2016 Wireless Days (WD)

Energy-Efficient Adaptive Forwarding Scheme for

MANETs
Atif A. Alghamdi, Robert J. Pooley and Peter J. B. King
School of Mathematical and Computer Sciences
Heriot-Watt University, Edinburgh, UK
{aaa148, R.J.Pooley, P.J.B.King}@hw.ac.uk
Several schemes have been proposed to alleviate this problem [6], [7], [8]. However, most of these schemes assume that
every node in the network has the same transmission range.
Therefore, these schemes have poor performance when the
nodes in the network have different transmission ranges [5],
[9]. Nevertheless, asymmetric/unequal radio range in wireless
communication can occur in MANETs for many reasons such
as power saving protocols [10]. It is believed that with variable
transmission ranges in Ad-hoc networks can perform more
complex tasks.

AbstractFlooding is the simplest way of broadcasting, in

which each node in the network retransmits an incoming message
once. Simple flooding technique in wireless Ad-hoc networks
causes the broadcast storm problem. However, this technique
is inefficient in terms of resource consumption such as bandwidth and energy. This paper presents a new hybrid scheme
that combines different techniques that collaborate to reduce
overhead and conserve energy. We propose an Energy-Efficient
Adaptive Forwarding Scheme, that utilizes the information of
the 1-hop neighbouring radios. In this scheme nodes do not
need a positioning system or distance calculation to determine
their location. In addition to the previous works, the proposed
protocol divides the network into different groups based on their
transmission-power levels. Therefore, the node which receives
HELLO message from different groups is considered a Gateway
node. This node efficiently participates in forwarding RREQ
packets and the unnecessary redundant retransmission is avoided.
The performance evaluation of the proposed protocol shows a
reduction in the routing overhead and in energy consumption,
when compared with the Pure-Flooding AODV and DynamicPower AODV using NS2.

This paper presents an energy-efficient adaptive forwarding

scheme that takes into account, the variable TransmissionPower levels of the neighbouring nodes. The proposed scheme
dynamically adjusts the RREQ probability of forwarding
at each node, based on the node position (Gateway/Nongateway). On the other hand, this base is ignored in sparse
region as compared to dense region to maintaining the network connectivity. The decision to forward RREQ packets
is made instantly after receiving a packet without any delay.
Asymmetric radio ranges have been considered in this work.
The new scheme not only reduces unnecessary redundant
retransmission but also increases reachability of data packet
and at the same reducing dropped packet ratio and packet
loss. Thus, reduces overall routing overhead and improves the
energy consumption. For the evaluation of our scheme, the
conventional AODV which uses pure flooding and DP-AODV
which uses power control are use to compare our scheme. The
rest of this paper is organised as follows: Section II provides
an overview of previous related work, Section III describes the
proposed work, simulation scenarios and performance metrics
are presented in Section IV, performance result provided in
Section V, and finally, the conclusion and future work is
provided in Section VI

Keywords- MANET, routing, broadcasting, forwarding, energy conservation.

I.

I NTRODUCTION

Mobile ad hoc network (MANET) is basically a group

of mobile nodes that communicate using multi-hops in a
non-infrastructure based setup. MANETs have been receiving
serious attention over the recent years due to their vast potential
in both military and civil applications [1]. The intermediate
nodes between source and destination play a significant role
in forwarding the control packets. This flooding mechanism is
fundamental to the route discovery process. However, flooding
can be redundant and energy consuming. Therefore, many attempts were made to restrict retransmission of control packets
to minimize unnecessary overhearing.

II.

Flooding is the most frequently used method for nodes

to exchange network information or deliver Routing Request
(RREQ) messages to a destination [2]. Moreover, the simplest
way to flood a network is through a technique called Simple
F looding or P ure F looding [3]. In simple flooding when a
node receives a packet, it blindly rebroadcast it immediately to
all its one-hop neighbours, it may cause a lot of unnecessarily
rebroadcast of packets. Thus, a lot of packet collisions occur
especially if the network is dense. Therefore, the main purpose
of different flooding techniques in MANETs is to avoid the
well-known Broadcast Storm P roblem [4] as much as
possible, while maintaining high network coverage and low
delivery latency [5].

