An Adaptive Energy Efﬁcient MAC Protocol for Wireless Sensor

Network

Van Thiep Nguyen, Matthieu Gautier and Olivier Berder

University Rennes 1, IRISA, INRIA,

6 Rue Kerampont, 22300 Lannion, France

1 RESEARCH PROBLEM

In last years, the wireless sensor networks (WSNs)

have take the interest of researchers since they have

a number of potential applications in many domains

of daily-life such as environment monitoring, healthy

monitoring, medicine, surveillance, military, etc. The

innovation of digital electronics, micro controller,

wireless communications have facilitated the devel-

opment of small tiny devices with these characteris-

tics: low-cost, low-power and multi-functional. Due

to the small size and the difﬁcult access of a sensor

node, the power source has limited capacity and can-

not be replaced (non-replaceable). Therefore, energy

consumption is an important constraint in WSNs. En-

ergy harvesting (from many natural environment re-

sources such as the heating, the sun light, the wind

etc.) provides an effective option to increase the life-

time of the sensor network but it is not a complete so-

lution. In some cases, these resources are not enough

to be ensure a sufﬁcient reliability. Thus the reduction

of energy consumption is still one important focus of

research in WSNs domain.

The reduction in energy consumption can be con-

sidered in both software as well as hardware ap-

proach. As far as the hardware components are con-

cerned, radio transceivers integrated circuit has kept

important role in reducing energy consumption. The

radio chips are classiﬁed into two types, one focuses

on optimizing receive current consumption but the

transmit current is higher such as 8mA for receive cur-

rent and 25mA for transmit current. The second type

keeps the balance between receive and transmit cur-

rent, which are from 17mA to 19mA. Although the in-

novations of technologies in radio transceivers in term

of reduction in energy consumption the transceivers

are still required to be improved. In the software point

of view, the protocols not only within a single layer

but also through cross-layer have achieved major im-

provements for extending lifetime of the sensor node.

Moreover, the impact of the software controlling the

hardware (also called as ﬁrmware) is indispensable,

so there is room for the software part of a sensor node

to perform better. This thesis focuses on lower layers

of WSNs, especially on the access protocols to reduce

energy consumption.

2 OUTLINE OF OBJECTIVES

In the context of energy aware WSNs, the main

causes of energy consumption are: collision, over-

heads, overhearing, idle listening (Bachir et al.,

2010). Based on these causes of energy consump-

tion, the various MAC protocols could be classiﬁed

into these groups: reducing collision, reducing over-

heads, reducing overhearing, reducing idle listening.

In these regards, our objective is to reduce energy con-

sumption in approach of reducing idle listening. An

energy-efﬁcient MAC protocol for cooperative strate-

gies in WSNs will be proposed. Our works are sepa-

rated into two steps:

First step is to propose an adaptive energy-

efﬁcient MAC protocol for the direct communica-

tion between two sensor nodes. The receiver tries to

adapt its wakeup interval according to the trafﬁc of

the sender.

Second step is to apply the MAC protocol pro-

posed in ﬁrst step to the cooperative strategies. In this

approach, there is a cooperative relay node, which re-

sends the data packet on demand of receiver if needed.

The network prototype is described in the Figure 1. In

cooperative strategy, the receiver needs to adapt with

not only the sender but also the relay node.

Figure 1: Cooperative relay strategy.

Nguyen V., Gautier M. and Berder O..

An Adaptive Energy Efﬁcient MAC Protocol for Wireless Sensor Network.

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

3 STATE OF THE ART

The Medium Access Control (MAC) layer is software

integrated in a sensor node, which allows this node to

efﬁciently share the wireless medium with others in

the network. In MAC layer, the main causes of energy

consumption are idle listening, overheads, overhear-

ing and collision. Thus, in order to achieve the en-

ergy efﬁciency, these factors need to be optimized but

there are trade-off between them. For example, reduc-

ing idle listening and collision requires extra synchro-

nization and overheads, whereas, reducing the syn-

chronization and overheads causes the waste of en-

ergy in collisions. In the context of energy-efﬁcient

MAC protocols, an important mechanism for reduc-

ing energy consumption is duty cycling. In this tech-

nique, the radio is turned on periodically, switching

between awake and sleeping state. The duty cycle,

which is measured as the ratio of time a node is awake

to the total time, is used also to evaluate the perfor-

mance of a protocol. The recent duty cycling MAC

protocols can be grouped into two types: synchronous

and asynchronous.

The synchronous duty cycling MAC protocols

(such as SMAC(Ye et al., 2002), TMAC(van Dam

and Langendoen, 2003)) reduce energy consumption

by synchronizing the sleep & wakeup time of sen-

sor nodes. After the synchronization, the idle lis-

tening problem is resolved but in the synchronization

process, sender and receiver must exchange control

packets. As an example in SMAC, the control pack-

ets CTS/RTS/SYNC are sent between sensor nodes to

synchronize. In addition, using ﬁxed time (time slot)

for sleeping and listening state is inefﬁcient with vari-

able trafﬁc rate.

