Content uploaded by Nadeem Javaid
Author content
All content in this area was uploaded by Nadeem Javaid on Feb 09, 2019
Content may be subject to copyright.
Adaptive-Reliable Medium Access Control
Protocol for Wireless Body Area Networks
By
Mr. Azizur Rahim
Registration Number: CIIT/SP11-REE-027/ISB
MS Thesis
In
Electrical Engineering
COMSATS Institute of Information Technology
Islamabad – Pakistan
Spring, 2012
Adaptive-Reliable Medium Access Control
Protocol for Wireless Body Area Networks
A Thesis presented to
COMSATS Institute of Information Technology
In partial fulfillment
of the requirement for the degree of
MS (Electrical Engineering)
By
Mr. Azizur Rahim
CIIT/SP11-REE-027/ISB
Spring, 2012
COMSATS Institute of Information Technology
ii
Engineering).
Adaptive-Reliable Medium Access Control
Protocol for Wireless Body Area Networks
A post Graduate Thesis submitted to Department of Electrical Engineering as
partial fulfillment of the requirement for the award of Degree of M.S
(Electrical
Supervisor:
Dr. Nadeem Javaid,
Assistant Professor,
Department of Electrical Engineering,
COMSATS Institute of Information Technology (CIIT)
Islamabad Campus
June, 2012
Name Registeration Number
Mr. Azizur Rahim CIIT/SP11-REE-027/ISB
iii
Final Approval
This thesis titled
Adaptive-Reliable MediumxAccessxControl
Protocol forxWirelessxBodyxAreaxNetworks
By
Mr. Azizur Rahim
CIIT/SP11-REE-027/ISB
Has been approved
For the COMSATS Institute of Information Technology, Islamabad
External Examiner: __________________________________
Supervisor: ________________________
Dr. Nadeem Javaid /Assistant professor
Department of Electrical Engineering
Islamabad Campus
Co-supervisor: ________________________
Dr. Safdar H.Bouk / Assistant professor
Department of Electrical Engineering
Islamabad Campus
HoD: ________________________
Dr. Shafayat Abrar / Associate professor
Department of Electrical Engineering
Islamabad Campus
iv
Declaration
I Mr. Azizur Rahim, CIIT/SP11-REE-027/ISB herebyxdeclare that I havexproduced
the workxpresented inxthis thesis, duringxthe scheduledxperiod of study. I also declare
that I havexnot taken anyxmaterial from anyxsource exceptxreferred toxwherever due
that amountxof plagiarism isxwithin acceptablexrange. If a violationxof HEC rulesxon
research hasxoccurred in thisxthesis, I shall be liablexto punishablexaction under the
plagiarismxrules of the HEC.
Date: ________________
________________
Mr. Azizur Rahim
CIIT/SP11-REE-027/ISB
v
Certificate
It is certified that Mr. Azizur Rahim, CIIT/SP11-REE-027/ISB hasxcarried out
allxthe work relatedxtoxthis thesisxunder myxsupervision at the Departmentxof
ElectricalxEngineering COMSATSxInstitute ofxInformation Technology, xIslamabad
and thexwork fulfills thexrequirements for award of MSxdegree.
Date: _________________
Supervisor:____________________
Dr. Nadeem Javaid /Assistant professor
Department of Electrical Engineering
CIIT Islamabad Campus
Head of Department:
____________________________
Dr. Shafayat Abrar/Associate professor
HoD Electrical Engineering
vi
DEDICATION
Dedicated to my elder brother, Muhammad Rahim
vii
ACKNOWLEDGMENT
I am heartily grateful to my supervisor, Dr. Nadeem Javaid, whose patient
encouragement, guidance and insightful criticism from the beginning to the final level
enabled me have a deep understanding of the thesis.
Lastly, I offer my profound regard and blessing to everyone who supported me in
any respect during the completion of my thesis especially Dr. Safdar H.Bouk and my
friends in every way offered much assistance before, during and at completion
stage of this thesis work.
Mr. Azizur Rahim
CIIT/SP11-REE-027/ISB
viii
ABSTRACT
This thesis presents Adaptive-Reliable MediumxAccessxControl (AR-MAC)
protocolxfor WirelessxBody AreaxNetworks (WBANs). In WBANs, small battery-
operated on-body or implanted biomedical sensor nodes are used to monitor
physiological signs such as temperature, blood pressure, ECG, EEG etc. Proposed
protocol is based upon fixed topology of WBAN to use TimexDivisionxMultiple
Access (TDMA) approach for channel access with a novel scheme of synchronization.
All nodes remain in sleep mode until the time slot assigned by Central Node, to avoid
idle listening and overhearing. An adaptive guard band algorithm is used to avoid
collision due to clock drift of nodes. Simulationxresults showxthat proposed AR-MAC
outperformsxthan IEEE 802.15.4 inxterms of energyxconsumptionxand reliability.
ix
TABLE OF CONTENTS
1 Introduction 2
1.1 Wireless Body Area Networks (WBANs) . . . . . . . . . . . . . . . 2
1.2 Motivation................................ 3
1.3 Scope .................................. 3
1.4 Research methodology . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Research Contribution . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 Thesisorganization........................... 4
2 Introduction to WBANs 6
2.1 Wireless Sensor Networks . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.1 Network Topology . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.2 Transmission Media . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Wireless Body Area Networks . . . . . . . . . . . . . . . . . . . . . 8
2.3 WirelessStandards ........................... 8
2.3.1 IEEE 802.15.1/Bluetooth . . . . . . . . . . . . . . . . . . . 10
2.3.2 IEEE 802.15.4 and ZigBee . . . . . . . . . . . . . . . . . . . 10
2.4 Comparison of wireless networking standard . . . . . . . . . . . . . 12
3 Related Work 14
3.1 WBANs ................................. 14
3.2 WBANArchitecture .......................... 15
3.2.1 Overview ............................ 15
3.2.2 Design Requirements for WBANs . . . . . . . . . . . . . . . 16
3.2.2.1 Energy efficiency . . . . . . . . . . . . . . . . . . . 17
3.2.2.2 Reliability....................... 17
3.2.2.3 Scalability....................... 17
3.2.2.4 Quality of Service (QoS) . . . . . . . . . . . . . . . 17
3.3 Sources of Energy Dissipation in WBANs . . . . . . . . . . . . . . . 18
3.4 Classification of MAC protocols for WBANs . . . . . . . . . . . . . 19
3.4.1 Contention-Based MAC Protocols . . . . . . . . . . . . . . . 19
3.4.2 Contention-Free MAC Protocols . . . . . . . . . . . . . . . . 20
3.4.3 Low Power Listening (LPL) MAC Protocols . . . . . . . . . 21
3.5 MAC protocols for WBANs . . . . . . . . . . . . . . . . . . . . . . 23
3.5.1 IEEE 802.15.4 MAC protocol . . . . . . . . . . . . . . . . . 23
3.5.2 Battery-aware TDMA protocol . . . . . . . . . . . . . . . . 24
3.5.3 Priority guaranteed MAC protocol . . . . . . . . . . . . . . 24
3.5.4 Energy-Efficient Low Duty Cycle MAC Protocol . . . . . . . 25
x
3.5.5 A power-efficient MAC protocol for WBAN . . . . . . . . . 26
3.5.6 Energy Efficient Medium Access Protocol . . . . . . . . . . . 27
3.5.7 BodyMAC............................ 29
3.5.8 MedMAC ............................ 30
3.5.9 Heartbeat-Driven MAC protocol . . . . . . . . . . . . . . . . 31
3.6 Discussion and Open Research Issues . . . . . . . . . . . . . . . . . 34
4 Proposed MAC Protocol 37
4.1 Protocoldesign ............................. 37
4.1.1 Channel Selection . . . . . . . . . . . . . . . . . . . . . . . . 38
4.1.2 Time Slot Assignment . . . . . . . . . . . . . . . . . . . . . 38
4.1.3 Synchronization . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.1.4 FrameFormate ......................... 41
4.2 Energy Consumption Analysis . . . . . . . . . . . . . . . . . . . . . 42
4.2.1 Switching Energy . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2.2 Transmission Energy . . . . . . . . . . . . . . . . . . . . . . 43
