Energy Saving and Connectivity Tradeoff by Adaptative Transmission Range
in 802.11g MANETs
Fatiha Djemili Tolba Damien Magoni Pascal Lorenz
University of Haute Alsace, ULP – LSIIT University of Haute Alsace,
68008 Colmar, France 67400 Illkirch, France 68008 Colmar, France
firstname.lastname@example.org email@example.com firstname.lastname@example.org
Abstract in the same transmission range. This last range
determines the range over which the signal can be
Power conservation is a crucial problem in mobile ad coherently received, and is therefore crucial in
hoc wireless networks knowing that, each mobile node determining the performance of the network such as
has a limited amount of energy concentrated in a delay and energy consumption. In such network, each
battery. The main objective of our paper is to use a node is characterized by a well defined quantity of
variable transmission range in order to save some energy. The source of this energy is a battery
energy and to keep the connectivity of the network. implemented in each node. If the battery is discharged
Our algorithm is implemented at the data link layer of the node can not receive or send any packet. So, it is
the OSI model and thus can be used by all MANET necessary to control the transmission range for both
routing algorithms such as AOVD and DYMO. minimizing energy consumption and extending battery
Simulation experiments were conducted to evaluate life.
the performance of our algorithm in terms of energy To seek the best value of transmission range that
used and connectivity. We show that our algorithm preserves connectivity and conserves the energy is an
has an energy gain between 35% to 85% at reasonable important problem for network functionality. A large
speeds below 20 m/s and a high enough network value of transmission range cause consumption of
density. increased energy of the node but preserves the
connectivity, while smaller value causes preserve of
Keyword: mobile ad hoc networks, IEEE 802.11, energy but can adversely impact the connectivity of the
transmission power, transmission range, connectivity. network by reducing the number of active links and,
potentially, partitioning the network , . For this, a
1. Introduction value should be found which makes the compromise
between the connectivity and the consumption of
Today there is a widespread utilization of Mobile Ad energy.
hoc Networks (MANET) in communications. MANET Several papers treat the minimization of transmission
is an autonomous system of mobiles nodes connected range and power control. The first category of paper
by wireless links. The nodes can act as both hosts and aims to find an optimal transmission range in order to
routers since they can generate as well as forward control the connectivity , , ,  or the power
packets. These nodes are also free to move and consumption . The second category aims to find the
organize themselves into a network. MANET does not shortest path with a power based metric using various
require any fixed infrastructure (i.e. a wired or a fixed parameters such as energy consumed per packet or the
wireless base station). The principal characteristics of energy cost per packet . Finally, the third
MANET are the dynamic topology and the limited category aims at modifying the MAC layer  .
energy of mobiles nodes. The interest in such network In this paper, we propose a new protocol for
architecture is focused on battlefield and military controlling the transmission range in mobile ad hoc
applications such as voice and video communications, networks. The first objective of this protocol situated
and also for disaster relief situations. in the conservation of energy knowing that the node is
Ad hoc networks are usually modeled by unit graphs, powered only by a battery. The second objective
where two nodes are connected if and only if they are situated in the preservation of connectivity between the
0-7695-2629-2/06/$20.00 (c) 2006 IEEE
nodes knowing that the connectivity plays an important to 5 m/s). Whereas our algorithm is more general and it
role in the route discovery. The novelty in our is tested for a high mobility (1 to 80 m/s).
proposition lies in the possibility to exploit our In the second category, the power control in the
protocol in all MANET routing algorithms i.e. our routing in ad hoc networks is used by Kawadia and
protocol is generic and completely distributed. Kumar . Each node runs several routing layer agents
The remainder of this paper is organized as follows. that correspond to different power levels. In this
Section 2 reviews some work in this area. Section 3 protocol each node along the packet route determines
presents our contribution and details our proposed the lowest power routing table in which the destination
algorithm. Simulation results presented in section 4 is reachable. However, this protocol is more suitable
demonstrate that the proposed algorithm is better in for network with lowest mobility and the resultants of
terms of energy conservation. Finally, we present our simulation are given only for a fixed density of nodes
conclusion and we discuss future work of (i.e. number of nodes is invariable). In  Spyropoulos
investigation. and Raghavendra proposed an energy-efficient
algorithm for routing and scheduling in ad hoc network
with nodes using directional antennas. The first step of
2. Related work this algorithm consists in finding the shortest cost
paths, using the metric “minimize energy consumed
In the last decade a lot of researchers have per packet”. Next, find the maximum amount of time
contributed in the controlling of energy in MANET. each link can be up, using the metric “maximize
Consequently, a several algorithms using transmission network lifetime”. In the end scheduled nodes’
range have been proposed. An overview of these transmissions by executing a series of maximum
algorithms is presented as follows: weight matching. However, since each node is
In the first category Dai and Wu  proposed three assumed to have a single beam directional antenna, the
algorithms, Prim’s Minimum Spanning Tree, Prim’s sender and the receiver must redirect their antenna
MST with Fibonacci heap implementation and the beam towards each other before transmission and
area-binary in order to find the minimum uniform reception can take place .