In contrast, the asynchronous duty cycling pro-

tocols do not require any synchronization period

and can be categorized into two groups: sender

initiated and receiver initiated. In sender initiated

Figure 2: TAD-MAC protocol.

MAC protocols, e.g. BMAC (Polastre et al., 2004),

XMAC(Buettner et al., 2006), WiseMAC(El-Hoiydi

and Decotignie, 2004), the sender initiates the com-

munication by sending the preamble-sampling packet

before a data transmission to notify the receiver of up-

coming packet. Instead of using long preamble as

in BMAC, XMAC protocol uses the series of short

preamble packet and ACK packet is used just right

after the reception of ﬁrst preamble packet from re-

ceiver side to signify this sensor node is awake. On

the other hand, the preamble-sampling packet is re-

placed by wakeup beacon, which is sent from re-

ceiver in the receiver initiated MAC protocols such

as PW-MAC(Tang et al., 2011), RICER(yi A. Lin

and Rabaey, 2004), RI-MAC(Sun et al., 2008), TAD-

MAC(Alam et al., 2012). The wakeup beacon is

shorter than preamble so the wireless bandwidth us-

age and collision are reduced. The basic principle of

TAD-MAC protocol is described in. The ﬁgure is sep-

arate into two phases (i.e., before convergence and af-

ter convergence) that outline the results of a simple

network with two nodes (Tx1 and Tx2) attempting to

transmit data to a receive node. During the evaluation

phase, before reaching the convergence, the receive

node tries to adapt its wakeup-interval (I

) to the

data transmission rate of each transmit nodes. After

several wake-ups, the receive node will adapt its I

based on the statistics of trafﬁc that it receives from

each individual transmit node. The second phase, af-

ter convergence, indicates that the I

of receive node

has been adapted to trafﬁc of each transmit node in a

way that the idle listening is minimal.

4 METHODOLOGY

To design energy-efﬁcient WSNs, we must be able to

explore all the parameters idle-listening, overheads,

overhearing and collision (Bachir et al., 2010). The

general methodology that can be applied to this ex-

ploration is shown in Figure 3. We consider the con-

Figure 3: General methodology.

SENSORNETS2015-DoctoralConsortium

text of an already established network, not in the net-

work development phase and the network topology

is ﬁxed. With this predeﬁned network topology, we

choose one parameter to optimize and observe the in-

ﬂuence on other parameters by estimating energy con-

sumption. After several repetitions, we can ﬁnd a set

of optimal parameters. To evaluate the performance

of proposed protocol, TAD-MAC is chosen and im-

plemented also to compare the simulation results.

5 STATE OF RESEARCH

5.1 Introduction

In a WSNs platform, typically, the radio transceiver

consumes most of energy. The radio activities are

controlled by Medium Access Control (MAC) layer.

Therefore, it is necessary to design an energy-efﬁcient

protocol suitable for WSN. In the context of energy-

efﬁcient WSN, low duty-cycle protocols such as

preamble sampling MAC protocols are very efﬁcient

for low trafﬁc applications because these protocols

improve the lifetime of the network by reducing the

unnecessary energy waste. In the ﬁrst step of our

work, we proposed an adaptive energy-efﬁcient MAC

protocol called FTA-MAC (Fast Trafﬁc Adaptive).

FTA-MAC is an asynchronous duty cycling protocol

with receiver initiated mechanism.

5.2 FTA-MAC Protocol Design

The basic principle of FTA-MAC is described in Fig-

ure 4. As other receiver initiated protocols, the re-

ceiver sends the wakeup beacon packet right after its

wakeup to notify to other sensor nodes. But unlike

TAD-MAC protocol, in FTA-MAC, the Wakeup Bea-

con (WB) does not contain the destination address, it

uses broadcast WB. At the sender side, when it re-

ceives WB from the receiver, the sender senses the

wireless medium for an interval of time called CCA

(Channel Clear Assignment) to make sure that the

channel is free to send the data. Normally the time

of CCA is ﬁxed as in other MAC protocols, but for

FTA-MAC protocol, if the CCA is ﬁxed, there exist

collisions caused by broadcast WB. In the case there

are two or more senders, which receive a same WB,

they will be in conﬂict for data transmission. To avoid

this collision, a function is proposed for senders to

calculate the time of CCA each time they receive the

WB as

CCA

= t

CCAmax



1 −

idle

W Bmax



(1)

Figure 4: Basic principal of FTA-MAC protocol.

where t

CCAmax

is the constant value maximal for the

time of CCA, idle is the interval of time that the

sender waits for WB and t

W Bmax

is the constant value

for the maximal time waiting WB.