4.2.3 Receiving Energy . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2.4 Time-OutEnergy........................ 44
4.3 SimulationResults ........................... 44
5 Conclusion 47
5.1 Conclusion................................ 47
References 48
xi
LIST OF FIGURES
2.1 MICAzSensornode........................... 7
2.2 Architecture of Sensor node . . . . . . . . . . . . . . . . . . . . . . 7
2.3 WBAN with on-body sensor nodes . . . . . . . . . . . . . . . . . . 9
3.1 Communication Architecture of WBANs . . . . . . . . . . . . . . . 15
3.2 Algorithm of CSMA/CA . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 TDMAFrame.............................. 21
3.4 Normalized Throughput Versus NC . . . . . . . . . . . . . . . . . . 24
3.5 TDMA Frame Structure . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Superframe Structure of Priority-Guaranteed MAC . . . . . . . . . 25
3.7 TDMA Frame Structure . . . . . . . . . . . . . . . . . . . . . . . . 26
3.8 SuperfameStructure .......................... 27
3.9 Power Compared to Sleep Time and Number of Retransmits . . . . 28
3.10 BodyMAC Frame Structure . . . . . . . . . . . . . . . . . . . . . . 29
3.11 Multi-Superframe Structure for MedMAC Protocol . . . . . . . . . 30
4.1 WBANTopology ............................ 37
4.2 Channel Selection Procedure . . . . . . . . . . . . . . . . . . . . . . 38
4.3 Time Slots Assignment with Guard-band Time . . . . . . . . . . . . 39
4.4 MAC Layer Frame Formate . . . . . . . . . . . . . . . . . . . . . . 41
4.5 Energy consumption of AR-MAC and IEEE 802.15.4 for N= 1000 45
xii
LIST OF TABLES
1.1 Comparison of CSMA/CA and TDMA . . . . . . . . . . . . . . . . 3
2.1 IEEE 802.15.4 - Frequency bands and data rates . . . . . . . . . . 11
2.2 Comparison of Wireless Networking Standards . . . . . . . . . . . 12
3.1 Comparison of MAC protocols based on channel access mechanism . 22
3.2 Comparison of MAC Protocols . . . . . . . . . . . . . . . . . . . . . 32
3.3 Comparison of MAC Protocols . . . . . . . . . . . . . . . . . . . . . 33
4.1 Simulation Parameters Value . . . . . . . . . . . . . . . . . . . . . 44
xiii
Chapter 1
Introduction
1
1
Introduction
1.1 Wireless Body Area Networks (WBANs)
Pervasive and mobile healthcare are emerging technologies for long time pa-
tient monitoring using biomedical sensors. These biomedical sensors are small
in size and battery-operated devices with limited computational and communica-
tion capabilities. WBANs have enabled deployment of wearable and implantable
biomedical sensorsxto provide ubiquitous health monitoring services. Biomedical
sensorsxarexused toxmonitor the physiological parameters ofxhuman bodyxwith
throughput ranging fromxseveral bitsxper hourxup to 10xMbps. WBANs have
some similar demands and challenges like other wireless networks. Sensor nodes
in WBANs have small batteries because of size limitation. In most cases these
small batteries cannot be recharged or replaced. WBANs require energy efficient
mechanism for long time patient monitoring. Thus, energy efficiencyxis onexof
theximportant factors of the MAC design. Similarly, other requirements are min-
imum latency and fair bandwidth management.
For fair access of the shared medium, MAC protocols forxWireless SensorxNetworks
(WSNs) and other short range wireless technologies use TimexDivision Multi-
ple Accessx(TDMA) or CarrierxSense MultiplexAccess with Collision Avoidance
(CSMA/CA). Due to complex hardware and high computational power require-
ments, FrequencyxDivision Multiple Access (FDMA) and CodexDivision Multi-
plexAccess (CDMA) are not suitable approaches for medium access in sensor net-
works. CSMA/CA approach out performs in dynamic types networks. It is pre-
sumed that WBANs are not dynamic. TDMA approach is well suited for WBANs.
However, TDMA-based MAC protocols require extra energy consumption for syn-
chronization. Comparison of CSMA/CA and TDMA is shown in Table 1-1.
Packet collision, idle listening, overhearing, protocol overhead state switching,
etc are the major causes of energy dissipation in WSNs. Corrupted packets af-
ter collision are discarded and followed by retransmission. Re-transmission of
packets leads to extra energy consumption. Packet collision also increases latency.
Transceiver remains operational and continuously monitors medium for data pack-
2
ets during idle listening. In many sensor networks, nodes remain in idlexmode
forxmost ofxthe time, i.e., listening to receive data or control packets that are
not sent. In overhearingxnodes receivexpackets thatxare destinedxto other nodes.
Protocols with high control packet overhead, lead to complexity and high energy
consumption. However, frequent switching of transceiver to avoid idle listening
and overhearing is also energy consuming. Energy efficiency can be improved by
avoiding such wastage causes in efficient way. TDMA is the best approach to
avoid these major sources of energy dissipation. AR-MAC is based upon TDMA
approach to overcome these sources of energy waste. Adaptive assignment of time
slots and guard band improve efficiency of WBANs in terms of energy consumption
and bandwidth utilization.
Table 1.1: Comparison of CSMA/CA and TDMA
Feature CSMA/CA TDMA
Power Consumption High Low
Bandwidth utilization Low Maximum
Traffic level support Low High
Mobility(Dynamic) Good Poor
Synchronization N/A Necessary
1.2 Motivation
Extensive energy is consumed by transceiver communication operation. Exist-
ing research on MAC layer focuses to maximize battery-powered sensor node’s life.
Bottleneck ofxMAC layer protocolxdesign for WBAN isxto achieve highxreliabilityxand
energy minimization. Majority of MAC protocols designed for WBANs are based
upon TDMA approach. However, a newxprotocol needs toxbe definedxto achieve
high energy efficiency, fairness and avoid extra energy consumption due to syn-
chronization.
1.3 Scope
The scope ofxthis research isxto define and developxa newxMAC layer protocol
for WBANs that is more energy efficient and can be implemented on the current
nodes.
3
1.4 Research methodology
An extensive study of the existing protocols and factors was done that influence
the performance of those protocols. A new protocol was designed taking into
consideration the common flaws and problems faced by those protocols. We use
MATLAB for performance evaluation of proposed protocol.
1.5 Research Contribution
We summarize our contribution in this thesis as follows:
•We describe the communication architecture of WBANs with major require-
ment and dominant sources of energy dissipation.
•Then we analyze the existing MAC protocols for WBANs with emphasis on
energy minimization at MAC layer.
•We discuss open research issues with direction for future research.
•Then propose Adaptive-Reliable Medium Access Control (AR-MAC) proto-
col
•Finally we evaluate the performance of AR-MAC with respect to IEEE
802.15.4 in termsxof energy consumptionxandxreliability
1.6 Thesis organization
We provide a detail discussion about the WBANs architecture with emphasis
on requirements, architecture and energy dissipation sources in chapter 2. Chapter
3 presents the proposed MAC protocol design. Chapter 4 describes the analytical
analysis of energy consumption with simulation results. We conclude our research
work in chapter 4 with performance analysis of proposed protocol with IEEE
802.15.4.
4
Chapter 2
Introduction to WBANs
5
2
Introduction to WBANs
2.1 Wireless Sensor Networks
Wireless Sensor Network was first introduced during the Cold War by US [4].
