An Opportunistic AODV Routing Scheme : A Cognitive Mobile Agents Approach

International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.8, No.1/2/3, June 2017
DOI : 10.5121/ijasuc.2017.8301 1
AN OPPORTUNISTIC AODV ROUTING SCHEME: A
COGNITIVE MOBILE AGENTS APPROACH
Seema B Hegde1
and B.Sathish Babu2
1,2
Department of Computer Science and Engineering,
Siddaganga Institute of Technology, Tumkur, Karnataka, India.
3
Pallapa Venkatarm,
PET Unit, Electrical Communication Engineering,
Indian Institute of Science, Bangalore, Karnataka, India.
ABSTRACT
In Manet’s Dynamics and Robustness are the key features of the nodes and are governed by several routing
protocols such as AODV, DSR and so on. However in the network the growing resource demand leads to
resource scarcity. The Node Mobility often leads to the link breakages and high routing overhead
decreasing the stability and reliability of the network connectivity. In this context, the paper proposes a
novel opportunistic AODV routing scheme which implements a cognitive agent based intelligent technique
to set up a stable connectivity over the Manet. The Scheme computes the routing metric (rf) based on the
collaboration sensitivity levels of the nodes obtained based through the knowledge-based decision. This
Routing Metric is subsequently used to set up the stable path for network connectivity. Thus minimizes the
route overhead and increases the stability of the path. The Performance evaluation is conducted in
comparison with the AODV and sleep AODV routing protocol and validated.
KEYWORDS
Mobile Adhoc Network’s, AODV, Opportunistic network, Cognitive agents, Routing factor, Sleep AODV
1. INTRODUCTION
The information diffusion dynamics has become seamless with the increased use of pervasive
computing devices. The MANET is one among the important technologies supporting pervasive
computing by enhancing the mobility in the dynamic environment. MANET's [1] are distributed,
self-configurable, infrastructure less Ad hoc class of wireless networks with no fixed
apparatus. MANET's find applications in many areas ranging from commercial to critical as well
from consumer to the military. MANET's diversified use is due to its features such as robustness
to cope up with dynamic network topology; fluctuating, high bit error rate and various links;
collaboration among the nodes and so on [2]. The typical applications are in dynamic unplanned
situations [3] requiring emergency management such as recovery from disasters; an earthquake,
flood, chemical accidents and so on. MANET's can handle the volatile topology [4] through
dynamic routes. It has hosted routing protocols [5] such as AODV, DSR, DSDV, and TORA and
so on. Even though an investigation says no specific protocol is sufficient for all the scenarios,
AODV [6] is the popular imploded on-demand reactive, table driven protocol.

2
AODV consists of two phases of routing, i.e., route discovery and route maintenance. AODV
discovers routes when required and are maintained just as long as needed. When a node wishes to
send a packet to some destination, on the unavailability of the current path initiates the route
discovery phase by broadcasting route request (RREQ) packet. On comparison, AODV
outperforms DSR protocol [7] [8] in stressful situations having increased mobility and traffic due
to fewer delay requirements.
However, the AODV protocol exhibits several pitfalls that need to overcome for better
connectivity during the information diffusion. The major limitation of the protocol is the routing
overhead [9]. It only maintains the link information between the communication nodes. Link
failures will be more at higher mobility, triggering more route request packets, and thus the
network has to suffer from the high frequency of route discovery phases which degrades the
performance. Mostly, these link breakages are due to lack of resources such as energy, memory,
and width, computing power and the mobility of nodes.
There is a need for a novel scheme that minimizes the route request overhead by identifying the
resource requirement and takes knowledge based decision prior the link breakage. An efficient
resource pooling technique is required to identify, manage and track the resource on time. Thus
the probability of link breakage can be determined and proactively notified to the mobile nodes.
Using the information the node needs to take an intelligent decision towards maintaining the
connectivity without triggering the route re-discovery phase. But due to the dynamic mobility of
the nodes, getting the resourceful node for computation and connectivity will be difficult with
conventional AODV scheme, as it requires an end-to-end connectivity between the source and
destination. Hence a computing paradigm is needed that makes use of the intermittent
connectivity within the network based on opportunistic contacts between nodes. The paper
proposes a novel intelligent scheme to reduce the routing overheads of the AODV scheme using
opportunistic computing technique based on cognitive theory.
1.1 OPPORTUNISTIC COMPUTATION
An opportunistic computing is a kind of pervasive computing [10], which takes the benefits of
randomness and uncertainty in the environment to establish the communication opportunistically
between pairs of diverse devices and applications. The opportunistic computation takes place in
the nodes that are part of an opportunistic network.
1.2 OPPORTUNISTIC NETWORK (ON)
It is the class of MANET's which exploits the contact opportunity among the nodes for the
communication between the source and destination without the end-to-end connectivity [11]. The
favorable circumstances such as interconnectivity, context-activity and so on, create the contact
opportunities.
Formulation of opportunistic network:
Let us consider a graph of MANET, G{V, E}, with ‘V' number of nodes and ‘E' wireless links
among them as shown in figure 1.