978-1-5090-2494-0

R ELATED W ORK

As the type broadcast scheme used plays an important

role in the performance of the network, it has to be selected
carefully. In MANETs there are several broadcasting protocols
designed to achieve an efficient flooding to avoid unnecessary
redundant retransmission.
Williams and Camp [11] classified the existing efficient
flooding schemes into simple flooding, probability based
schemes, counter-based schemes, distance based, locationbased schemes and neighbour knowledge schemes.
A notable work of efficient flooding that uses 1-hop
neighbor information is Edge F orwarding. Ying et al [12]

2016 Wireless Days (WD)

Fig. 1.

The proposed Gateway Forwarding

introduced this work attempt to reduce flooding traffic by

exchanging location knowledge between neighbouring nodes.
So that retransmission is limited to nodes near the boundary of
each other. In this scheme, the transmission area of the node
N is divided into six equal-size regions and labelled as NP 1,
NP 2, NP 3, NP 4, NP 5 and NP 6 respectively. The authors
considered that their edge forwarding technique was novel at
the time. Nine years later Raiyani and Lathigara [8], refer back
to them when explaining this technique in their survey of the
latest work on flooding mechanism for AODV. This means that
the idea of edge forwarding can be considered relatively new
and is open for more investigation, and concluded that Edge
F orwarding works well with a large network domain area and
is more suitable for flooding without extra control overheads.
However, using nodes transmission range in the expressions to
calculate the forwarding probability is undependable in reality
[6]. This parameter can be very different from the nominal
value defined by the manufacturer. Moreover, external factors
such as external noise and interferences can also effect the real
value.

reduction in the number of transmission collisions will increase

energy saving. Transmission-Power control is very common
issue of researches to reduce collision and congestion in high
dense network [17], thus reduce energy consumption.
Bamhdi and King [18] developed a DP -AODV protocol.
This protocol automatically adjusts the transmission power at
each node so as to keep its number of neighbours within
a specified range. This work attempts to prevent unwanted
interference and unnecessary overhearing by other nodes.
Thus increasing the throughput in the network and reducing
energy consumption. However, there are no restrictions on
rebroadcasting in the DP -AODV , which means unnecessary
broadcast occurs, which in turn increases overhead.
As a result, the researcher chooses the applied forwarding
scheme, motivated by the following reasons: First, it is easily
implemented, because it only needs a few lines of code to
design it. Second, no additional hardware or calculation is
needed to operate it. Finally, the node need not wait for a
random delay period; node can immediately decide to forward
or rebroadcast the RREQ or not.

Qayyum et al [13] proposed a sender-based mechanism

called M ultipoint Relay (MPR) for efficient broadcasting.
This mechanism restricts the number of retransmissions by
checking the 2-hop neighbors that can only be reached by
1-hop neighbors and select them as MPRs. Then, from the
remaining 1-hop neighbors, select those that cover the most
2-hop neighbors that are not covered yet. Continue until all
2-hop neighbors are covered. However, in order to calculate
the multipoint relays, every node in MPR has to collect the
set of 1-hop neighbors and the 2-hop neighbors, which results
in heavy overhead.

III.

The forward node choice is a vital part in designing

the proposed algorithm. Thus, in this work, we classify the
mobile nodes in the network as Gateways and Non-Gateways
with respect to the neighbouring Transmission-Power level
(P tLevel ). Note that, Gateways are those nodes which have
different values of P tLevel in their neighbouring table are
more likely to forward RREQ packets. On the other hand,
Non-Gateways are those nodes which have same values of
P tLevel in their neighbouring table (see Fig. 1). Therefore,
any retransmission by Non-Gateway should be exempted in
order to avoid early unnecessary redundant retransmission and
decrease overhead of routing packets. So the decision which
node is Gateway and which is Non-Gateway is taken on the
basis of node position. The nodes which are located between
different groups ( can receives packets from different values
of P tLevel ), will be categorized as Gateways; else NonGateways.