The main principle of FTA-MAC protocol is to

adapt the wakeup interval (I

) of the receiver ac-

cording to the data transmission rate of each transmit

node. A Trafﬁc Status Register (TSR) is used to store

the status of data transmission in receiver. Each time

the receiver receives a data packet, a bit 1 is inserted

into TSR by shifting left one bit. In contrast, if the re-

ceiver does not receive a data packet in one wakeup,

the bit 0 is inserted. A list of TSR called TSR-Bank is

used to estimate the data transmission of all neighbor

nodes as in Figure 5. The next I

is calculated by an

adaptive function as

(i + 1) =

(

(i) + n

(i).t

re f

TSR(i) = 0

∑

j=k

( j)+idle

−idle

wbMissed+1

TSR(i) = 1

(2)

where n

(i) is the number of bits 0 in TSR, t

re f

the system clock factor, k index stands for the last

moment when the receive node received data from

sender, idle

& idle

are the idle time of sender for

two last time it received WB and wbMissed is the

number of wakeup without receiving WB. The three

values idle

, idle

and wbMissed are calculated in

sender side and sent to receiver by including in data

packet. The two ﬁrst bytes in the payload part of data

packet are used to store these values. These parame-

ters are described more detail in Figure 6 and Figure

AnAdaptiveEnergyEfficientMACProtocolforWirelessSensorNetwork

5.3 Implementation

In this section, the implementation details of FTA-

MAC protocol in a network simulator are explained.

One of the network simulators, which support wire-

less sensor network, is OMNeT++/MiXiM. OM-

NeT++ is an object-oriented modular event network

simulation and it provides the infrastructure for writ-

ing simulations with the component architecture. All

elements are called modules and the modules can be

connected via gates (or ports) and communicate by

exchanging messages. MiXiM is a modeling frame-

work for wireless network (WSNs, body area net-

work, ad-hoc network, vehicular network, etc).

In OMNeT++/MiXiM, there is a deﬁnition of net-

work node called WirelessNodeBattery which con-

tains the speciﬁc elements for a sensor node such as

the network interface WirelessNicBattery, the power

supply SimpleBattery and the module BatteryStats

which calculates the consumption of energy for each

element. The main implementation of FTA-MAC

protocol is FTAMACLayer, which is a module of

the MAC layer that is integrated in a new net-

work interface NicFTAMAC (based on WirelessNic-

Battery). As the other modules in MiXiM, FTA-

MACLayer contains the gates to connect with other

modules/other layers. The messages are used not only

as the packets to send between the layers of a sen-

sor node or between two nodes but also as the events

to change the state of a node. The state machines

Figure 5: TSR-Bank containing N registers for N neighbor

nodes.

Figure 6: Adaptive function to calculate I

Table 1: Simulation conﬁguration.

Scenario 1 Scenario 2 Scenario 3

simTime 1000s 1000s 1000s

W Bmax

250ms 500ms 500ms

TxInt 500ms 500ms variable

of transmit node and receive node shown in Figure

8(a) and 8(b) describe all the self-messages and the

network packets that are used in FTAMACLayer to

change the state of a sensor node.

The new structure of MAC frame is proposed in

Figure 7. The two ﬁrst bits of frame control byte (the

ﬁrst byte of MAC frame) are used to deﬁne the type

of frame (WB, DATA or ACK). In WB frame, the

address information needs 4 bytes to store only the

source address because the WB is broadcast frame.

The size of WB frame is very small, only 5 bytes of

length (1 byte for frame control and 4 bytes for source

address) so the energy consumption is reduced. In

DATA frame, the frame length is displayed in 6 bits of

frame control byte and the two ﬁrst bytes of payload

part are used to store the idle and wbMissed variables.

5.4 Performance Evaluation

To evaluate the performance of FTA-MAC protocol,

the simulation scenarios are executed in OMNeT++

and compare with TAD-MAC. The conﬁguration of

simulation scenarios are shown in Table 1. In each

scenario, the value I

is variable to study the perfor-

mance of protocol by observing the parameters: con-

vergence time of I

, energy consumption, specially,

in the scenario 3, the trafﬁc is variable.

The ﬁrst results in Figure 9(a) and 9(b) show that

FTA-MAC reach the steady state faster than TAD-

MAC, specially, in scenario 1 with the t

W Bmax

250ms, the convergence value of I

in TAD-MAC

is wrong if the value I

initial is equal or greater than

700ms. It converges at 750ms instead of 250ms

When comparing energy consumption, FTA-

MAC can optimize the energy consumption in trans-

mit node (as in Figure 10(a)) but on the other side, in

Figure 7: New MAC frame structure.