The acoustic sensors were placed in the bottom of ocean to detect the movement of
Soviet submarines. This system of acoustic sensors was called Sound Surveillance
System (SOSUS) [4]. During the same time US also used sensors for radars system.
Both of these networks were wired nature and having no constraint of bandwidth
or energy.
The modern sensor research was first started by Defense Advanced Research
Project Agency (DARPA) in early 1980’s. DARPA introduced Distributed Sensor
Networks (DSN) program, where a network is composed of many independent and
low cost nodes that are able to collaborate with each other. In the mid of 1980’s
the Massachusetts Institute of Technology (MIT) developed a DSN consisting of
acoustic sensors designed to detect and track low-flying aircrafts [4]. Figure 2.1
shows a MICAz node.
The four basic components of each and every node are power source, process-
ing unit, sensing unit and transceiver. Some sensor nodes also contain optional
components like location finding system (GPS), Mobilizer and power generator.
Figure 2.2 shows the basic components of a sensor node.
An optional power generator can be used to support the power unit; solar cells
can also be used for this purpose. The processing unit consists of a processor
and memory. This unit is responsible for managing the tasks of sensor unit. The
sensing unit is generally consistsxof axSensor and Analogue toxDigital Convertor
(ADC).xThe ADCxconverts the analogue data to digital data so that node can
process it before transmitting. The Transceiver connects the node to the network
either through Radio Frequency (RF) or optical communication such as infrared.
The optional location finding system may have a low power Global Positioning
System (GPS). Mobilizer is used to enable the node movement, if mobility is
required for a node to perform its task. All of these components must be fitted in
6
Figure 2.1: MICAz Sensor node
a smaller module like matchbox.
Figure 2.2: Architecture of Sensor node
2.1.1 Network Topology
In WSN network topology changes and maintenance can be viewed in three
phases i.e. deployment phase, post-deployment and re-deployment [9]. The initial
stage is the deployment phase, in which nodes are deployed in a certain territory
either by placing these nodes one by one or dropping from airplane. Topology
changes may occur due to nodes failure and mobility, the post deployment phase
7
are used here to manage such a situation. Sometimes additional nodes are deployed
in the network; this process is known as re-deployment.
2.1.2 Transmission Media
WSNs use wireless medium for RF communication. Most of the WSNs use the
IndustrialxScientific and Medicalx(ISM) frequency band for communication. ISM
frequency band is globally available, unlicensed and is centered around 2.4 GHz.
2.2 Wireless Body Area Networks
Number of small and smart devices increasing due to advancement in wireless
and storage technologies. These small devices are capable of long time health mon-
itoring with in hospital or outside. Wireless Body Area Networks (WBANs) enable
us to use portable, small and lightweight sensor nodes for long time health moni-
toring. Using sensing capabilities, these small energy constrained devices measure
human body parameters and communicate with some external monitoring sta-
tion for diagnose or prescription from a physician. Data streaming from human
body to monitoring station using wireless communication channel is an energy
consuming process. Low power signal processing and energy efficient communi-
cation mechanisms prolong lifespan of these small devices. For Low-Rate Wire-
lessxPersonal AreaxNetworks (LR-WPANs), IEEE 802.15.4 definesxspecification
for PhysicalxLayer and Data LinkxLayer.
In WBANs, sensor nodes of small size with low power and limited compu-
tational capabilities are attached or implanted to human body for measurement
of physiological signs. These physiological signs include; respiratory patterns,
heartbeat, temperature, posture, breathing rate, ElectroCardioGram (ECG), Elec-
troEncephaloGraphy (EEG) and many more. Transmission data rates for these
physiological parameters vary from 1Kbps to 1Mbps. Sensor nodes collect informa-
tion from human body and communicate with a central device called Coordinator.
2.3 Wireless Standards
In March 1999, the IEEE established the 802.15 working group as part ofxthe
IEEE ComputerxSociety’s 802 Localxand MetropolitanxArea NetworkxCommittee.
The 802.15 working group was established with the specific purpose of developing
8
Figure 2.3: WBAN with on-body sensor nodes
short range wireless networks, also known as Wireless Personal Area Networks
(WPANs).
Task Group 1 (802.15.1) defines a standard for WPANs based on Bluetooth
specification for physical and MAC layer. The goal of Task Group 2 (802.15.2) is
to develop axmodel for thexcoexistenceof WLAN (802.11) and WPAN (802.15).
The Task Group 3 (802.15.3) is responsible to develop standards for high data
rate WPANs (20 Mbps or greater). The goal of Task Group 4 (802.15.4) is to
define a low data rate and less complex PHY and MAC layer standards that
will save energy and will achieve a battery life time of months to years. Task
Group 5 working in mesh networks with emphasis on interoperability, stability
and scalability. IEEE 802.15.6, TaskxGroup 6 works to define standards for BAN
xtechnologies.
Inxthe following section we will discuss the 802.15.1 and 802.15.4xstandards.
The 802.15.4 is especially important as it is aimed for sensors and other devices
needing long battery life.
9
2.3.1 IEEE 802.15.1/Bluetooth
Bluetooth was designed to replace the short range cable technology and pro-
vides communication between computer and its peripherals. Bluetooth isxa shortxrange
(10m), low power (1 to100 mW) and low cost device whose transceiver operate
in 2.4 GHz of ISMxband. Bluetooth uses FrequencyxHopping SpreadxSpectrum
(FHSS) with the hop rate of 1600 hops/s [12].
Bluetooth forms piconet, a small network of Bluetooth devices. A piconet can
have two to eight nodes. One of the nodesxact as a master whilexthe remaining
nodes are connected to it as slave. The limit of seven slaves is because of three
bit addressing scheme in piconet. Three bit allow eight different addresses in
which zero address is reserved for broadcasting, so a piconet can maximally have
seven slaves. The master clock is used for synchronization and all communication
within piconet is routed via master. When a node participates in more than one
network, scatternet is formed. A scatternet is a network ofxtwo orxmore piconets.
A node participating in more than one piconet is called gateway node and uses
Time Division Duplex (TDD) in order to be active in one piconet at a time.
Based on Bluetooth specificationxthe IEEE 802.15.1 definesxMAC and PHYxlayers
standard for WPANs. The radio layer of the Bluetooth protocol stack forms the
PHY layer of 802.15.1 while the Logical Link Control and Adaptation Protocol
(L2CAP), Link Management Protocol (LMP), and Baseband layers of Bluetooth
protocol stack form the 802.15.1 MAC layer. The PHY layer specifies the com-
munication band (2.4 GHz) while the MAC layer is responsible for the time syn-
chronization of the FHSS communication.
2.3.2 IEEE 802.15.4 and ZigBee
An Alliancexwas formed by an association of several companies in 2002, called
the ZigBee Alliance. The main goal of the Alliance was to develop monitoring
and controlling devices that are reliable, low power, low cost and are wirelessly
networked using an open global standard. The IEEE 802.15 task group four has
already started working on a standard for low data rate WPANs. The IEEE and
ZigBee Alliance joined and decided that ZigBee would be the commercial name of
the technology.
Some ofxthe applications ofxthe 802.15.4 standard include sensors, remote con-
trols, home automation, smart badges and interactive toys [13]. The standard use
three license free frequency bands with two Directed Sequence Spread Spectrum
10
(DSSS) PHY layers. One PHY layer operate at 868/915 MHz and uses the 868-870
MHz band with one channel and 902-928 MHz band with ten channels. This PHY
layer achieves a dataxrate of 20 kbps in 868-870 MHz frequency band and 40 kbps
in the 902-928 MHz band. The second PHY layer operates atx2.4 GHz and uses
the 2.4-2.4835 GHz band with sixteen channels and achieves a data rate of 250
kbps. The Table 2.2 summarizes the frequency band and data rates of 802.15.4
standardx.
