An Improved TMAC Protocol for Wireless Sensor

Networks

Pankhuri Singh Dr. Shirshu Varma

Indian Institute of Information Technology (lIlT) Indian Institute of Information Technology (lIlT)
Allahabad, India Allahabad, India
[email protected] [email protected]

Abstract- In this paper we introduce a new MAC protocol Lifetime of the network can be increased by many ways as
(Improved TMAC) which inherits the features of TMAC illustrated in fig. I but in this paper we only focus on
protocol (Timeout MAC) and makes it more energy-efficient by Contention-based low duty-cycle MAC protocol
adding a new feature to it called POWER SAVER MODE in
implementation.
which TA is reduced by 10% when remaining energy of the
Improved TMAC[I] protocol inherits the features of
sensor node falls below certain threshold value (i.e 15% of its
node's initial energy) and when no activity is taking place.
TMAC protocol and adds POWER SAVER MODE to make it
more energy-efficient and optimized. The extra listen time or
Key words- TMAC, TMACR, Power saver mode, TA, WSN. active time is cut down when no activity is going on and also
when the remaining energy of the particular node falls below
I. INTRODUCTION certain threshold value then its Activation timeout is reduced
by 10%. This reduces the energy wastage in Idle Listening and
Wireless Sensor Network has become an interesting topic of
increases the network's lifetime. The protocol is implemented
research in current scenario. As the network is deployed in a in Omnet++[S] with Castalia -3.2[7] extension.
hostile environment and also sensor nodes used are not
battery operated so saving the energy of the nodes by making II. RELATED WORK
them awake/sleep at appropriate times is really important.
Broadly there are two types of MAC protocols : Contention
This makes the network energy-efficient and increases its
Based and TDMA based.
lifetime.
In TDMA based time is divided into slots and each node is
Some of the major causes of the energy consumption and
assigned predefined slot in which they could transfer the data.
lifetime reduction of the network is as follows:
Z-MAC, TRAMA, TreeMac are some examples of it.
(1) Collision: two nodes transmission interferes with each
Contention based is further divided into Asynchronous and
other.
Synchronous. In this each node contends for the access of the
(2) Overhearing: node receiving the packets not intended
wireless channel and for this it uses CSMA technique. In
for it.
synchronous type nodes form cluster and share common
(3) Idle Listening : No activity for the node but still the
sleep/awake schedules and in asynchronous no common
node is awake.
schedule is synchronized. B-MAC[5], A-MAC, WiseMAC,
RC-MAC are examples of asynchronous protocols and S­
MAC[2], T-MAC[I], R-MAC, DS-MAC[6] are examples of
synchronous protocols.
---L
�_ __""\

III. IMPROVED TMAC PROTOCOL DESIGN

In Wireless Sensor Networks a good MAC Protocol should
contain following features:
Reduction -Mobile (1) Energy-Efficiency : It increases the lifetime of the
-Data Relay network.
Acquisition
(2) Latency: Less delay which is the basic requirement in
real time applications.
Contention based
(3) Fairness : Each node in the network gets the equal
MAC protocol chance to transmit the data to the sink.
TDMA based (4) Throughput: Maximum number of packets should be
transmitted and received successfully.
(Fig I: Lifetime increasing schemes in WSN)

978-1-4799-3140-S/14/$31.00 ©2014 IEEE 91

 Improved TMAC inherits the features of TMAC Protocol and cannot reply because of overheard CTS from other
adds some additional attributes to reduce the idle listening node. And if for TA time the sending node does not
time. receive any CTS then it goes to sleep which would not
(I) Dynamic Duty-Cycle : Under varying load be fair in last two cases. This results to decrease in
conditions fixed Duty-Cycle will lead to missing of successful transmissions. So, at least two RTS retries
packets if traffic is high and idle listening if traffic is must be made.