3
Figure 1: MANET set represented by graph G {V, E} with ‘n’ number of V and ‘n-1' number of E.
Then H{P, L} be the subgraph of G with P ={ nk} number of nodes and L={ li,j}, where i≠j
wireless links as shown in figure 2, where S represents the opportunistic network area.
Then H is called an opportunistic network over a Euclidean plane ‘S’ iff the affinity of
opportunistic contact between the nodes ni P and nj P is greater than the average degree of
opportunistic contact ( ) between the two nodes as shown in figure 2.
i.e., ,
P {}
{ , } ( ) ( , ) avg
i j i j
if L
H P L n n S if n n





∅ = =
= ∆ ≥ ∆U
∆ denotes the degree of opportunistic contact between the nodes and .
The degree of opportunistic contact of node ni w.r.t its neighboring nodes (k) is, .
Similarly for nj, .
Let D {N, W} the weighted contact graph with N number of neighbor nodes and W weight
matrix= {wij} be the sub graph of H. Weight matrix indicates the relationship strength between
the nodes i and j. The weight wij is a correlated matrix of the parameters contact frequency fij and
the sum of all contact durations tij.
Where, σf and σt are empirical standard deviations and i, j ∈ N.
The wij is transformed to a scalar value Wij using the principal component.
The average degree of opportunistic contact , N is the number of connected neighbors.

4
Figure 2: Formation of an Opportunistic network represented by sub graph H (P, L) within the MANET
graph G (V, E)
1.3 OPPORTUNISTIC CONTACT (OC)
The nodes which are within the reachable region are said to be in opportunistic contact if they
share any one or all of the characteristics such as the common goal, shared context activity, same
background, common destination and so on. The degree of opportunistic contact (∆) indicates the
affinity between the nodes to have an opportunistic contact.
Therefore the Opportunistic contact between the nodes ni and nj is { , } max { }
0
kOC n n
i j ii
= ∆
=
U
where, the degree of opportunistic contact ∆ for OC
n nodes
Number of nodes having the affinity
Total umber of
=
The number of nodes having affinity for OC is given by,
1
( )
n
k
k
a nnrcπ
=
∑ ,
where a(n) represents the affinity of node n towards the OC and is given by,
( ) E { }
S
a n Cn Rn Ln
∈
= ∪ ∪
Ε
a U
Ε denotes the MANET environment with s ⊆ Ε .
Let Cn be the set of context activities in E at time NT.
Let Rn be the set of resource available in E at time NT.
Let Ln be the set of nodes towards same destination in E at time NT.
The ∆ is parameterized using the metric called Opportunistic Contact level (OCL). The various
context activities at the Kth
node will choose the OCL into any one of the 3 set levels within range
of {0-2}, where 0 is the highest degree and 2 is the least.
The jth
context activity of the Kth
node is given by, ( ) ( )
( ) (l , )k K
j ij ij
C k x=
where lij is the link between the nodes ni and nj. xij is the context similarity metric between the
nodes ni and nj. Similarly, Rj(k) and Lj(k) can be computed.

5
The xij is a function of the three context parameters. The age of last context match (Am), which is
obtained from the timestamp attached with each context match. The number of context matches
between the nodes (Nm) obtained from log history. The location traces in the history log gives the
similar context location (Lm).
Let {Tri} be the set of location traces for node ni and {Trj} be the set of location traces for node nj
then,
The contact similarity metric xij
(k)
is given by,
( )
( )
( )
0 j=1,2,....n
if 0 x 0.5 k , then
0.5 x 1 k then
k
ij
k
ij
k
ij
x
if
ε
δ







≥ ∀
≤ ≤ ∀ ∈ Ε
≤ ≤ ∀ ∈ Ε
Where ε denotes the Vagrant context similarity and δ denotes the maximum context similarity.
The computation performed on nodes by opportunistically availing the other resources of the
environment, along with resources available on its device within the opportunistic network is
called opportunistic computation. Here each node exploits the transient unpredicted contacts to
cooperate and share each other's content, resources, and services [11, 12] (M. Conti, 2010).
1.4 COGNITIVE THEORY
Cognitive Agent [12, 13], is an intelligent software agent. It is an autonomous and responsive
entity which reflects the state of information of the environment which is represented using
cognitive terms. Similar to human perceptual experience and understanding of the surrounding
environment, CA's can provide an immediate or a controlled response to the changes in the
environment [14, 15]. Cognitive agents have self-learning, self-improving and self-instructible
ability [17, 18] and therefore, reduce the network overhead. Cognitive agents are suitable for
adaptive, dynamic, context-based behaviourism, decision making, human mobile network
interaction system [16, 19, and 20] and so on.
1.5 PROPOSED COAODV
The proposed Cognitive agents based Opportunistic AODV (COAODV) scheme uses Mobile
cognitive agent (MCA). The mobile nodes in the network embed the MCA with them. The MCA
is having communication with all reachable MCA's track the availability of resources through the
Behavior-Observation-Belief (BOB) cognitive model (b s bsbu et al., 2009) and generates the
recommendations about resource availability based on the routing factor. Upon the
recommendation, the AODV routing protocol will opportunistically route the messages towards
the destination before losing the current connectivity and avoids the triggering of the route
discovery process.
Organization of the paper: The rest of the paper is organized as follows, Section 2 gives some
of the related works, Section 3 discusses the proposed scheme, section 4 illustrates the results,
and finally, Section 5 draws conclusions.