Table 2.1: IEEE 802.15.4 - Frequency bands and data rates
PHY Band Channels Chip
Rate
Modulation Bit
Rate
868/915 MHz 868-
870MHz
0 300
Kchip/s
BPSK 20 kbps
902-928
MHz
1 to 10 600
Kchip/s
BPSK 40 kbps
2.4 GHz 2.4-
2.4835GHz
11 to 26 2
Mchip/s
O-QPSK 250 kbps
The IEEE 802.15.4 supportsxtwo addressing mechanism namelyx16bit short
and 64bit IEEE addressing. The PHY layer also has features for link quality
indication, receiver energy detection and clear channel assessment. MAC layer
support both contention free and contention based access with a maximum packet
size of 128 bytes, containing a payload of 104 bytes maximally. The MAC layer
uses full handshaking for reliability and uses CSMA/CA for carrier sensing.
ZigBee defines three software layers [10] (network, security and application)
on top of PHY and MAC 802.15.4 layers. Network layer supports three net-
workxtopologies namely star, mesh orxpeer-to-peer, andxcluster based topologies,
as shown in Figure 2.4. The 802.15.4 definesxtwo types of nodes i.e. Fully Func-
tionalxDevice (FFD) andxReduced Functional Device (RFD). An FFD canxroute
data while an RFD cannot, this standard also specify that a network must have
at least one FFD.
A start topology saves energy and increase network lifetime since every RFD is
directly connected with the coordinator. A mesh or p2p topology brings reliability
and scalability since all nodes are FFDs and directly interconnected so it introduces
multiple routing paths. The cluster tree topology combines both the start and
mesh topologies and trying to extend network lifetime with a reasonable reliability
and scalability [10].
11
2.4 Comparison of wireless networking standard
There are different standards defined for wireless networks. These standards
divide wireless networks into different categories based on factors like network
size, transmission range, data rate and network lifetime. Below table shows a
comparison of the three important wireless network standards.
Table 2.2: Comparison of Wireless Networking Standards
Market Name Wi-Fi Bluetooth ZigBee
Standard IEEE 802.11b IEEE 802.15.1 IEEE 802.15.4
Type of Network WLAN WPAN WPAN
Application Focus Web, email,
Video
Cable Replace-
ment
Monitoring and
control
System Resources 1MB+ 250KB+ 4KB-32KB
Battery Life (days) 0.5 5 1-7 100 1,000+
Network Size 32 7 255/65,000
Data rate (kbps) 11,000+ 720 20-250
Transmission
Range
1-100 (meters) 1 10+ (meters) 1-100+ (meters)
Success Metrics Speed, Flexibil-
ity
Cost, Conve-
nience
Reliability,
Power, Cost
12
Chapter 3
Related Work
13
3
Related Work
3.1 WBANs
Number of small/smart devices increasing due to advancement in wireless and
storage technologies. These small devices are capable of long time health monitor-
ing with in hospital or outside. Wireless Body Area Networks (WBANs) enable
us to use portable, small and lightweight sensor nodes for long time health moni-
toring. Using sensing capabilities, these small energy constrained devices measure
human body parameters and communicate with some external monitoring sta-
tion for diagnose or prescription from a physician. Data streaming from human
body to monitoring station using wireless communication channel is an energy
consuming process. Low power signal processing and energy efficient communi-
cation mechanisms prolong lifespan of these small devices. ForxLow-Rate Wire-
lessxPersonal Area Networks (LR-WPANs), IEEEx802.15.4xdefines specification
for PhysicalxLayer and DataxLink Layer [1].
In WBANs, sensor nodes of small size with low power and limited compu-
tational capabilities are attached or implanted to human body for measurement
of physiological signs. These physiological signs include; respiratory patterns,
heartbeat, temperature, posture, breathing rate, ElectroCardioGram (ECG), Elec-
troEncephaloGraphy (EEG) and many more. Transmission data rates for these
physiological parameters vary from 1Kbps to 1Mbps. Sensor nodes collect informa-
tion from human body and communicate with a central device called Coordinator.
Energy efficiency is the mostximportant requirement of a goodxMAC proto-
colxfor WBANs. To improve energy efficiency of WBANs, a versatile MAC pro-
tocol should have the capabilities to reduce power dissipation due to collision of
packets, overhearing of nodes, idle listening to receive probable data packets and
control packet overhead of communication. Similarly Qualityxof Servicex(QoS) is
an important goal to achieve in WBANs. This includes latency, jitter, guaranteed
communication and security.
14
3.2 WBAN Architecture
In this section, we describe communication architecture of WBANs. Then
we describe the design requirements for WBANs. Finally, we briefly present the
energy dissipation sources in WBANs.
3.2.1 Overview
Figure 2.1 shows 3-level architecture of WBANs for non-medical and medical
applications. Lowest layer consists of small sensor nodes attached or implanted to
a human body for long time monitoring of physiological or biomedical signs or hu-
man body postures. Two types of nodes are used in WBANs: (1)Biosensors: used
to measure ElectroEencephalGgram (EEG), ElectroCardioGram (ECG), Heart-
beat, continues blood sugar, Human body temperature, Blood Pressure (BP); Bio-
kinetic Sensors: used to measure acceleration and human body mobility. These
on, in or around the body sensor nodes are organized in the most common star
topology for communication of sensed information to a central device. The cen-
tral device communicates the received information for diagnose and prescription
from health services provider. For communication, central node uses the existing
technology of Level 1 and Level 2 as shown in Figure 2-1. Communication pattern
in Level 1 is termed as IntraBAN. However, Level 2 and Level 3 communication
are termed as ExtraBAN communication.
WiFi
GPRS
Central Node
ECG
ECG
PH
Hearbeat
Motion
GSM/GPRS BTS
Aceess Point
Bluetooth
/ WiFi
Figure 3.1: Communication Architecture of WBANs
The number and nature of sensor nodes vary according to the application re-
quirements. In deployment of these nodes the human body structure and mobility
are kept in consideration for reliable communication. Human body tissues are
15
sensitive to the electromagnetic radiation of the transceiver. To avoid the harmful
effects, the transceiver power is adjusted to minimum level as possible. Sensor
nodes placed on head and torso do not observe mobility as compared to nodes
placed on head and legs. However, attached sensor nodes to legs and arms scru-
tinize high mobility. For data communication in Level 1, WBANs usexIndustrial,
Scientificxand Medical (ISM) frequency band, Ultra-WideBandx(UWB) and Wire-
less MedicalxTelemetry Services (WMTS) frequency band. MICS (402-405 MHz)
and WMTS (14 MHz) are licensed frequency bands. However, ISM (2.4 MHz) is
an unlicensed frequency band. WMTS is highly secure and only authorized and
trained physicians/technicians can use this spectrum. However, WMTS cannot
support audio and video streaming. MICS is especially dedicated to implant com-
munication. The most common frequency band used in WBANs is ISM, WiFi,
Bluetooth and ZigBee [Sanaullah] also use this specific frequency band for wireless
communication.
The medical applications and consumer electronics applications depend on pro-
tocols design at Level 1. The small battery operated sensor nodes collect the crit-
ical and non-critical information from environment or human body. Traffic from
these nodes classified into; Normal traffic, Emergency traffic and On-demand traf-
fic. Normal traffic is generated periodically under normal conditions. Coordinator
or central node collects the normal traffic periodically. The on-body or implanted
sensor nodes initiate the emergency traffic whenever the measured value exceeds a
predefined threshold value. Emergency traffic is unpredictable and not generated
on regular basis. However, central node or coordinator originates on-demand traf-
fic to acquire some information needed by the physician or monitoring station for
treatment or network management. Overall performances of WBANs, especially
energy efficiency, reliability, robustness, wear ability and scalability is related to
Level 1. For energy efficiency and reliability of communication, design of MAC
layer protocols at Level 1 play a vital role. With a good MAC design at Level
1, high throughput, high energy efficiency and minimum delay can be archived.