Cluster-based flooding scheme have suggested by many

researchers as a way to improve broadcasting reliability. In the
previous works [14], [15], algorithm divides an entire network
into groups, each of which being represented by a Cluster
Head (CH). The CH dominates all other members in the group
and retransmits broadcast packets. However, they require the
large amount of information exchange and reconstruction of
the clusters whenever the network topology changes.
It has been found, in the entire wireless network and
MANET in particular that, transmission collision is a major
source of energy loss. Huang et al [16] argued that a successful

978-1-5090-2494-0

P ROPOSED A LGORITHM

The method for determining the different values of P tLevel

are employed [18] by dynamically adjusted the Transmission

2016 Wireless Days (WD)

Power at each node, so as to keep its number of neighbours

within a specified power level. In the physical layer, the
node adjusts its Transmission-Power value (P tV alue ). Three
different power levels are used; High power if low number
of neighbours (fewer than 7 neighbours); Medium (7 to 15
neighbours); and Low power if high number of neighbours
(more than 15 neighbours). Table I shows the values of levels
that were used in our simulation.
TABLE I.

T RANSMISSION -P OWER L EVELS

P tLevel

P tV alue

P tRange

High
Medium
Low

0.28183815
0.24169726
0.20191908

250 m
200 m
170 m

The basic idea is that only those nodes have to forward

the packet, which bridge groups of nodes with different
P tLevel . Intuitively, the node that has the same P tLevel as its
neighbours is most likely located within the same group. Thus,
all neighbours may receive a broadcast packet. On the other
hand, this base of gateway determination is ignored for the
node that is located in a high-sparse or high-dense network
region (all its neighbours have same P tLevel and there is
no node determined as gateway), to maintaining the network
connectivity. Therefore, the P tLevel is utilized in the proposed
method in order to attain more efficient and reliable flooding.

Fig. 2.

When the nodes knowledge is used to determine if it should

participate in RREQ forwarding, two goals are achieved. First
of all, energy that is the main resource of nodes will be
reserved, because only specific nodes can participate in retransmitting RREQ packets. Secondly, the impact of increasing
control overheads and congestion in wireless network, which
could lead to degrading of the networks performance, are
avoided. The proposed forwarding scheme will be presented
in detail in this section.

and there is no available route, the source node will broadcast

a RREQ packet to find a path to destination.
Intermediate Node: If any intermediate node within the
source transmission range receives the RREQ packet for the
first time, a node implementing the proposed mechanism that
shown in Fig. 2, and it behaves as follows:

A. Control Packets
To determine whether the node in the network is a
Gateway or not, we need to exchange packets and for this
we modify the existing control packets in AODV protocol
[19] to broadcast the P tLevel (High/M edium/Low) and
Status(Gateway/N onGateway) instead of adding new control packets.
HELLO packet: All nodes in the proposed protocol make use
of the HELLO packet in AODV to exchange the information
of their transmission-power level and status among each other.
In order to accumulate such information, a new fields called
Transmission-Power Level (P tLevel ) and Status are added
to the HELLO packet. With the help of the HELLO packet,
every node creates a neighbour table which store the P tLevel
and Status values of all 1-hop neighbours.

If receives the RREQ for the first time.

It checks if the RREQ sender is currently on the

neighbouring table.

If the RREQ sender currently exists, the node then

determines whether it is a Gateway node or not by
checking its neighbouring table.

If its neighbouring table has nodes with different

values of P tLevel , it then determines that it is a
Gateway node and then it forwards the RREQ packet.

Else if all its neighbours have same value of P tLevel

and there is no neighbour holds Gateway value of
Status, it then operates as a gateway and forwards
the RREQ packet.