6
2. RELATED WORKS
Some of the research works conducted in AODV routing technique to discuss the pitfalls of the
protocol are mentioned here.
Sung-Ju Lee et al., (2000) proposed a scheme to keep the alternate backup paths during link
establishment to avoid failures without using additional control messages. The
mesh structure uses an alternative route table with the alternate
route utilization mechanism like DSR protocol and tested the performance of AODV protocol.
The mechanism increased the robustness in mobility but degraded at high traffic load due to
multiple collisions.
Naif Alsharabi et al., (2005) proposed a Packet Received Time (PRT) based prediction technique
in AODV route maintenance phase by measuring the power of the received packets to predict the
link breakage in the priory and rebuild the route. The simulation results show that AODV_PRT
delivers more packets at higher mobility, but the end-to-end delay will be more than the original
AODV and remains same as original at lower mobility as well the control overhead will be more.
C. T. Cuong et al., (2014), Proposed routing algorithm MAR-AODV (Mobile Agent- AODV) to
avoid the link breakage due to congestion using mobile agents which are added to AODV for
updating traffic density at each node, thereby improving the network performance.
Xin Ming Zhang et al., (2013), proposes a neighbor coverage-based probabilistic rebroadcast
(NCPR) protocol, to reduce the routing overhead through avoiding the frequent link breakages. It
exploits the dynamically calculated rebroadcast delay metric, rebroadcast probability and the
neighbor coverage knowledge got from 1-hop neighborhood information. The performance
evaluation shows it reduces the redundant packet retransmissions and alleviates routing overhead.
But, the protocol has considered only the packet collision not connectivity loss due to the random
change in the topology.
D.G. Reinaa et al., (2012), have shown the importance of MANET's in the disaster management
scenario. It proposes a computational approach using reach ability metric to optimize the
connectivity through modeling the MANET's mobility during disaster using a genetic algorithm
for both Intra and intercommunication.
Tahar abbes Mounir et al., (2013), propose sleep-AODV protocol in the disaster scenario to keep
the connectivity during the communication. Based on the mobility of the nodes the energy
consumption rate is divided into three periods and limits excessive routing messages. However,
the performance evaluation shows the degradation in the functionality under high-level traffic due
to more alternate paths.
C. A Ntuen et al., (2011), give a framework to simulate the MANET's network as an intelligent
agent using the cognitive and behavior based modeling in the battlefield scenario. The proposed
scheme uses an OODA (Observe, Orient, Decide, Act) model to represent the MANET's
intelligent behavior. The paper provides a foundation for modeling agent behaviors in a way
plausible on human behavior.

7
3. PROPOSED COGNITIVE AGENTS BASED OPPORTUNISTIC AODV ROUTING
SCHEME (COAODV)
The COAODV, like AODV is a reactive on demand scheme. It has both route discovery and
route maintenance phase. The difference is, the path is set up only with those nodes having the
resource availability level more than the required threshold so that the link breakage due to
resource starvation is not found and reduce the often triggering of the route discovery phase.
Once the path is set up, the message will be forwarded opportunistically towards the destination
(Without the End-to-End connection between source and destination) and reduce the dropping of
packets.
3.1. COAODV scheme architecture
This section gives the architecture of the COAODV scheme by discussing its functional
components and their working.
The design of COAODV consists of a mobile cognitive agent on the node along with the Contact
Analyzer (CA) and the Resource Record (RR) as shown in figure 3.
Figure 3. Architecture of the COAODV
The contact analyzer classifies the external parameters obtained from the surrounding
environment and group the close contact nodes to different levels of opportunistic contact based
on their level of sensitivity towards the formation of opportunistic contact between the nodes. The
Resource Record (RR) is a tuple stores the resource information such as <Type, ID, total>
connected to the node where Type signifies the resource type, ID gives its unique identification
and total give the number of such resources.
The main components of MCA include the belief formulator logic, belief record, route handling
logic and the opportunistic contact level obtained from contact analyzer. MCA's of a node will be
monitoring the resource availability of the host node as well as stores the resource capabilities of
the neighbouring nodes in the form of beliefs. The belief formulator logic generates these beliefs