A number of MAC protocols for WBANs are proposed so far, we discuss these
protocols with their pros and cons in Section IV.
3.2.2 Design Requirements for WBANs
In WBANs, sensor nodes collect the critical and non-critical information from
different parts of the patient body and communicate with coordinator. Latency
and transmission reliability are important requirements for effective patient health
monitoring systems. Similarly for long time monitoring, WBANs required high
16
energy efficiency and scalability at Level 1.
3.2.2.1 Energy efficiency
Energy efficiency is the first goal to achieve in WBANs since sensor nodes are
small and battery operated. For long time patient monitoring, it is an obligatory
goal to play down energy dissipation at Level 1 as much as possible. Multiple
and dynamic power management schemes can be used to prolong lifespan of sen-
sor nodes. In WBANs, sensor node’s transceiver is one of the dominant energy
dissipation sources. Optimization of PHYsical (PHY) and MAC layer processes
result in reduced power consumption of transceiver. PHY layer has some limita-
tion for power optimization. However, MAC layer provides higher level of energy
savings by introducing multiple transmission scheduling schemes, optimal packet
structure, smart signaling techniques and enhanced channel access techniques.
3.2.2.2 Reliability
Reliability of WBANs depends upon transmission delay of packets and packet
loss probability. Packet transmission procedures at MAC layer and Bit Error Rate
(BER) of channel influence packet loss probability. Appropriate channel access
techniques, packet re-transmission schemes, packet size, and enhanced scheduling
schemes at MAC layer improve reliability.
3.2.2.3 Scalability
Scalability is the essential requirement for WBANs. Number of nodes, to col-
lect life critical and non-critical information, varies according to patient monitor-
ing requirements. Easily configuration of WBANs by adding or removing sensor
nodes is required to support the scalability. MAC layer has the potential to achieve
scalability.
3.2.2.4 Quality of Service (QoS)
MAC layer play a vital role to achieve high QoS. Medium access techniques at
MAC layer like TDMA and polling put forward deterministic packet loss, packet
delay. However, contention based protocols like CSMA allocates the transmission
channel to node only when it is free and the node has data to transmit. Random
access techniques result in variable packet loss and delay. Adaptive sleep cycles in
17
contention based protocols enhance energy efficiency at the cost of increase latency
and packet drops.
3.3 Sources of Energy Dissipation in WBANs
Sensor nodes have small batteries with limited power capabilities. Replace-
ment or recharging of batteries by energy scavenging is not possible. Due to
limited energy resources the power consumption of sensor nodes needs to be con-
trolled tightly. Thus minimization of energy consumption is a major issue in
WBANs. Power consumption of sensor nodes can be decreased with low power
MAC protocols. Collision of packets, overhearing of nodes, idle listening to re-
ceive the possible data packets , communication control packet overhead, packet
forwarding and transceiver state switching are the foremost sources of energy dis-
sipation in Wireless Sensor Networks (WSNs). In [12] authors identify the first
four sources of energy dissipation.
Transmission ofxdata packets on single channel byxtwo or morexsensor nodes
simultaneously results in packet collision. Collision of the packets occurs at the
receiver end. These packets are dropped and sender nodes retransmit these pack-
ets. Re-transmission of the dropped packets results in extra energy dissipation. In
overhearing, sensor nodes receive the packetsxthat are destinedxfor otherxnodes.
Those received packets are dropped and energy is dissipated. In idle listening,
nodes listen to idle channel toxreceive the possiblexpackets transmitted by other
nodes which results in extra energy consumption. If the control packets used in
communication are maximumxeffective throughput decreases.xTransmission and
receptionxof these maximum control packets consume more energy. Energy is con-
sumed in packet forwarding, when router nodes consume energy to forward a data
packet from source to destination. However, energy consumption due to packet
forwarding is ignored in WBANs due to single-hop communication in star topol-
ogy. The last source is state switching, which occurs when a sensor node switch
its transceiver from sleep mode to active mode for data transmission and then
back to sl eep mode to avoid idle listening and overhearing. Frequent switching of
transceiver is also energy consuming. Energy efficiency can be improved by avoid-
ing such energy wastage sources in efficient way. In this chapter, we briefly discuss
the classification of MACxprotocols based upon medium access techniques. We
also discuss the existing proposed protocols in details with their pros and cons. A
comprehensive table of comparison of these protocols is given at the end of this
chapter.
18
3.4 Classification of MAC protocols for WBANs
For fair access of the shared medium, MACxprotocols developed forxWBANs
are classified intoxthree categories based on channel access mechanism; xContention-
Based, Contention-Free and Low Power Listening (LPL) or Polling. This section
provides brief description about these channel access mechanisms.
3.4.1 Contention-Based MAC Protocols
In contention-basedxchannel access mechanism, sensor nodes contend for shared
medium to communicate with other nodes or coordinator. There is no predefined
scheduled for the end nodes to communicate in contention-based mechanism re-
sulting in variable latency and packet loss. CSMA is a contention based mech-
anism to access the available shared medium for data transmission. However,
CSMA/CA is a modification of CSMA to avoid packet collision. In CSMA/CA
with no RTS/CTS exchange, before transmission of data packets, nodes listen to
shared medium/channel to find out whether the shared channel is idle or not.
In case of idle situation, node starts transmission of data packets. However, if
channelxis sensed busy, transmission is rescheduled for axrandom period of time.
Figure 2.2 shows the CSMA/CA simplified algorithm.
In some cases, we need a scheduled based contention channel access mecha-
nisms called scheduled-contention. A common schedule is used for data commu-
nication to ensure the reliability and collision avoidance. Scheduled-Contention
mechanisms required periodic synchronization. To maintain synchronization, sched-
ules are exchanged on regular basis leading to extra energy consumption. The
synchronization of nodes is highly sensitive to clock drift. Periodic sleep of nodes
in this mechanism reduces the idle listening and overhearing to improve the power
efficiency.
Contention-based mechanisms are well suited in dynamic and sealable net-
works. However, in WBANs such mechanisms do not provide reliable and efficient
communication due to high energy consumption for Clear Channel Assessment
(CCA) and poor handling capabilities for emergency and on-demand traffic.
19
START
Generate Data
Packet
Is the
Channel Idle?
Yes
No Wait for a
Random Backoff
Time
Transmit RTS
CTS Received? No
Transmit Data
Packet
END
-- Using RTS/CTS Exchange --
Not using RTS/CTS Exchange
Figure 3.2: Algorithm of CSMA/CA
3.4.2 Contention-Free MAC Protocols
InxContention-Free MAC protocols, sensor nodes are assigned Guaranteed
Time Slots (GTS) for data communication. These protocols provide determin-
istic delay with no packet loss due to communication in guaranteed time slots
with out contention period. TDMA is a Contention-Free channel access mecha-
nism where channelxis divided into multiple time slots of fixed orxvariable length.
These time slotsxare allocated to end nodes for communication. However, multiple
time slots can also be assigned to a single node depending upon the requirements
and data volume. Pre-defined and dedicated time slots in TDMA provide a col-
lision free environment for data communication. Synchronization is the key issue
in TDMA based MAC protocols. However, TDMA base MAC protocols are effi-
cient than CSMA/CA based protocols in terms of energy efficiency and bandwidth
utilization.
For limited number of sensor nodes in WBANs with fixed data rate, TDMA
is suitable. Sensor nodes only wakeup in specified time slots for communication
otherwise, they remain in sleep mode to avoid idle listening and overhearing.
Assigning time slots to sensor nodes with different data rates, non-periodic data
and scalability are the key issues in implementing TDMA in WBANs
20
TDMA Frame With N User Time Slots
Figure 3.3: TDMA Frame
PG-MAC protocol in [4] is based upon TDMA approach where dedicated time
slots or assigned to sensor nodes. Performance evaluation shows that proposed
protocol out performs than IEEE 802.15.4 with respect to power consumption.