:
less. So Dynamic duty-cycle is used which introduces RTS DATA ACK
CTS
a new term TA (Activation Timeout). If no event

SZ\\'\/
occurs of time TA then the active period will end . ro",,"d

\. "
TA> CI+RTS+T --+eqn(J)

Where CI is Contention Interval, RTS is length of ""I,,d

\ \\\
RTS packet and T is Turn-Around time (time between C
RTS packet's end and CTS packet's start). TA

(Fig 3: Data exchange through RTS/CTS)

NORMAL

Active period
(6) FRTS (Future request to send) : In unidirectional
communication pattern problem of Early sleeping
Improved
TMAC arises. In this due to overheard CTS or RTS the node
TA TA Sleep period TA does not take part in communication and goes to sleep
state and may miss the packets destined for it.
RTS CTS DATA ACK
(Fig 2: Dynamic duty-cycle of Improved TMAC)

(2) Power Saver mode : When the remaining energy of

the node falls below some threshold level (15% of the
initial energy of the node) then TA is reduced by
10%. This is called Power saver mode which extends
C

+
the lifetime of node. And also when the channel is
busy, Carrier sense is not valid, no event is taking
place TA is reduced to avoid idle listening. Active Sleep
------���
D �4==�==�_�------ RTS?
TA
(3) Virtual Clustering : Frame synchronization is
(Fig 4: Problem of Early sleeping of node D)
achieved by virtual clustering. When a node joins the
network, it waits and listens. And if for a certain
In Fig 4 the communication flow is unidirectional
amount of time it does not hear any SYNC packet
(from node A to node B and so on). According to
from any node which contains the schedule, it creates
node C , node D also sets it active and sleep periods
its own schedule and transmits it to the rest nodes in
by overhearing CTS from node B. Therefore, node D
SYNC packet. A node adopts both schedules if a node
misses the RTS destined for it. To avoid this problem
having its own schedule receives the schedule in
of Early Sleeping of nodes due to overhearing of
SYNC packet from other node. In such a case,
CTS/RTS we use FRTS.
Activation event will occur at start of both the frames
but actual data transmission will take in their
In Fig 5 use of FRTS is shown. Node C receives CTS
respective active periods.
packet from node B and then immediately sends
FRTS packet to node D to inform it that it is the future
(4) RTS/CTS : Data exchange is by using RTS, CTS,
receiver for RTS so it must be awake by that time.
ACK and SYNC packets. Node C does the interfere
While sending FRTS packet to node D, it may collide
between the communication between nodes A and B
with the DATA packets, so at that time DS (Data­
because it overhears the CTS and sets its sleep period
send) packet is send which does not contain any
accordingly. This avoids collisions.
relevant information and is of short duration which
only reserves the medium for data transmission and
(5) RTS Retries: In this 2 RTS retries is given to avoid
avoids collision between DATA packets and FRTS
the decrease in throughput. When CTS is not received
packet.
in reply of RTS then may be node is in sleep state, or
due to collision it cannot hear the RTS , or the node

92 20 14 International Conference on Signal Propagation and Computer Technology (ICSPCT)

 RTS CTS DS DATA ACK • For Entering into POWER SAVER MODE and
saving the Energy consumed.

fW"'�\Z�
!
_ '_
� _ Ac
_t �
==V _
:'\ \.
...

_ _ _
... .... ....

i
: \....

i
...
. .... .. .... :/1
:

i �4�=�=�TA=:'::::�.;-·4�:�FRTS::.� ____ RTS

__

(Fig 5 : Use FRTS to avoid problem of early sleeping of nodes)

IV. ALGORITHM USED

• For Creating and updating the Schedule for Node.
Node's
From Sensor Node Energy:S
15% initial
energy

(Fig 7: Flow Chart for Power saver mode)

Insert Primary
Schedule in Schedule
V. SIMULATION RESULTS
Table
Only one scenario is tested for both protocols :

Broadcast Time of
TMAC and Improved TMAC in Omnet++ with Castali-3.2
next Frame
extension. 100 number of nodes are deployed randomly in
field size of 1OOx100m2•
Rest parameters used for both the protocols are mentioned
in table below:
From Another
Sensor Node Paramet ers TMAC Improved TMAC

Initial energy 18720 Joule 18720 Joule

of each node
Activation 15ms Varying (reduction
Timeout required in power
Create Secondary saver mode)
Schedule
Use FRTS NO YES
Ack packet 11 bytes 11 bytes
size
Update Schedule SYNC packet 11 bytes 11 bytes
Stay on Primary Table size
Schedule of the node
CTS/RTS 13 bytes 13 bytes
packet size
(Fig 6: Flow Chart for creating schedule)

20 14 International Conference on Signal Propagation and Computer Technology (lCSPCT) 93

 Contention 10ms 10ms Reception of Packets. There is not much difference between
Period both of them but Improved TMAC is able to receive more
Sync Time 6ms 6ms packets successful1y as compared to TMAC. Therefore,
Throughput is little bit more in TMACR.
Frame time 610 ms 610ms
RX pkt breakdcmn
Simulation 600s 600s 4500
TMAC_
Time TMACR _
4000
Power saver Not available Available
mode
3500
Threshold Not required 15% of initial
value energy 3000