transmission range that ensure network connectivity at The idea to change MAC layer is presented in .
same time. However, in these algorithms either each The authors proposed a power control schemes where
node has all information about the network or a the principle is to use two power levels to transmit
specific node has the information about the MST and each data packet: the maximum transmission power for
diffuses it. While, it is more interesting that each node RTS-CTS and the minimum transmit power for
has local information about its neighbors. DATA-ACT. This work has been implemented using
In  Althaus and al. study the problem of omni-directional antenna. Therefore, the scenario is
transmission range in goal to minimize the power completely changed when we use directional antenna
computation for ensure network connectivity. The to transmit and receive signals. Although, Saha and
authors give a minimum spanning tree (MST) based 2- al. propose to use two levels of transmission power
approximation algorithm for Min-Power Symmetric using an antenna operating at omni-directional and
Connectivity with Asymmetric Power Requirements. directional mode. Their work helps to conserve the
In the same problem Santi  proves that the Critical transmission power when the directional transmission
Transmission Range (CRT) in the mobile case is at is used.
least as large as the CRT in case uniformly distributed
Narayanaswamy and al.  proposed a distributed
3. Our contribution
protocol for power control and provided a
conceptualization of this control. This algorithm aims In this section, we introduce our contribution in
to find the smallest common power (COMmon which we give the basic idea. After this, we discuss the
POWer) level at which the network is connected. In the details of the algorithm.
same category, Elbatt et al.  proposed to use the
notion of power management where the study of the 3.1 Basic idea
impact of using different transmission powers on the
The main objective of our protocol is to propose a
average power consumption and end-to-end network
generic solution that can be used by various routing
throughput by controlling the degree of a node.
algorithms such as AOVD, DYMO, etc. Moreover, our
However, this solution is includes in clustering
solution aims at both preserving the energy and
algorithms and it is more suitable for a low mobility (1
maintaining the connectivity of the mobile nodes. The
0-7695-2629-2/06/$20.00 (c) 2006 IEEE
idea proposed in this paper is that each node uses the same time (see figure 3). Note here that a smaller
variable values of transmission range according to the value of transmission range consumes less of energy.
distance between itself and the other nodes.
Our protocol is completely distributed and it takes
into account some features such as transmission range 3 6
and position of the node. In the following, we explain
the choice of each feature.
Transmission range: plays an important role in the
communication between two nodes as mentioned
previously. However, in the mobility model the 5
nodes are free to move in or out the transmission
range that return the precision of its value difficult.
Moreover, a large value of transmission range Figure 2. Transmission range according to the
accrues a consumption of the battery energy. In connectivity.
order to prolong the life span of this last it is
necessary to choose the better value. In the other
hand, the transmission range influenced on the
connectivity of the node. For these reasons the 3
transmission range is a paramount feature in our 5
Position of the nodes: while we work in an
environment where the nodes are mobiles, we must
update the coordinates of the nodes every period of
time. In our protocol, each node broadcasts its
address which is registered by all its neighbors. It is
assumed that a node receiving a broadcast from Figure 3. Variation of transmission range
another node can estimate their mutual distance
from the power level of the signal received. The
3.2 Description of the proposed algorithm
Global Position System (GPS) can be another
solution, but its disadvantage is the consumption of Application layer
Initially and for any topology for ad hoc networks, Transport layer
each node has the same value of transmission range
(i.e. maximum). Of course, this range gives a
maximum connectivity for the nodes (see figure 1).