3.4.3 Low Power Listening (LPL) MAC Protocols
In LPL mechanism, sensor nodes periodically listen to the channel. Nodes
go into sleep mode if channel is sensed idle, other wise keep the transceiver in
active mode to receive data packets. This mechanism is also known as Polling.
A long preamble is sent before the message to detect the pooling at receiver side.
LPL mechanisms avoid idle listening and overhearing. Synchronization is not
required here. Based on hardware complexity and listening of long preamble, LPL
mechanisms are not well suited in WBANs. LPL mechanisms support simplex
communication. However, WBANs required duplex channel communication to
accommodate periodic, on-demand and emergency traffic.
21
Table 3.1: Comparison of MAC protocols based on channel access mechanism
MAC Features Contention-Based MAC
protocols
Contention-Free MAC
Protocols
Low Power Listening
MAC Protocols
Network Scala-
bility
Highly Scalable Poor Scalability Good but limited by delay
Packet delay Variable, depends on traffic
Load, priority and applica-
tion
Deterministic Deterministic but varies
with Traffic load
Packet loss Variable, depends on traffic
Load, priority and applica-
tion
Deterministic Deterministic but varies
with Traffic load
Energy Effi-
ciency
CCA in high traffic ends up
with high energy consump-
tion
Guaranteed time Enable
collisions free communica-
tion and periodic sleep
which improves energy effi-
ciency
High Energy efficiency
Traffic Handling Handle periodic, non-
periodic and on demand
traffic
Handle periodic traffic Handle periodic traffic with
capabilities for non-periodic
Throughput Low Excellent Good
Synchronization Synchronous Synchronization is required NA
Sensitivity to
clock drift
Sensitive to clock drift Highly Sensitive to clock
drift
NA
22
3.5 MAC protocols for WBANs
Inxthis section wexdiscuss some of well known existing MAC protocols pro-
posed for WBANs. This discussion covers the pros and cons of these proposed
protocols in context of energy minimization and resource utilization. The fol-
lowing subsections provide detail operation of these protocols with emphasize on
energy consumption. We also discuss, how these protocols tackle energy ineffi-
ciency sources like collision, idlexlistening,xoverhearing and controlxpacket over-
headxwhich are widely addressed in literature.
3.5.1 IEEE 802.15.4 MAC protocol
IEEE 802.15.4 is designed for low data rate wireless applications [1]. This
protocol operates in threexfrequency bands: 868 MHz,x915 MHz andx2.4 GHz
frequency bands. These frequency bands are further divided into 27 sub-channels
i.e., 2.4 GHz frequencyxband is divided into 16 sub-channels, 915 MHs frequency
band into 9 sub-channels and onexsub channel in 868 MHz frequency band. Two
operational modesxare defined for IEEE 802.15.4: beaconxenabled mode and non-
beaconxenabled mode.
In beaconxenabled mode, coordinator controls device synchronization, associ-
ation and data transmission using periodic beacons. Beacon enabled mode use
a super frame. Thisxsuper frame consists of active and inactive periodsx. Ac-
tive period of super frame isxdivided into three parts: Contention Access Period
(CAP) using slotted CSMA/CA, beacon and a Contention Free Period (CFP). A
maximum ofxseven Guaranteed Time Slots (GTS) are assigned to end nodes to
accommodate time critical data in CFP. This mode of operation ofxIEEE 802.15.4
is not suitable for WBANs due to its asymmetric transmission support.
Non-beaconxenabled mode ofxIEEE 802.15.4 uses un-slotted CSMA/CA. Au-
thors in [2], analyze slotted and un-slotted CSMA/CA and presented their re-
sults. These results show that un-slotted mechanism out performs well in terms
of bandwidth utilization and latency as shown is Figure 2.4. However, inxnon-
beacon enabled mode the Clear ChannelxAssessment (CCA) leads to high energy
consumption.
23
Figure 3.4: Normalized Throughput Versus NC
3.5.2 Battery-aware TDMA protocol
In [3], authors propose a battery-aware TDMA based MAC protocol with
cross-layer design to maximize the network life. This protocol takes the following
parameters into account for medium access: electrochemicalxproperties of battery,
time-varying wirelessxfading channel, and packetxqueuing characteristics. The
operation of this protocolxis similar toxIEEE 802.15.4 beacon enabled mode, where
the modes listen periodically to beacons from coordinator. The time axis is divided
into three parts; beaconxslot, active time slots andxinactive period as shown in
the Figure 2.5 [3]. The frame length is adaptive and can be changed according
to application requirements. Sensor nodes wake up at the beginning of beacon
period. Each node has its own distinct time slot Tsto transmit data in active
period after receiving the beacon. To avoid extra energy consumption, nodes
remain in sleep mode for the inactive time. This protocol prolongs the lifespanxof
wireless sensorxnodes.xReliable and timelyxdelivery of packets is achieved using
GTS. However, there is no mechanism defined for emergency data. Similarly
holding of packets in buffer for long time, leads to high average delay and packet
drop rate.
3.5.3 Priority guaranteed MAC protocol
In [4], authors propose a priority-guaranteedxMAC protocol. This protocol
uses a new superframe structure as shown in Figure 2.6. The activexperiod is di-
videdxinto five parts; a beacon, Control Channel AC1, Control Channel AC2,
Time SlotxReserved for Periodic (TSRP) traffic, and Time Slot Reservedxfor
24
Figure 3.5: TDMA Frame Structure
Bursty (TSRB) traffic. AC1 is used for uplinkxcontrol of life-critical medical
application while AC2 is used for uplinkxcontrol of Consumer Electronics (CE)
applications. Randomized ALOHA is used for these two control channels. Pro-
posed protocol is based upon TDMA approach to assign Guaranteed Time Slots
(GTS) within twoxdata channels TSRP and TSRB. These time slots are allocated
on-demand to using the control channels. As shown in simulation results [4],
this protocol outxperforms than IEEE 802.15.4 inxterms of energyxconsumption.
However, complexxsuperframe structure and inadaptability to emergencyxtraffic
are major drawbacks of this protocol.
Figure 3.6: Superframe Structure of Priority-Guaranteed MAC
3.5.4 Energy-Efficient Low Duty Cycle MAC Protocol
Authorsxpropose a new MAC protocol based upon the static nature of BAN
[5]. TDMA approach is used for streamingxlarge amount ofxdata. The Static
nature and TDMA approach are being utilized efficiently to maximize the network
life. In target topology a Master Node (MN) collects data from on body nodes
and communicates with a Monitoring Station (MS). The received data is being
analyzed by MS while the on-body network coordination and synchronization is
being performed by MN. As shown in Figure 2.7, the total frame is dividedxinto
multiplextime slots. Timexslots S1 to Sn are allocated to sensor nodes while
time slots RS1 to RS2 are reserved which are being assigned when requested.
25
The number of these extra time slots depends upon the targeted packet drop,
packetxerror rate and numberxof sensor nodes.
Figure 3.7: TDMA Frame Structure
To avoid the collision/overlapping of packets transmission due to clock drifts,
guard band time is inserted between two consecutive time slots. Two types of
communication models are being discussed in the paper. First, the MN has one
transceiver. In this case, enough time is reserved for communication of MN with
MS. In second case, where the MN has two transceivers, simultaneously commu-
nication of MN with MS and sensor nodes is possible. The communication uses
the different physical layer communication models for transparency. From energy
consumption analysis in [5], proposed protocol out performs in term of energy for
high communication data rates as well as for short burst of data. However, this
protocol uses a Network Control (NC) packet for periodic synchronization after
Nnumber of time frames which leads to an extra consumption of energy. Other
shortcoming includes; fixed frame structure based on pure TDMA, no CAP to
accommodate small burst of data, and no mechanism is defined for on-demand
traffic.