Destination Node: Whenever the RREQ packet has reached

the intended destination or an intermediate node that has a
fresh route to the destination, it unicasts a reply by sending a
Route Reply (RREP) packet along the reverse path established
at intermediate nodes during the route discovery process.
However, we observed too many data packets dropped by the
routing layer when implementing the proposed algorithm on
the simulator.

B. Route Discovery
The route discovery process has operations on all nodes
which include source node, intermediate nodes and destination node. The route discovery process of proposed scheme
operates as follows:
Source Node: The source process is same as used in AODV
that whenever a source needs a route to the destination. When a
source node is going to transfer packets with a destination node

978-1-5090-2494-0

Description of the proposed forwarding RREQ

Since the network operates on variable Transmission-Power

ranges (P tRange ), Asymmetric Link P roblem [20] occurred
regularly. As illustrated in Fig. 3, node b can hear Hello
message from node a, but node a cannot hear from node b.

2016 Wireless Days (WD)

Fig. 3.

It then handles the (P tV alue ) to the Physical Layer in

proportion for forwarding RREP packet (see Fig. 4-b).

Asymetric link problem

Asymmetric links can cause complete node failure; although

node b can receive RREQ packets from node a, its own
RREP packets may never reach node a. Therefore, many data
packets would be lost, thus degrading the network throughput.
Moreover, to addresses this problem, this paper proposes
a technique that attempts to deliver the RREP packets to
its destination by introducing the following modification in
sending and forwarding RREP process:

Fig. 5.

Asymetric link solution

Fig. 6, shows the effect of the proposed modified RREP

process on the number of dropped data packets. In the case
of asymmetric link the proposed RREP process decreased
the number of dropping data packets, which means that the
P tLevel used to send or forward the RREP packets, was well
selected.

Fig. 6. Effect of the proposed sending and forwarding RREP process on the
number of data packets drop

IV.

Fig. 4.

The simulation and performance evaluation are applied

using NS2.35 framework. The specific simulation parameters
are shown in Table II. Node mobility used a random waypoint
model with different speeds (10, 20, 30 m/s), and the mobility
had no significant effect of performance. The connection model
used in the simulation was UDP (User Datagram protocol) with
Constant Bit Rate (CBR). The simulation of different number
of nodes (75, 100, 150 and 200) that form a network over
a fixed area of 1000m x 1000m is used for evaluation. The
simulation time is fixed at 100 seconds. Moreover, the MAC
layer used was the IEEE 802.11. These parameters were widely
used in previous research papers, which in turn enabled the
results of this simulation to be compared with results from
those studies.

Description of the proposed sending and forwarding RREP

Send RREP:

If receives RREQ packet.

It checks whether it is a destination or it has a fresh

route to the destination.

It selects the temporary appropriate P tLevel according

to the next hop P tLevel in the neighbouring table (see
Fig 5).

It then handles the (P tV alue ) to the Physical Layer in

proportion for send RREP packet (see Fig. 4-a).

The following metrics are normally used to evaluate routing

protocols in ad hoc networks. Moreover, they can also be used
to evaluate broadcast schemes as well, especially that they are
an important part in the performance of the routing protocols.
Routing Overhead: is the total number of control packets
(include route requests, replies, error messages and Hello message) transmitted for route discovery and route maintenance
needed to deliver the data packet.
End-to-End Delay (EED): is the average time that a packet

Forward RREP:

If receives RREP packet.

It checks if it has a fresh route to the destination.

It selects the temporary appropriate P tLevel according

to the next hop P tLevel in the neighbouring table.

978-1-5090-2494-0

S IMULATION PARAMETERS AND P ERFORMANCE

M ETRICS

2016 Wireless Days (WD)

TABLE II.