Data link layer Our protocol
Trmax Physical layer
4 5 2
1 Figure 4. Position of our protocol in the OSI layers.
Initially, we note that our work is focused on level 2
(the Data link layer) of the OSI layers (see figure 4). In
Figure 1.Network topology. the following we describe the proposed algorithm in
After some time we notice that the node number 1 mobile ad hoc networks. Before proceeding with the
changes its transmission range in order to keep always presentation of the various steps of the algorithm we
the same number of neighbors (see figure 2). In the describe the system model. We consider a network
same way, the node number 3 changes its transmission topology which is represented by a graph G = (V, E)
range after a time t. Applying the strategy of variation where V is the set of mobile nodes ( V =m) and e =
of transmission range these nodes (1 and 3) preserve (u, v) ∈ E will model wireless link between a pair of
their connectivity and use the minimum of energy in node u and v only if they are within wireless range of
0-7695-2629-2/06/$20.00 (c) 2006 IEEE
each other. The procedure consists of seven steps as Note here that the update of data is carrying out every
described below: period Δt .
Step 1 In the previous steps, we show in first that the
Each node broadcast data packet with some transmission range is based on the distance between
information’s about its address, position and time the receiver and the sender that allow economizing the
stamp. So, each node will have local information about battery energy. The fact that the node changes its
their neighbors. Initially the transmission range Tr transmission range according to its need (distance) the
takes the value used by the 802.11g for a throughput of battery life span can be prolongs. The realization of
54Mbps. this last is the heart of our work. In the other hand we
Step 2 find that fixed the same connectivity for all nodes
Each node receives this packet, calculates the allows both to form a graphs related and do not charge
distance d between itself and its neighbors using the the node. This facilitates the communication between
received information, as the nodes.
d= (x1 − x 2 )2 + ( y1 − y 2 )2 (1) 4. Simulation
4.1 Simulation topology
Where ( x1 , y1 ) and ( x 2 , y 2 ) are the coordinates of The performance evaluation of our algorithm is
sender and receiver node respectively. made via simulation using the Network Simulator (NS-
Step3 2). We consider a network of n nodes. The nodes are
Recalculate the distance d1 taking into account the uniformly distributed and moved by using the random
speed of the node s max for the time Δt in order to waypoint mobility model . The nodes move in all
envisage the future position of the node. possible directions with speed varying between 1 m/s
to a maximum value (see table 1).
d1 = d + 2 * s max * Δt (2) Parameters Value
Number of nodes 10, 20 and 40
Step4 Area 1000 x 1000 m
Calculate the necessary time for the packet arrived Minimum reception power -70 dBm
to the receiver. Maximum transmission power 18 dBm
Minimum connectivity 2 – 16
t = t current − t stamp (3) Pause time 0s
Maximum speed of the nodes 5 – 80 m/s
Where t current and t stamp are the time current and the
Table 1. Parameters used in simulations.
Step 5 4.2 Performance evaluation
Compare the time t necessary for sending a packet
In these results, each simulation experiment is given
with the time Δt .
for three different node density for networks.
If the time t is inferior or equal to the time-necessary of
Moreover, these results are obtained with a confidence
update all information, so add the sender to the list of level equal to 0.95 and a maximum error threshold
neighbors of the receiver when this list is not saturated
equal to 5%.
(inferior to the minimum number of neighbors).
We measured the following characteristics:
- The energy used for various node density and speed.
If the list of neighbors is empty, so set transmission
- The connectivity factor for various node density and
range to maximum Trmax in order to have a maximum speed.
number of neighbors. Else, set transmission range to
the farthest neighbors distance in order to reduce the Figure 5 shows the speed nodes according to the
transmission range by maintaining the sufficient ratio between the quantity of energy used and the
number of neighbors. maximum energy used. The energy maximum used in
Step 7 our simulation is energy spends if we use a fixed
In the final step, set power level transmission range (that is 802.11 g). We observe that
Ptrans corresponding to the current transmission range. the quantity of used energy increases with the increases
of the speed of the nodes. This is due to the fact, that
0-7695-2629-2/06/$20.00 (c) 2006 IEEE
the moving speed needs more power as distances vary Figure 7 shows the relationship between the energy
more. We conclude that the lower speeds save more used and the connectivity factor. We observe that in
energy. the mathematic terms it exists a proportional relation
i.e. the used energy increases with the increase of the
connectivity factor whatever the value of speed or the
value of nodes number. Therefore, in reality the
relation between these parameters is absolutely the
opposite. This is due to the increase of connectivity
factor that requires an augmentation of the energy used
that must be economized as possible. For this reason, it
is necessary to find a compromise between these two
Figure 5. Energy used vs maximum node speed
The figure below shows the node speed according
to the ratio between the connectivity factor and the
maximum connectivity factor. The connectivity factor
is equal to the inverse of the number of connected
components in the network (i.e related graph).