3.5.5 A power-efficient MAC protocol for WBAN
In [6], authors propose a new mechanism at MAC layer to accommodate nor-
mal, emergency,xand on-demand traffic.xFor reliable transmission twoxwakeup
mechanisms are defined: traffic-based wakeupxmechanism for transmission of nor-
mal traffic and wakeupxradio mechanism for emergency/on-demandxdata trans-
mission. Normal traffic is generated periodically by sensor nodes to monitor rou-
tine physiological parameters. The unpredictable emergencyxtraffic is initiated by
26
on-body sensor nodes in emergency situation. However, the coordinator generates
on-demand traffic to acquire information from sensor nodes. A new superframe
structure is defined where the time axis is divided intoxthree parts: axbeacon mes-
sage, a Configurable ContentionxAccess Period (CCAP) to accommodate short
burst of data, and axContention Free Period (CFP) where Guaranteed Time Slots
(GTS) are assigned to end nodes for collision free communication. In CCAP,
proposed protocol uses slotted ALOHA. Superframe structure for this protocol is
shownxin Figure 2.8 [6].
Figure 3.8: Superfame Structure
Coordinator organizes the traffic-based wakeup table according to application.
Periodic sleep/wakeup mode avoids the unnecessary energy dissipation due to
idle listening and overhearing. To compensate the clocks drift at coordinator
and sensor nodes, sensor nodes wake up in advance for a time period of TK=
2θTWwhere TWis the beacon period. For emergency traffic sensor nodes send
wake up radio signal to coordinator while coordinator sends a wake up signal to
sensor nodes for on demand traffic. Simulation results based upon Monte Carlo
method for poisson and deterministic traffic. Performance of proposed protocol
in terms of energy and delay are compared with that of WiseMAC [7], where it
performs better. However, use of Low Power Listening (LPL) is not an optimal
choice for implanted and on-body sensor nodes communication due to strict power
capabilities.
3.5.6 Energy Efficient Medium Access Protocol
In [8], authors propose a new MAC protocol based upon centrally controlled
wakeup and sleep mechanisms to maximize energy efficiency. Some upper layer
functionalities are incorporated to reduce power dissipation due to overhead. This
protocol is based upon basic assumption of sensor nodes with a star topology
where a central node (master node) coordinates with on-body/implanted sensor
27
nodes (Slave nodes). Maximum number of slave nodes for a single master nodes is
defined to be 8. Due to high power and computational capabilities, more activities
and processes are assigned to central node.
Basic operation of this MAC protocol involves three processes. First one is
link establishment, where a slave node wants to join a cluster. After success-
ful link establishment, each nodexis assigned with a unique sleep time to avoid
idle listening and overhearing. Second one is the wakeupxservice process, where
master and slave nodes communicate. Exception process, also called an Alarm
process is initiated by slave node to communicate with master node for emergency
data. For guaranteed and reliable communication, a novelxconcept of Wakeup
FallbackxTime (WFT) is introduced. In case of failure in assigned wakeup pro-
cess, sensor node enters into sleep mode for a specific time interval calculated by
WFT. During this sleep time, senors node buffers data packets for future commu-
nication. Similarly, master node also goes into sleep mode set by WFT if it fails
to communicate with slave nodes. Overlapping of time slots is being avoided by
this mechanism.
Figure 3.9: Power Compared to Sleep Time and Number of Retransmits
From simulation results for different applications such as glucose monitoring,
human body temperature, EEG, and ECG, power consumption depends on sleep
28
interval and number of retransmissions as shown in Figure 2.9. The centrally
controlled process reduces efficiently the extra energy consumptionxdue to idle
listeningxand overhearing. However, implementation of this protocol is highly
complex and has no proper mechanism to handle on-demand traffic. Other draw-
backs include: limitation of nodes in one cluster, communication is only initiated
by mater node and only one node goes into link establishment process at a time.
3.5.7 BodyMAC
In [9], authors propose a TDMA-based MAC protocol where they define uplink
and downlink subframes to facilitate sleep mode with emphasize on energy min-
imization. Nodes remain in sleep modexwhen they havexno data to send.xSleep
mode performs well for low duty cycle sensor nodes. Different data communica-
tion models are accommodated using 3 bandwidth management procedures; Burst
Bandwidth procedure, Periodic Bandwidth procedure and Adjust Bandwidth proce-
dure. This efficient and flexible bandwidth management procedure improves the
network stability and improves transmission of control packets.
Figure 3.10: BodyMAC Frame Structure
As shown in Figure 2.10 [9], the MAC frame isxdivided intoxthree parts; a
beacon, a downlink and uplink. Synchronization is archived by beacon. To ac-
commodate on demand traffic, downlink is used for data communication from
coordinator node to sensor nodes. However, the uplink frame isxdivided into Con-
tention Access Period (CAP) and Contention Free Period (CFP). CAP is based
on CSMA/CA, where nodes compete to send control packets to coordinator for
Guaranteed Time Slots (GTS). However, nodes can also communicate for small
data packets during CAP. Coordinator assigns GTS to sensor nodes in CFP to
avoid collision. Communication using CFP improves energy effecting. However,
for uplink frame in CAP, CSMA/CA ends up with high energy consumption due
to Clear Channel Assessment (CCA) and collision issues.
29
3.5.8 MedMAC
In [10], authors propose Medical Medium Access Control (MedMAC) protocol
for WBANs to improve channel access mechanism and reduce energy dissipation.
MedMAC using TDMA approach for time slots assignments to end nodes for data
communication. However, these assigned time slots are of variable length and
vary according to sensor nodes requirements. A novel scheme is introduced for
synchronization. MedMAC uses multi-superframe structure, where beacons are
used for synchronization as shown in Figure 2.11 [10]. For network initialization,
emergency traffic, and low data communication it uses an optimal contention
period.
Figure 3.11: Multi-Superframe Structure for MedMAC Protocol
To maintain clock synchronization of nodes and coordinator, MedMAC uses
timestamp scavenging with Adaptive Guard Band Algorithm (AGBA). Collision
of data packets is avoided using unique GTS for each sensor node. Similarly AGBA
maintain the synchronization of devices to avoid collision due to clocks drift. Us-
ing AGBA guard band time is inserted between two consecutive time slots. This
guard band time is adjustable and based on clock drift of devices. Drift Adjust-
ment Factor (DAF) monitors the guard band and avoid waste of bandwidth using
extra guard bands.
Authors use OPNET for simulation. They comparexthe performancexof Med-
MAC with thatxof IEEE 802.15.4 with respect to power dissipation. For applica-
tions with low data rates like pulse (8 bps),respiration (640 bps), and temperature
(16 bps), and medium data like ECG, simulation are performed. From the simu-
30
lation results in [25], it is concluded that it out performs than IEEE 802.15.4 with
respect to energyxconsumption. Collision is being avoided using GTS. However,
MedMAC takes low data traffic into consideration which is not suited in WBANs
where date rates for wearable and implanted sensors may be high.
3.5.9 Heartbeat-Driven MAC protocol
In [11], authors propose a TDMA based protocol for WBANs with utilization
of Heartbeat-Rhythm for synchronization. Network topology for proposed pro-
tocol is star topology where a central node coordinates the network. To avoid
collision, H-MAC assigns dedicated time slots to sensor nodes for communica-
tion. Using Heartbeat Rhythm, H-MAC maintains the synchronization required
for TDMA approach without using periodic control messages. This mechanism
leads to minimize overall energy consumption. Each biosensor extracts Heartbeat
Rhythm information from its sensory data. For detection of peaks in the heart-
beat rhythm, authors use the algorithms proposed in [12,13]. Synchronization is
archived by these peaks. H-MAC uses the peek intervals for data communication.
Time slots assignment and frame cycles for synchronization are calculated by co-
ordinator. The coordinator also utilizes Heartbeat Rhythm information from its
own sensory data.