SIMULATION PARAMETERS

Mac Layer
Bandwidth
Propagation model
Antenna model
Simulation area
Simulation time
Number of nodes
Node speed
Pause time
Traffic type
Number of connections

DP-AODV increased dramatically when the number of nodes

increases in comparison to the Proposed Algorithm which
managed to maintain a noticeably low overhead, approximately
80% lower than the others. This is because, when network
density increases (resulting from the increase of nodes) there
is more chance for the network being divided to a large
number of groups working with different P tLevel . The small
area of overlapping generates a small set of gateway nodes,
which greatly reduces the number of forwarding RREQ. Thus,
unnecessary retransmission is avoided and routing overhead is
reduced. We can conclude that our proposed flooding algorithm
is more suitable for networks with high density.

IEEE 802.11
2 Mb/s
Two-ray ground
Omni-Antenna
1000 m x 1000 m
100 (sec)
75, 100, 150, 200
10, 20, 30 (m/s)
40 (sec)
CBR
40

experienced to send from the source until it reaches the

destination (measured in second).
Packet Delivery Ratio (PDR): is the ratio of packets that
were received successfully by the destination, to the packets
originated by the source.
Throughput: is the amount of data that is successfully received by their destinations divided by simulation time (measured in bits per second).
Energy Consumption: is the average energy consumed by a
node in the network (measured in joule), as energy is a limited
resource in ad hoc networks.
V.

40

Packet Delivery Ratio (%)

35

P ERFORMANCE R ESULTS

25

20

15

10

PF-AODV
DP-AODV
Proposed Algorithm

To analyse the performance of our scheme, we compare our Proposed Algorithm with Pure-Flooding (PF-AODV)
which use blind flooding to disseminate RREQ packets, and
Dynamic-Power (DP-AODV) which use Transmission-Power
control mechanism to reduce the signal collision. We study
the effect of node density on the performance of each protocol.
Our simulation results are obtained from 5 different movement
scenarios and 5 different traffic scenarios which means that
each metric value is the mean of the 25 runs. Generally, 95%
of the confidence interval was small compared with the value
that was being reported.

5
50

100

150

200

Number of Nodes

Fig. 8.

Packet Delivery Ratio

Fig. 8 shows the packet delivery ratio (PDR). The PDR

is the most important performance metric for flooding mechanism, considering that the goal of broadcasting is to deliver
packets to all nodes. Our Proposed Algorithm has a significant
impact in improving PDR it comes as a result of that our
technique mitigates more network collision and contention.
On the other hand, while increasing the nodes number in the
network, the PDR of PF-AODV and DP-AODV are noticeably
decreased, this is because pure broadcast induces unnecessary
redundant retransmission of RREQ packet, collisions in the
network, and causes the broadcast storm problem.

300000

PF-AODV
DP-AODV
Proposed Algorithm
250000

Routing Overhead (packets)

30

200000

150000

220

200
100000
180

Throughput (bps)

50000

0
50

100

150

200

Number of Nodes

160

140

120

100

Fig. 7.

Routing Overhead

80

PF-AODV
DP-AODV
Proposed Algorithm

60

Fig. 7 illustrates the routing overhead incurred by the three

protocols: PF-AODV, DP-AODV, and our Proposed Algorithm.
It shows that the Proposed Algorithm is significantly better
than performance and has achieved less overhead than the
PF-AODV and DP-AODV. Both the curves of PF-AODV and

978-1-5090-2494-0

40
50

100

150

Number of Nodes

Fig. 9.

Throughput

200

2016 Wireless Days (WD)

Fig. 9 shows the performance of the protocols based

on data throughput. The Proposed Algorithm shows superior
performance compared to the other two protocols. This is
because in the Proposed Algorithm there is restrictions on
RREQ rebroadcasting. This reduces the amount of unnecessary
signalling for route discovery. Thus, reducing the overhead that
can cause collision.

and has high probability in RREQ forwarding to maintaining

connectivity. Thus more energy consumption spending.
Finally, we measured the effect of maximum node speed
on the performance of each protocol; the Proposed Algorithm
was still not affected. This is because in our technique the
node depends on its knowledge of having at least two different
values of Transmission-Power level in its neighbouring table
to determine whether it is a Gateway/Forward node or not.
Therefore, our Proposed Algorithm is not affected by node
mobility.