8
and stores in the belief record. The route handling logic will initiate a route discovery by flooding
Route Request (RREQ) packets.
The opportunistic contact level and the belief generated together will set the parameter called
collaboration sensitivity level (CSL); The ability of the node to collaborate with the host node to
coordinate the communication. Using the CSL the reliability of the connection is computed using
the parameter called routing factor (RF) to set up the path.
3.2 Components of COAODV
The COAODV architecture consists of the following components.
3.2.1 Contact Analyzer
CA will analyze the sensitivity of the incoming transactions called contact parameters (CP) to
find whether the close contact node can form an opportunistic contact with the node. It extracts
the multiset of context activities ‘C i ‘having the Multiplicity m (c) from the CP and compute the
OCL as shown in algorithm 1. The obtained OCL will be concatenated with the beliefs got from
the belief record to get a CSL value.
Algorithm 1: Working of contact analyzer
1: Begin
2: Initialize the session
3: Initialize the Log.
4: INPUT: Contact parameters CP;
5: OUTPUT: Opportunistic contact levels grouped into different groups.
6: Set multiset S to ø, contains the set of context-activity [C , C , C ]
1 2
m n tS
n
= ⋅⋅⋅⋅⋅⋅⋅⋅ each occurring m, n· · · · · ·t
number of times.
7: while Not end of session do
8: Capture the contact parameters CP externally and from log.
9: for each time multiset S≤ n do
10: Extract the context activity Cn to multiset S.
11: End for
12: Set multiset Q to ø, contains the set of resource availability
13: [R , R , R ]
1 2
m n tQ
n
= ⋅⋅⋅⋅⋅⋅⋅⋅ each occurring m, n· · · · · ·t number of times.
14: End for
15: Set multiset V to ø, contains the set of node links [L , L , L ]
1 2
m n tV
n
= ⋅⋅⋅⋅⋅⋅⋅⋅ each occurring m, n· · · · · ·t
number of times.
16: End for
17: Set T to ø, set of all the multisets, i.e., T=[s, Q, V]
19: compute the various combination of contact opportunities
( )
( , ) = = 2
! !0
i T iO T n
n T nn
∑
−=
13: Based on the intensity of contact 2i
output will group in to different opportunistic contact levels (OCL)
ranging from 0 to 2
14: end while
15: Return OCL
16: Periodically refresh Log
17: End

9
3.2.2 Belief Formulator Logic
The logic is based on the BOB (Behavior-Observation-Belief) model [B S babu, 2009]. The three
constructs namely Behaviors identifier, Observations generator, and Beliefs formulator are put
together to form a Belief formulator logic as shown in figure 4.
The belief formulator will formulate the belief over the availability of resources on a node in an
ongoing session. The logic makes use of three essential knowledge quantifiers called Behaviour
(The actions or reactions of nodes while formulating and executing the transactions), Observation
(The summarization of various behaviors exhibited by the mobile nodes during transaction
execution) and Belief (Representation of information about the world or an entity from the
obtained observation perception). The BOB model will capture the behaviour parameters from
external environment from which it deduce the observations and generate the beliefs. It works as
in algorithm 2.
The belief record stores the belief. The beliefs generated are classified based on the availability of
resource in the host node. The beliefs are grouped into any one of the four classes, Casual, Slack,
Patron or Vargant based on the rate of mobility and resource availability of the host node as
shown in Table 1. The beliefs need not be mutually exclusive.
Figure 4: Belief formulation logic

10
Algorithm 2: working of Belief formulator
Table 1: Classes of Beliefs
Class Characteristic
Patron High profile i.e., capable to give large amount of resources with High mobility
node
Casual Low profile i.e., capable to offer small amount of resources with High mobility
node
Slack high profile with low mobility node
Vargant low profile with low mobility node
Belief Record is a tuple <Node ID; Belief generated> which stores the generated beliefs with the
corresponding node ID. The belief generated by the MCA running on the host node will be stored
as self-belief. Beliefs obtained from all the reachable neighbouring nodes are stored along with
their node ID.
These beliefs with the same collaboration sensitivity level are exchanged among reachable nodes
on demand to make the decision about having collaboration with those nodes when required in
the form of resource request and reply.
3.2.3 Route handling Logic
The logic obtains the Belief (BL) and the Opportunistic contact level (OCL) as the inputs.
1: Begin
2: Initialize the session
3: Initialize the Log.
4: INPUT: behavior parameters bh;
5: Intermediate outputs: Behavior (BH), Observation (OB)
5: OUTPUT: Generated belief BL.
6: Set the belief set BF to ø, contains the set of beliefs generated in belief formulator
7: initialize Behavior Identifier (BI)
8: while Not end of BI do
9: Capture the behavior parameters bh externally and from log.
10: Compute the weightage αbhk for bhk
11: αbhk = f(bhk; bhl;m)
12: BHi = f(αbhk ;Current value of (bhk);Max: possible value of(bhk))
13: end while
14: Behavior set BP=BP U BHi
15: initialize Observation generator (OG) with
16:for all observation obi , select favorable BHj
17: while set OB ≥ 0 do
18: OB = obi U OBi occurred most no. of times during the session.
19: end while
20: Initialize the Belief generator (BG)
21: Pass OB to BG.
22: for all BLi€ BL, select favorable OBj
23: BLi = BLi U OBj
24: while BL ≥ 0 do
25: BL = BL U BLi
26: end while
27: Return BL
28: Periodically refresh Log
29: End