Concerning the maximum connectivity factor is chosen
when used 802.11g. We observe that the connectivity Figure 7. Energy used vs connectivity factor
factor increases with the increase of node speed. This
due to the fact that when the node moves quickly it
risks to loss their neighbors (i.e. the node can leave its 5. Conclusions and future research
transmission range). Moreover, we observe that if the
connectivity factor bring closer to 1 the nodes can In this paper we presented a new approach to
better communicate in the network. controlling the energy used in ad hoc networks. This
approach is based on the variation of the transmission
range. The transmission range varied according to the
position of the node in the network. So, if the node has
no neighbors, the transmission range takes the
maximum value. Therefore, if the node has a sufficient
number of neighbors (i.e. minimum connectivity) the
transmission range takes the value of the distance
between a sender and a receiver.
The proposed algorithm is simulated in NS-2
environment. Simulation results show that this
approach improves the conservation energy between
35% and 85% when the density of the network exceeds
20 nodes and the speed node is below 20m/s. Contrary
to the energy conservation when we use a fixed
transmission range that is 802.11 g.
In order to continue our investigation in this track,
we will exploit this approach in a routing algorithm.
Figure 6. Connectivity factor vs maximum node speed
0-7695-2629-2/06/$20.00 (c) 2006 IEEE
 F.J. Ovalle-Martinez, I. Stojmenovic, F. Gracia-Nocetti
and J. Solano-Gonzalez, “Finding minimum transmission
radii for preserving connectivity and constructing spanning
trees in ad hoc and sensor networks”, Journal of Parallel and
Distributed Computing, . 2005, pp. 132-141.
 M. Krunz, A. Muqattash and S. J. Lee, “Transmission
Power Control in Wireless Ad Hoc Networks: Challenges,
Solutions, and Open Issue”, Network IEEE, sept-oct 2004,
Vol. 18, pp.08-14.
 Q. Dai and J. Wu, “Computation of Minimal Uniform
Range in Ad Hoc Wireless Networks”, Cluster Computing,
2005, No 8, pp. 127-133.
 E. Althaus, G. Calinescu, I.I. Mandoiu, S. Prasad, N.
Tchervenski and A. Zelikovsky, “Power Efficient Range
Assignement in Ad-hoc Wireless Networks”, IEEE Wireless
Communications and Networking Conference, New Orlean
USA, March 2003.
 P. Santi, “The Critical Trasmitting Range for
Connectivity in Mobile Ad Hoc Networks”, IEEE
Transactions on Mobile Computing, Vol. 4, No. 3, May/June
2005, pp. 310-317.
 S. Narayanaswamy, V. Kawadia, R. S. Sreenivas and P.
R. Kumar, “Power Control in Ad-hoc Networks: Theory,
Architecture, Algorithm and implementation of the
COMPOW Protocol”, proceedings of the European Wireless,
2002, pp. 156-162.
 T. A. Elbatt, S. V. Krihnamurthy, D. Connors and S. Dao,
“Power Management for Throughput Enhancement in
Wireless Ad-Hoc Networks”, IEEE International Conference
on Communications, 2000, pp. 1506-1513.
 V. Kawadia and P. R. Kumar, “Power Control and
Clustering in Ad Hoc Networks”, IEEE INFOCOM, 2003.
 A. Spyropoulos and C. Raghavendra, “Energy Efficient
Communications in Ad Hoc Networks Using Directional
Antennas”, IEEE INFOCOM 2002.
 E. S. Jung and N. H. Vaidya, “A Power Control MAC
Protocol for Ad Hoc Networks”, ACM MOBICOM, 2002.
 D. Saha, S. Roy, S. Bandyopadhyay, T. Ueda and S.
Tanaka, “A Power-Efficient MAC Protocol with Two-Level
Transmit Power Control in Ad Hoc Network Using
Directional Antenna ”, 5th International Workshop on
Distributed Computing IWDC, India, december 2003.
 W. Navidi and T. Camp, “Stationary Distributions for
the Random Waypoint Mobility Model”, IEEE Transaction
on Mobile Computing, vol. 3, n°. 1, January-march 2004.
0-7695-2629-2/06/$20.00 (c) 2006 IEEE