3.6

3.4

End-End-Delay (second)

3.2

VI.

2.6

PF-AODV
DP-AODV
Proposed Algorithm

2.4

2.2

1.8

1.6

1.4
50

100

150

200

Number of Nodes

Fig. 10.

End-to-End Delay

Fig. 10 as we can see from the results that while the

number of nodes increases, our Proposed Algorithm achieves
low average of End to End delay compared to PF-AODV and
DP-AODV. This is because our Proposed Algorithm have lower
control overhead than the others.
2.6

PF-AODV
DP-AODV
Proposed Algorithm

Energy Consumption (Joule)

2.5

In our future work, more optimization and further improvement for reducing congestion and energy consumption will be
planned. For instance, once gateway node set is determined, the
problem of how to schedule those gateway nodes to efficiently
forward RREQ packets requires further study.

2.4

2.3

2.2

2.1

R EFERENCES
2

[1] R. Huo, D. Li, Q. Zhu, and H. Yang, An Efficient Distributed

Flooding Scheme Using One-Hop Information in Heterogeneous Ad
Hoc Networks, in Communications and Mobile Computing, 2009.
CMC 09. WRI International Conference on, vol. 1, Jan. 2009, pp.
513517.
[2] H. Jeong, H. Jeong, and Y. Yoo, Dynamic probabilistic flooding
algorithm based-on neighbor information in wireless sensor networks,
in International Conference on Information Networking, 2012, pp. 340
345.
[3] A. Qayyum, L. Viennot, and A. Laouiti, Multipoint relaying for
flooding broadcast messages in mobile wireless networks, in System
Sciences, 2002. HICSS. Proceedings of the 35th Annual Hawaii International Conference on, Jan 2002, pp. 38663875.
[4] Y. Tseng, Y. Chen, and J. Sheu, The broadcast storm problem in a
mobile ad hoc network (manet), in Proceeding of the 5th ACM/IEEE
International Conference on Mobile Computing and Networking, NY,
USA, 1999, pp. 51162.
[5] A. Al-Qawasmeh and S. Bettayeb, A proactive distance-based flooding
technique for MANETs with heterogeneous radio ranges, in Local
Computer Networks, 2008. LCN 2008. 33rd IEEE Conference on, Oct.
2008, pp. 648654.

1.9

1.8
50

100

150

200

Number of Nodes

Fig. 11.

C ONCLUSION AND F UTURE W ORK

Although energy consumption can be reduced at device

level, at transmission level or by using power aware routing
protocol [17], this paper presented a novel hybrid scheme
that combines different techniques that collaborate to reduce
overhead and conserve energy. The techniques used are: 1)Using Transmission-Power control mechanism in order to reduce
interference between nodes and increase the throughput in
the network, thus reducing the overall energy consumption.
2)Developing Transmission-Power aware routing algorithm in
order to restrict rebroadcast RREQ packets to avoid unnecessary overhearing, and selecting the appropriate power level
to deliver the RREP packets to its destination; especially
in asymmetric link situations. 3)Avoiding processes that majorly depend in their function on Global Positioning System
(GPS) [21] and distance calculators which wastes energy and
computation resources at device level. As a result, working
with different transmission ranges caused asymmetric link
problem to occur. Consequently, this problem was addressed
by modifying the proposed RREP process to attempt deliver
the RREP packets to its destination optimally. Thus, most data
packet drop cases in routing layer is avoided which eventually
increase the packet delivery ratio and throughput.