11
Figure 5: The CSL derived from Cartesian product of OCL and Belief
It is responsible to set the nodes collaboration sensitivity level (CSL) of the host node.
The Cartesian product of the two inputs will generate the CSL as shown in figure 5. The
CSL can be of levels Highly recommended (2), Recommended (1) or Not recommended
(0) based on the Cartesian parameters. The CSL is derived from the two inputs OCL and
BL as in algorithm 3.
Algorithm 3: working of Route handling logic
3.2.4 Route table
The route table is a tuple; it stores the information required to assist the routing of the packets
towards the node having the better stability such as CSL value. The node initiating the routing to
transfer the message is called as Source node; The intermediate nodes that assist the source node
to set up the path are called Host nodes; The intended node to transfer the message is called as
Target node. The fields of the routing table are as shown in Table 2.
1: Begin
2: INPUT: Belief BL, Opportunistic contact OCL;
3: OUTPUT: Collaboration sensitivity level CSL;
4: initialize the parameters P and Q with 0
5: while P ≤ 2 do
6: call the contact analyzer and extract the value of the OCL to OCL P
7: end while
8: while Q ≤ 3 do
9: call the belief formulator and extract the value of BL to BLQ
10: end while
11: for all (OCL P , BLQ) do
12: obtain I = ∏(OCLP × BLQ), Where ‘×’ represents the Cartesian product of set OCL and BL
13: end for
14: If I == <(H,P)||(H,C)> then CSL ⇐ Highly recommended
15: If I== <(H,V)||(M,V)||(L,S)||(L,V) > then CSL ⇐ Not recommended else CSL⇐Recommended
16: Return CSL
16: End

12
Table 2: Route table of COAODV
UID CSL Source
UID
Source
CSL
Target
UID
RR RFs
01 2 00 2 03 I1-D 2
00 2 00 2 03 S-I1 1
Where, UID = The unique number of a host node.
CSL= Collaboration sensitivity level of the host node.
Source UID = The unique number of a source node.
Source CSL = The collaboration sensitivity level of source node
Route record (RR) = The forward path the packet may traverse.
Destination node = Target node for data packet.
Routing factor (RFs) = Indicates the connectivity reliability of the route on source.
3.3. Routing Procedure
When a node initiates the data routing, on the unavailability of the route, the route discovery
phase starts. During the route discovery phase, the source node clones the MCA along with the
RREQ packet to all its neighbor nodes as in figure 6.
Figure 6: Route request packet
The belief formulator of the cloned MCA will generate the belief about the resource availability
on the node and sends it to route handling logic.
The route handling logic extracts the OCL from the contact analyzer of the node over which it
resides. It will generate the CSL of the host node and copies to the RREP packet. The RREP
packet will be sent back to the source node.
Several RREP packets will reach the source node and updates its routing table. The MCA of the
source node will compute the Routing factor (RF) based on the established CSL value and stores
in the route table.
Where, CSLsource = The collaboration sensitivity level of the computing node.
CSLhost = The collaboration sensitivity level of the neighboring host considered.
The source node will select only the path which has routing factor greater than 1 and initiates the
data transfer as shown in figure 7.

13
Figure 7: Route path setup in COAODV
The belief formulation is a proactive process running on the MCA and each time the node
allocate the resources the belief over the source node changes and thereby the CSL of the
node varies. MCA's on each node will be monitoring its level and during the data packet
transmission, with decreasing CSL (long before the completion of N/2 packets) the MCA
Proactively stimulate on demand belief formulation process and updates the belief record.
Again the route handling logic will verify its CSL with the requested CSL level and on no
match computes the routing factor. If required (if it’s RF<1) handovers the routing
responsibility to a neighbouring node having RF>1.
As in figure 8, initially with sparse nodes being isolated the degree of opportunistic
contact will be less. With the increase in node density in both with and without agents
scenario the ∆ increases. However with further increase in the node density, the due to
resource starvation ∆ decreases in network without agents. As the resource availability is
predicted and the route is setup based on recommendations ∆ increases with increase in
node density.
The plot in figure 9 gives the rate of generation of different classes of collaboration
sensitivity levels. In can be observed that the class of ‘Recommended’ will be more
compared to ‘Highly recommended’ and ‘Not recommended’ signifying that a controlled
recommendations are given so a reliable path is set up for routing and reduces the
frequent triggering of route discovery phase.

