)
[6] G.G Md. Nawaz Ali, “Co-operative data accessinmultiple
Road Side Units (RSUs)-based Vehicular Ad Hoc
Networks (VANETs),” ResearchGate publications
Conference Paper, January (2011).
[7] Uichin Lee, Eugenio Magistretti , Biao Zhou,MarioGerla,
Paolo Bellavista, Antonio Corradi, “MobEyes: Smart
Mobs for Urban Monitoring with a Vehicular Sensor
Network,” IEEE Wireless communications (vol. 13) pp.
1-3. October (2006).
[8] Oscar Urra, Sergio Ilarri, Thierry Delot and Eduardo
Mena, ”Mobile Agents in Vehicular Networks: Taking a
First Ride,” Springer Berlin Heidelberg publications,
(2010).
[9] S.S. Manvia , M.S. Kakkasageri and Jeremy Pitt,
“Multiagent based information dissemination in
vehicular ad hoc networks,” ACMpublications,journal in
mobile information systems, (vol.4) pp. 363-389 dec
(2009).
[10] B. Yu, J. Gong, and C.-Z. Xu, “Catch-up: Adata aggregation
scheme for VANETs,” ACM international workshop on
Vehicular Internetworking, New York, NY, Sep. (2008).

[11] T. Nadeem, P. Shankar, and L. Iftode, “A comparative
study of data dissemination models for VANETs,” IEEE
Annual International Conference on Mobile and
Ubiquitous Systems: Networking and Services
(MobiQuitous), San Jose, CA, Jul. (2006).
[12] I. Salhi, M. O. Cherif, and S.-M. Senouci, “A new
architecture for data collection in vehicular networks,”
IEEE International Conference on Communications
(ICC), Dresden, Germany, Jun. (2009).
[13] W.-R. Chang, H.-T. Lin, and B.-X. Chen,:Trafficgather,“An
efficient and scalable data collection protocol for
vehicular ad hoc networks,” IEEE Consumer
Communications and Networking Conference (CCNC),
Las Vegas, NV, Jan. (2008).
[14] B. Brik, N. Lagraa, H. Cherroun, and A. Lakas, “Token-
based clustered data gathering protocol (TCDGP) in
vehicular networks,” International Conference on
Wireless Communications and Mobile Computing
(IWCMC), Sardinia, Italy, Jul. (2013).
[15] B. Brik, N. Lagraa, M. B. Yagoubi, and A. Lakas, “An
efficient and robust clustered data gathering protocol
(CDGP) for vehicular networks,” ACM international
symposium on Design and analysis of intelligent
vehicular networks and applications (DIVANet), New
York, NY, Oct.(2012)
[16] Mohammad Jalil Piran, G. Rama Murthy, G. Praveen
Babu, “vehicular ad hoc and sensor networks;principles
and challenges,” International Journal of Ad hoc, Sensor
& Ubiquitous Computing (IJASUC) (Vol.2)pp-38-40June
(2011).
[17] Samira Harrabia, Walid Chainbib, Khaled Ghedirac, “A
Multi-Agent Approach For Routing On Vehicular Ad-
HocNetworks,” The 4th International conference On
Ambient systems, Networks and Technologies (ANT
2013).
Received B-tech degree in
Computer ScienceandEngineering
from Cochin University of Science
And Technology (CUSAT) in 2013.
In 2016 completed M-Tech in
Network Engineering from Kerala
University, and presented national
and international conferences in
these areas.

A Data Collection Scheme with Multi-Agent Based Approach for VSNS

More Related Content

What's hot (20)

Similar to A Data Collection Scheme with Multi-Agent Based Approach for VSNS (20)

More from IRJET Journal (20)

Recently uploaded (20)

A Data Collection Scheme with Multi-Agent Based Approach for VSNS