2.8

Energy Consumption

Fig. 11 shows that, while the number of nodes increases,

the Proposed Algorithm consumes less energy compared to
the other two protocols. The fact behind this result, that
the Proposed Algorithm combines different techniques that
collaborate to conserve energy. Thus, the Proposed Algorithm
can perform its functions with high level of energy efficiency.
As shown, in sparse network, the proposed algorithm consume
higher power as compared to dense network. This is because,
while a node in sparse network it has a small number of
neighbours, thus a node operates with high level of power

978-1-5090-2494-0

2016 Wireless Days (WD)

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]
[14]

[15]

[16]

[17]

[18]

[19]
[20]

[21]

D. G. Reina, S. L. Toral, P. Johnson, and F. Barrero, A survey on

probabilistic broadcast schemes for wireless ad hoc networks, Ad Hoc
Networks, vol. 25, pp. 263292, 2015.
H. Liu, X. Jia, P.-J. Wan, X. Liu, and F. F. Yao, A distributed and
efficient flooding scheme using 1-hop information in mobile ad hoc
networks, Parallel and Distributed Systems, IEEE Transactions on,
vol. 18, no. 5, pp. 658671, 2007.
A. G. Raiyani and A. M. Lathigara, Probabilistic and neighbour knowledge based flooding mechanism for AODV, in Advanced Computing
& Communication Technologies (ACCT), 2014 Fourth International
Conference on. IEEE, 2014, pp. 216221.
D. Kim and H. Yoo, Stateless broadcasting to support heterogeneous
radio ranges in mobile ad hoc networks, in Computer Communications
and Networks, 2006. ICCCN 2006. Proceedings. 15th International
Conference on. IEEE, 2006, pp. 5662.
K. A. Rahman and K. E. Tepe, Analyzing extended sliding frame
reservation-aloha MAC protocol against asymmetric radio link problem
in MANET/VANET, in Wireless Communications and Mobile Computing Conference (IWCMC), 2013 9th International, 2013, pp. 1635
1639.
B. Williams and T. Camp, Comparison of broadcasting techniques for
mobile ad hoc networks, in Proceedings of the 3rd ACM international
symposium on Mobile ad hoc networking & computing. ACM, 2002,
pp. 194205.
Y. Cai, K. A. Hua, and A. Phillips, Leveraging 1-hop neighborhood
knowledge for efficient flooding in wireless ad hoc networks, in Performance, Computing, and Communications Conference, 2005. IPCCC
2005. 24th IEEE International, Apr. 2005, pp. 347354.
A. Qayyum, L. Viennot, and A. Laouiti, Multipoint relaying: An
efficient technique for flooding in mobile wireless networks, 2000.
J. Wu and H. Li, On calculating connected dominating set for efficient
routing in ad hoc wireless networks, in Proceedings of the 3rd
international workshop on Discrete algorithms and methods for mobile
computing and communications. ACM, 1999, pp. 714.
C. R. Lin and M. Gerla, Adaptive clustering for mobile wireless networks, Selected Areas in Communications, IEEE Journal on, vol. 15,
no. 7, pp. 12651275, 1997.
Y.-C. Huang, S.-Y. Chuang, and S.-D. Wang, A dynamic node degree
management scheme for energy-efficient routing protocols in wireless
ad hoc networks, in Parallel and Distributed Systems, 2005. Proceedings. 11th International Conference on, vol. 2. IEEE, 2005, pp. 711.
N. Tantubay, D. R. Gautam, M. K. Dhariwal, and Others, A review
of power conservation in wireless mobile adhoc network (MANET),
IJCSI, vol. 8, no. 4, p. 1, 2011.
A. M. Bamhdi and P. J. B. King, AODV with Dynamic Power
Enhancement: Performance Evaluation in MANETs, Journal of Communications Engineering and Networks, vol. 2, no. 1, pp. 2333, 2014.
C. Perkins, E. Belding-Royer, and S. Das, Ad hoc on-demand distance
vector (AODV) routing, Tech. Rep., 2003.
G. Wang, Y. Ji, D. C. Marinescu, and D. Turgut, A routing protocol
for power constrained networks with asymmetric links, in Proceedings
of the 1st ACM international workshop on Performance evaluation of
wireless ad hoc, sensor, and ubiquitous networks, 2004, pp. 6976.
G. Dommety and R. Jain, Potential networking applications of global
positioning systems (GPS), arXiv preprint cs/9809079, 1998.

978-1-5090-2494-0