14
Figure 8: Number of nodes Vs degree of Opportunistic contact (∆)
Figure 9: Number of nodes Vs CSL
4. VALIDATION OF COAODV
The COAODV is a stable routing protocol for MANET's where the stable route is obtained
between the source and destination based on the routing factor of the intermediate nodes which
depends on the resource availability for the data transfer. The protocol is validated through
comparison of the proposed protocol with a sleep-AODV protocol [26].
4.1 Sleep-AODV
It is an extended power aware routing protocol, which takes into account the behaviour of the
node (stable, unstable or isolated) to maintain the connectivity. The purpose of the sleep-AODV
protocol is to limit excessive routing messages caused by a large number of neighbours because a
routing table with many entrances involves a significant number of control packets, which
consumes more energy spent in the transmission, reception, and overhearing mechanism.
4.1.1 The principle of Sleep-AODV protocol
In Sleep-AODV protocol a node switch to standby mode for some time according to the network
status as shown in figure 10. The general operation of a node is represented by four states idle
(initial stat), Sleep, receive or transmit. Each node goes to sleep for a fixed duty cycle called sleep
time by turning off the radio. Then wakes up and listens in the idle state to see if any node
initiates communication either through send or receive state.

15
Fig 10: State diagram of Sleep-AODV protocol
Sleep-AODV protocol is designed to detect the unstable node, isolated nodes and in active paths.
On detecting such conditions it turns off the radio and reduces the energy resource consumption.
Hence, Sleep-AODV protocol avoids the reroute discovery due to energy deficient link failure by
reducing the passive listening period of the nodes.
4.2 Experimental Setup
An experimental setup is built to analyze the behaviour of the COAODV protocol in comparison
with Sleep-AODV. The scenario is setup for post-disaster management equipped with mobile
devices, in and around the disaster area. Immediately after the occurrence of the disaster, the
disaster relief operation will takeoff, and the rescue team will start the search and rescue activity.
The members of the rescue team will be working by grouping themselves into many subgroups.
Possible communication subnets:
• To have a personal connection between them, the team members will carry the mobile
devices such as smart phones, tablets, palm top, laptops and other devices having the
wireless capabilities. They can form the subnets in and around the incident location(
subnets created in and around the incident location due to the usage of computational
devices bundled with several wireless network interfaces).
• To clear the casualties and to transport the rescued victims for the medical supervision,
rescue vehicles will be moving throughout the incident area and can form the vehicular
network (The vehicular network created by the rescue vehicles moving around with the
devices come in contact with it on both the sides of its path).
Thus, the application scenario depicts that in spite of collapsed communication network there are
an intermittent group of subnets available with a various level of wireless capabilities and device
characteristics, which can be made use to setup a communication network among the personnel of
the rescue team.
Let us consider the scenario in which pooling of resources is done in a completely or partially
collapsed communication infrastructure formed amid the affected area, as shown in Figure 11. It
involves a mix of mobile devices with random mobility patterns and sporadic data transmission
rates having various computational capabilities. For example, the team of rescue members
(nodes) holding the wireless communication devices, such as smart phones, laptops, and so on,

16
will move with a group and have slower unpredictable movement. Thus, there will be much delay
in message passing. The rescue vehicles (nodes) will have faster and slightly predictable path,
leading to faster message transmissions and so on.
Figure 11: A Post disaster resource pooled MANET’s communication scenario
Initially, these mobile resources are registered into any one of the following four groups
based on their resource capacity and skills as shown in table 3.
Table 3: Types of mobile nodes
Node Description
Rescue Team A group of trained people allocated with pre-registered devices for faster
searching and triage.
Bystander The local volunteer individuals registered to assist the rescue team
Navigation
Controller
Registers the new devices and guides the search and triage operation
Vehicle
Controller
Tracks and coordinates the rescue vehicles used for victims triage
The model implemented with the COAODV has events such as registration,
quantification, mapping, and belief generation.
The MCA of the navigation controller, being the registration authority, will clone itself on the
newly registered device with the necessary initial handshaking. The copy of the agent from the
registration authority will migrate onto the new device, ready to capture the behaviours and
formulate the beliefs about resource availability. The MCA on the new device (node) will be
available for communication with the MCAs of the other nodes.

17
4.3 Test bed description
The evaluation of COAODV is done on a Riverbed opnet [28] test bed for a post-disaster scenario
as shown in figure 12. The simulation environment is considered in an area of 500 x 500 unit,
where mobile nodes are distributed randomly. The environment comprises 200 wireless mobile
nodes embedded with the MCAs designed using agent factory, based on 802.11b configured in
ad-hoc AODV routing mode.
Fig 12. The simulated Test bed
The nodes’ mobility is configured to a random waypoint with a communication range of 60 m
and pause time ranging from 0 to 150 ms. The nodes are running with the Linux kernel on the IP
networking protocol stack that is essential in studying the scenario, of which about two-thirds are
mobile nodes such as a WLAN iPad, Android mobile, WLan iPhone, MANET workstations,
WLAN workstations and so on supporting 1Mbps WLAN with CBR traffic. The nodes belonging
to the rescue team and bystanders have a reduced communication range of 10–30 m. The
remaining nodes are static for all transmission range such as survived infrastructure, vehicle
controllers, and navigation controllers emulated through Mac workstations with wireless cards.
The registered devices will attach to the node’s resource pool and be mapped with its class of
resource and will then update the RR. The test bed gives various possibilities of a node coming in
opportunistic contact (nodes in the same direction, toward same locations, executing the same
task) with other registered nodes for resource pooling. It uses the RFC 2501 [27] mobile IP
networking wireless protocol for node interaction.
The performance metrics used for the comparison and analyses of the protocol are as follows:
Number of control packets The amount of control packets generated/connection for route
discovery and rediscovery.
Packet delivery rate (PDR) The ratio of no. of packets sent to the no. of packets
received at destination.
Packet loss The number of CBR connection between the nodes to
the packet loss.