SENSORNETS2015-DoctoralConsortium

(a) transmit node. (b) receive node.

Figure 8: State machine of FTA-MAC.

(a) scenario 1. (b) scenario 2.

Figure 9: I

convergence time.

receive node, FTA-MAC consumes more energy. This

is because the receiver needs to keep awake longer to

wait more for data packet so it cannot switch off the

radio right after sending the ACK packet. The result

of comparing energy consumption in Figure 10 points

on the reducing of energy consumption in the network

(the total of transmit node and receive node) is efﬁ-

cient in FTA-MAC.

The results of simulation scenario 3 shows that

FTA-MAC protocol works well for variable trafﬁc.

The Figure 11(b), 11(c) and 11(d) are zooms of Fig-

ure 11(a) at the moment when trafﬁc is changed. After

few wakeups, the receiver can adapt to the new trafﬁc

rate of sender.

6 CONCLUSION AND

PERSPECTIVE

In the ﬁrst step of our works in this thesis, we pro-

posed an adaptive energy-efﬁcient MAC protocol for

WSNs. In this protocol, the sensor node adapts its

wakeup interval dynamically with the trafﬁc load it

receives and consequently optimizes the energy con-

sumption. This protocol is implemented in network

simulator OMNeT++ and the simulation result shown

that FTA-MAC outperforms other MAC protocols

such as TAD-MAC.

In last years, cooperative technology gains some

success result in term energy-efﬁcient for WSNs. Our

future works will focus on the cooperative strategies

and try to apply in FTA-MAC.

AnAdaptiveEnergyEfficientMACProtocolforWirelessSensorNetwork

(a) Energy consumption in sender. (b) Energy consumption in network.

Figure 10: Energy consumption for sending 2000 packets in 1000s.

(a)

(b) (c) (d)

Figure 11: I

with variable trafﬁc.

REFERENCES

Alam, M. M., Berder, O., Menard, D., and Sentieys, O.

(2012). Tad-mac: Trafﬁc-aware dynamic mac proto-

col for wireless body area sensor networks. IEEE J.

Emerg. Sel. Topics Circuits Syst., 2(1):109–119.

Bachir, A., Dohler, M., Watteyne, T., and Leung, K. K.

(2010). Mac essentials for wireless sensor networks.

Commun. Surveys Tuts., 12(2):222–248.

Buettner, M., Yee, G. V., Anderson, E., and Han, R. (2006).

X-mac: A short preamble mac protocol for duty-

cycled wireless sensor networks. In Proceedings of

the 4th International Conference on Embedded Net-

worked Sensor Systems, SenSys ’06, pages 307–320,

New York, NY, USA. ACM.

El-Hoiydi, A. and Decotignie, J.-D. (2004). Wisemac: An

ultra low power mac protocol for the downlink of in-

frastructure wireless sensor networks. In Proceedings

of the Ninth International Symposium on Computers

and Communications Volume 2, ISCC ’04, pages 244–

251, Washington, DC, USA. IEEE Computer Society.

Polastre, J., Hill, J., and Culler, D. (2004). Versatile low

SENSORNETS2015-DoctoralConsortium

power media access for wireless sensor networks. In

Proceedings of the 2nd International Conference on

Embedded Networked Sensor Systems, SenSys ’04,

pages 95–107, New York, NY, USA. ACM.

Sun, Y., Gurewitz, O., and Johnson, D. B. (2008). Ri-mac:

A receiver-initiated asynchronous duty cycle mac pro-

tocol for dynamic trafﬁc loads in wireless sensor net-

works. In Proceedings of the 6th ACM Conference

on Embedded Network Sensor Systems, SenSys ’08,

pages 1–14, New York, NY, USA. ACM.

Tang, L., Sun, Y., Gurewitz, O., and Johnson, D. B. (2011).

Pw-mac: An energy-efﬁcient predictive-wakeup mac

protocol for wireless sensor networks. In Proceedings

of the 30th IEEE International Conference on Com-

puter Communications, INFOCOM ’11, pages 1305–

1313.

van Dam, T. and Langendoen, K. (2003). An adaptive

energy-efﬁcient mac protocol for wireless sensor net-

works. In Proceedings of the 1st International Confer-

ence on Embedded Networked Sensor Systems, Sen-

Sys ’03, pages 171–180, New York, NY, USA. ACM.

Ye, W., Heidemann, J., and Estrin, D. (2002). An energy-

efﬁcient mac protocol for wireless sensor networks. In

INFOCOMM.

yi A. Lin, E. and Rabaey, J. M. (2004). Power-efﬁcient

rendez-vous schemes for dense wireless sensor net-

works. In In IEEE ICC, pages 3769–3776.

AnAdaptiveEnergyEfficientMACProtocolforWirelessSensorNetwork