Energy used 9.061 mJ 7.149 mJ

2500
in 600s
(FIg 8: Table of Parameters used) 2000

1500

(I) Sent Packets Breakdown :

1000

nn
100
Sent packets breakdo.vn 500
In
TMAC_
TMACR _ r"'II._
95

90
(Fig 10: Average number of Received packets by each
node)
85

80

75 (3) Consumed Energy :

Consumed Energy
70
10.5 mrrmmTTTTTmTl'ITT1TTmTTmm-rmm

65
10

60
9.5

55

50 r::::::J,r-
DATA SYNC
8.5
(Fig 9: Average number of DATA and SYNC packets sent by
nodes during respective protocols)

Fig 9 shows that SYNC packets sent is almost same in both

cases (100 SYNC packets) but the DATA packets sent slightly
differ. Somewhat more in TMAC as compared to TMACR
(Improved TMAC).

(2) Received Packet breakdown :

Six things can happen to receive packet: Nodes (0 -99)
• Failed with interference - 1
• Failed without interference - 2 (Fig II: Consumed energy by each node in both protocols)
• Failed below sensitivity - 3
• Failed , non RX state - 4 It is clear from fig 11 that more energy is consumed in
case of TMAC protocol as compared to TMACR
• Received despite interference - 5
(Improved TMAC). This is achieved by cutting the extra
• Received without interference - 6
Activation Timeout which led to reduction in idle
Graph in fig 10 shows the comparison between both Protocols
listening. And also when remaining energy is less than
in terms of failed Reception of Packets and Successful
15% of initial energy of the node then it goes Power saver

94 20 14 International Conference on Signal Propagation and Computer Technology (ICSPCT)

 mode and cuts the Activation Timeout by 10%. Therefore,
energy consumption is reduced by 25% in TMACR. [2] Ye W, Heidemann J, Estrin D. An energy-efficient MAC
protocol for wireless sensor networks. In: IEEE
(4). Latency : INFOCOM, New York; 2002. p. 1567-76.
More delay is observed in TMAC as compared with
TMACR. [3] Smriti joshi, Anant Kr. Jaiswal and Pushpendra Kr. TyagiJ
"A Novel Analysis of T Mac and S Mac Protocol for
A�llcalion lave/latency, In ms Wireless Sensor Networks Using Castalia" in International
Journal of Soft Computing and Engineering (USCE)
ISSN: 2231-2307, Volume-2, Issue-6, January 2013.

[4] Bhavana Narain, Anuradha Sharma, Sanjay Kumar and

Vinod Patle " Energy effiecient MAC protocols for
wireless sensor networks: a survey" International Journal
of Computer Science & Engineering Survey (UCSES)
Vol.2, No.3, August 2011.

[5] Polastre J, Hill J, Culler D. Versatile low power media

access for wireless sensor networks. In: Proceedings of
the second international conference on embedded
networked sensor systems (SenSys'04); 2004. p. 95-107.
0
(0,20] [20,4OJ [40,60] 160,SO] [SO,l00J [too,120) (120,140] (140,160] [160,18OJ [lSO,200] (200,220]
[6] Lin P, Qiao C, Wang X. Medium access control with a
(Fig 12: Latency or delay per node, ms)
dynamic duty cycle for sensor networks. In: Proceedings
of wireless communications and networking conference
(WCNC); 2004. p. 1534-9.
VI. CONCLUSION
From above results it is clear that Improved TMAC
[7] http://castalia.npc.nicta.com.aui
(TMACR) performs better than TMAC. Both are good
under varying traffic conditions but introduction of POWER
[8] www.omnet.org
SAVER MODE in TMACR makes it more energy-efficient.
There is not much difference if consider sent packets
breakdown, transmitted packets and received packets but if
we consider energy consumption and latency then TMACR
outperforms TMAC.
Future work will be the Hardware implementation of
TMACR on Tmote by varying different parameters and
observing the results.

VII. RFERENCES

[1] Tijs van Dam and Koen Langendoen, "An Adaptive

Energy Efficient MAC Protocol for Wireless Sensor
Networks ", in Proceedings of the first ACM conference
on embedded networked sensor systems (SenSys'03);
November 5-7, 2003, Los Angeles, California, USA.

2014 International Conference on Signal Propagation and Computer Technology (ICSPCT) 95