18
Overhead traffic rate(OTR) The ratio of the control traffic to the total traffic
generated over the number of connections.
End-to-End delay The average time between the start of packet transmission to the
time it reaches the destination with traffic density. The
breakdown time includes the delay due to route discovery
process and the queue in data packet transmission.
Initially, to establish the path between the nodes the route discovery phase is initiated. The RREP
packets are flooded over the neighbour nodes whose density increases with increase in connection
over the network as shown in figure 14. In sleep-AODV, with an increase in the connection the
control packet flooding is reduced by 20% as the nodes will go to sleep in a specific duty cycle.
In COAODV, control packets cloned with the MCA are flooded only during the initial route
discovery phase. The MCA’s will compute the belief on connectivity and reduce their discovery
phase. Due to this with an increase in the number of connections flooding of control packets is
been greatly reduced.
Figure 14, shows the rate of overhead traffic with the varying number of connections. At the
lower OTH, all the protocol behaves similarly. With an increase in the number of connections,
AODV suffers from highest OTR due to more rediscovery phases because of a number of
collisions leading to link breakages and more RREP packet broadcast. Sleep-AODV will detect
and separates unstable, isolated nodes and inactive path. So it will exhibit a reduced OTR
compared to AODV protocol. COAODV exhibits the minimum OTR because of least number of
link failures. COAODV will predetermine and the will setup the path only with the links having
required resources. It maintains the link with the obtained opportunistic contact and thereby
provides higher link stability.
Figure 15 gives the plot of Packet delivery rate which is highest in COAODV protocol. In
COAODV protocol, the packets will be sent only through the path which is having greater
stability and thereby reducing the number of packet retransmission. In AODV repeated link
breakages with the increase in a number of nodes have lesser PDR. Initially, Sleep-Aodv has least
PDR due to the number of inaccessible nodes and increases with increase in node density.
Fig 13. Number of CBR connections Vs Number of control packets generated

19
Figure 16 shows the plot of a number of connections Vs End to end delay. The Sleep AODV
protocol has a large delay, as most the time the nodes will be in sleep mode and are inaccessible.
The COAODV has least End to end delay due to the fact that the path will be established only
among the nodes having the routing factor greater than the threshold level and hence provides
greater stability and reliability.
Fig. 14 Number of connections Vs OTR
Fig 15. Number of nodes VS PDR

20
Fig 16. Number of connections Vs E-E delay
CONCLUSION
The Opportunistic computation in MANET’s exploits the contact opportunities to establish the
communication among the nodes without end- to-end connectivity. The OC with cognitive agents
can provide a better decision on route set up reducing the unnecessary route discovery phases.
The paper presents a novel COAODV scheme to improve the stability and reliability of the
network connectivity in MANET’s. The scheme uses CA-based OC principles to determine the
resource based routing stability of the path. The technique is implemented based on the BOB
model for a disaster management scenario. The scheme used works close to human cognitive
behaviour in determining the stable route reducing the frequent route discovery phase and is
therefore an efficient approach for MANET’s. The scheme on comparison with the AODV and
Sleep-AODV provides a reliable and adaptive and stable routing in a disconnected environment.
REFERENCES
[1] S. Corson , J. Macker, Mobile Ad hoc Networking (MANET): Routing Protocol Performance Issues
and Evaluation Considerations, RFC 2501 (1999)
[2] Jun Zhao Sun, Mobile Ad hoc Networking: An Essential Technology for Pervasive
Computing,proceedings of the International Conferences on Info tech and Info net, IEEE digital
Explore , DOI: 10.1109/ICII.2001.983076, 2001.
[3] Jeroen Hoebeke , Ingrid Moerman , Bart Dhoedt Piet Demeester, An overview of mobile ad hoc
networks: applications and challenges, Journal of the communications network. 3(3). P.60-66, 2004.
[4] Ripan Kumar, Performance Comparison of AODV and DSR Routing Protocols in manets, Thapar
University Digital Repository (tudr), http://hdl.handle.net/123456789/163, 2007.
[5] Elizabeth M. Royer, C.-K. Toh, A Review of Current Routing Protocols for Ad hoc Mobile Wireless
Networks, IEEE Personal Communications Magazine, April 1999, pp.46-55.
[6] Elizabeth M. Royer, Charles E. Perkins, An Implementation Study of the AODV Routing Protocol,
Proceedings of the IEEE Wireless Communications and Networking Conference, Chicago, IL,
September 2000.
[7] Per Johansson, Tony Larsson, NicklasHedman, , BartoszMielczarek., Routing protocols for mobile
ad-hoc networks - a comparative performance analysis. In Proceedings of the 5thInternational
Conference on Mobile Computing and Networking(ACM MOBICOM’99), August 1999

21
[8] Samir R. Das, Charles E. Perkins, Elizabeth M. Royer, Performance Comparison of Two On-demand
Routing Protocols for Ad hoc Networks. Proceedings of the IEEE Conference on Computer
Communications (INFOCOM), Tel Aviv, Israel, March, pp. 3–12, 2000.
[9] Emmanouil A. Panaousis, Behrouz Maham, Walid Saad, Efficient cooperative protocols for general
outage-limited multihop wireless networks, Proceedings in IEEE International Symposium on
Personal, Indoor and Mobile Radio Communications, 2010.
[10] M.Conti, Opportunities in opportunistic computing, IEEE Communication Magizine. 43, 42–50,
2010.
[11] C.M.Huang, C.-Z. Tsai and K.-c. Lan, A survey of opportunistic networks, IEEE Digital. Explorer.
DOI 10.1109/ WAINA.2008.292, 2008.
[12] C.Bradley, Cognitive agents interacting in real and virtual world. http://act- r.
psy.cmu.edu/wordpress/wp-content/ uploads/2012/12/622SDOC4694.pdf.
[13] J.Bradshaw, Software agents, AAAI Press, California, 2000.
[14] M.N.Huhns and M. P. Singh, Cognitive agents, IEEE Internet Comput. 2, 87–89, 1998.
[15] R.D.Beer, Framing the debate between computational and dynamical approaches to cognitive science,
J. Behav. Brain Science 21 , 630, 1998.
[16] C.Castelfranchi, Guaranties for autonomy in cognitive agent architecture, intelligent agents: theories,
architectures and languages, LNAI 890 , 56–70, 2005.
[17] D.B.Lange and M. Oshima, Programming and deploying Java mobile agents with Aglets, Addison-
Wesley, USA, 1998.
[18] K.Ahmed and S. Majid, Post-disaster communications: a cognitive agent approach, in: 7th
International Conference on Networking, 2008.
[19] B.S.Bsbu and P. Venkataram, Cognitive agents based authentication & privacy scheme for mobile
transactions (CABAPS), Comput. Commun. 31, 4060–4071, 2008.
[20] Cancer Research UK, Cancer statistics reports for the UK. , 2003 (accessed 13.03.03).
[21] M Gerla, SJ Lee, AODV-BR: Backup routing in ad hoc networks, Proceedings in IEEE Wireless
Communications and Networking conference, volume 3, 1311-1316(pp), 2000.
[22] Naif Alsharabi, Nasr AlSharabi, Li Ren Fa ,Mohammad H. Algamali, Applied Packet's Received
Time Prediction to Reactive Ad-hoc Routing Protocol, Journal of Applied Sciences 8 (6): 1035-1041,
2008.
[23] C. T. Cuong, V. T. Tu and N. T. Hai, “MAR-AODV: Innovative Routing Algorithm in MANET
based on Mobile Agent”, 27th International Conference on Advanced Information Networking and
Applications Workshops, pp.62-66, 25-28 March 2013.
[24] En Bo Wang, Jing Xia, Dan Keun Sung, Xin Ming Zhang, A Neighbor Coverage based Probabilistic
Rebroadcast for Reducing Routing Overhead in Mobile Ad hoc Networks. DOI:
10.1109/TMC.2011.277, IEEE Transactions on Mobile computing 12(3):424-433 · March 2013.
[25] M. Askalani, S. L. Toral, F. Barrero, E. Asimakopoulou, N. Bessis , D. G. Reina,A Survey on
Multihop Ad Hoc Networks for Disaster Response Scenarios,
http://dx.doi.org/10.1155/2015/647037, International Journal of Distributed Sensor networks, Article
ID 647037, 16 pages, 2015.
[26] Mounir Tahar Abbes, Belhirech Mohamed, Senouci Mohamed, Adaptation of a routing algorithm in
wireless video sensor network for disaster scenarios using JPEG 2000, DOI: 10.1007/s11276-015-
0979-z, Wireless Networks 22(2) · May 2015.
[27] Ntuen, Celestine A, Gwang M Kim. “An Agent-based Model Simulation of Multiple Collaborating
Mobile Ad Hoc Networks (MANET).” In 16th International Command & Control Research and
Technology Symposium Collective C2 in Multinational Civil-Military Operation, 51, 2011.
[28] https://support.riverbed.com/download.html
[29] S. Corson, J. Macker, “Mobile Ad hoc Networking (MANET):Routing Protocol Performance Issues
and Evaluation Considerations”, RFC 2501, IEFT, 1999.

An Opportunistic AODV Routing Scheme : A Cognitive Mobile Agents Approach

More Related Content

What's hot (14)

Similar to An Opportunistic AODV Routing Scheme : A Cognitive Mobile Agents Approach (20)

Recently uploaded (20)

An Opportunistic AODV Routing Scheme : A Cognitive Mobile Agents Approach