From simulation results, H-MAC prolongs the networks life as compared to
Lightweight MAC (L-MAC) [14] and Sensor MAC (S-MAC) [15]. This efficiency
is achieved by TDMA approach, where collisions are avoided by dedicated time
slots and reduced idle listening. Replacement of traditional synchronization with
Heartbeat Rhythm pattern also reduces the energy consumption. However, Heart-
beat Rhythm is not accessible by all sensors like accelerometer. In such cases
devices can not by synchronized. Integration of accelerometer with other sensors
or facilitating accelerometer to access the heartbeat leads to complexity. Similarly
the insertion of guard band time to avoid collisions leads to minimize bandwidth
utilization.
31
Table 3.2: Comparison of MAC Protocols
MAC Protocol Access Tech-
nique
Energy Efficiency Mech-
anism
Synchronization Performance
Compari-
son
Remarks
IEEE 802.15.4
Beacon Enabled
[1]
Slotted
CSMA/CA
NA Transmit Beacons NA Beacon enabled is not
suitable for long time
monitoring
Battery-aware
TDMA protocol
[3]
TDMA Periodic Sleep mode Using beacons like
IEEE 802.15.4
IEEE
802.15.4
and Blue-
tooth
No Mechanism for
Emergency Traffic.
Has high packet delay
Priority guaran-
teed MAC pro-
tocol [4]
Randomized
ALOHA and
TDMA
Differentiating control
and data channels and
introducing periodic
sleep mode
Beacons are used for
downlink synchroniza-
tion
IEEE
802.15.4
Complex Super
frame Structure and
inadaptability to
emergency Traffic
Energy-Efficient
Low Duty Cycle
MAC Protocol
[5]
TDMA Low duty cycle and long
sleep time to reduce
power consumption
Periodic synchroniza-
tion after N cycles us-
ing Network Control
Packet
FlexiMAC
[16]
No mechanism is de-
fined for on-demand
traffic, Periodic syn-
chronization is energy
consuming
A power-efficient
MAC protocol
for WBAN [6]
Low Power Lis-
tening (LPL) &
slotted ALOHA
Using Guarented time
slots for communiation
to avoid collisions, idle
listening and overhear-
ing are avoided by period
sleep
Early wakeup mech-
anism to compensate
clock drift
B-MAC,
X-MAC,
WiseMAC
and Zig-
Bee
Low Power Listening
(LPL) is not an op-
timal choice for im-
planted and on-body
32
Table 3.3: Comparison of MAC Protocols
MAC Protocol Access Tech-
nique
Energy Efficiency Mech-
anism
Synchronization Performance
Compari-
son
Remarks
IEEE 802.15.4
Non-Beacon
Enabled [1]
Unslotted
CSMA/CA
NA NA NA High throughput ,
Low latency and high
power consumption
Energy Efficient
Medium Access
Protocol [7]
CCA based on
Listen Before
Transmit (LBT)
centrally controlled pro-
cess reduces efficiently
the extra energy con-
sumption due to idle lis-
tening and overhearing
During every commu-
nication master ans
slave node share their
clocks informations
Zigbee,
Bluetooth
and IEEE
802.11
highly complex imple-
mentation and have
no proper mechanism
tohandle on-demand
traffic
BodyMAC [8] TDMA and
CSMA/CA
Idles listeing and over-
hearing are avoided by
periodic sleep
Beacon messages are
used for synchroniza-
tion
IEEE
802.15.4
CCA in uplink leads
to high energy con-
sumption
MedMAC [10] TDMA Dynamic adjustment of
resources for QoS
multi-superframe
structure is used for
synchronization
IEEE
802.15.4
Efficient for low data
rate
Heartbeat-
Driven MAC
protocol [11]
TDMA Using TDMA approach,
dedicated time slots for
sonsor nodes
utilization of
Heartbeat-Rhythm
for synchronization.
Lightweight
MAC (L-
MAC) [14]
and Sen-
sor MAC
(S-MAC)
[15]
Insertion of Guard
band reduces band-
width utilization
33
3.6 Discussion and Open Research Issues
Energy efficiency is one of the main goal to achieve in WBANs for mobile
and ubiquitous health monitoring with critical and non-critical conditions. Cur-
rent research work for energy minimization is focused at MAC layer. However,
other areas such as network layer and cross layer design need to be consider for
energy minimization. In cross layer design we can improve the energy efficiency
by integrating two or more protocol layers from communication protocol stack.
Therefore, research work using cross layer approach will be prominent field to
minimize energy consumption. Similarly Radio Frequency (RF) communication,
antenna design, and propagation modules effect performance of WBANs. Other
issues for researcher to be consider includes mobility of on, in or around human
body sensor nodes, transparency at MAC layer, interoperability, security and QoS.
A comparison of discussed MAC protocols is presented in Table III.
CDMA, FDMA, CSMA, and TDMA are multiple approaches for medium ac-
cess. However, each of them has some advantages and disadvantages. Collision free
communication is achieved by CDMA, but high computational and power require-
ments are major obstacles for implementation in WBANs where sensor nodes have
limited computational capabilities with constrained power. Hardware complexity
required for FDMA, to achieve collision free channel access, makes FDMA an
inappropriate solution for WBANs. CSMA based MAC protocols provide promis-
ing results such as low delay, reliable communication, and simple implementation
procedure in small dynamic networks. However, additional energy consumption
for collision detection or collision avoidance, and protocol overhead are major
shortcomings of CSMA. TDMA-based MAC protocols are contention free, nodes
transmit data in predefined time slots to avoid packet collision. For small networks
with low mobility and small number of sensor nodes and periodic data generation,
TDMA is the best approach for medium access. However, strict synchronization
requirement, non-adaptability and scalability are some issues faced by TDMA.
Based on topology and limited number of nodes in WBANs, TDMA could be
considered most suitable solution for medium access in WBANs.
Energyxefficiency is of utmost importance in WBANs. For high energy ef-
ficiency, a number of protocols have been proposed. However, MAC protocols
specifically for WBANs need to be developed. The aim of these protocols would
be to avoid energy dissipation due to collision, overhearing and idle listening with
reduced control packet overhead and implementation complexities. Fairness at
MAC layer, high bandwidth utilization, reliable communication, minimum delay,
and reduced synchronization cost are other objectives for multipurpose efficient
34
MAC protocol. The proposed protocol should have capabilities toxaccommodate
communication of normal,xemergency, andxon-demand traffic. However, selection
of MAC protocols is application and hardware dependent. This may be one of the
reasons that no proposed protocol is accepted as a standard for WBANs so far.
35
Chapter 4
Proposed MAC Protocol
36
4
Proposed MAC Protocol
4.1 Protocol design
Proposed MAC protocol, AR-MAC is based upon TDMA approach to min-
imize energy consumption. AR-MAC assigns Guaranteed Times Slot (GTS) to
each sensor node for communication based upon the requirements of sensor node .
To reduce overhearing and idle listening, proposed system uses periodic sleep and
wakeup according to node requirements. We assume a star topology; a Central
Node (CN) collects data from sensor nodes and communicates with a Monitoring
Station (MS), direct or through an Access Point (AP) as shown in the Figure 1.1.
Figure 4.1: WBAN Topology
CN is usually equipped with larger batteries and higher computational power.
One or two transceivers may be used within a single CN. In case of two transceivers
total time frame TF rame is allocated for communication with sensor nodes. We
assume CN with single transceiver where TF rame is divided into three parts: Con-
tention Free Period (CFP) for communication with sensors, Contention Access
Period (CAP) to accommodate emergency or on-demand traffic and time TM S
37
Start
Scans RF Channel
Wait For
Time TCP
Switch to another one Send ACK Send Channel Packet
Scans RF Channel
Start
Yes
Yes
Free
CP
Recieved Free
No No
No
Sensor Node Central Node
Yes
Figure 4.2: Channel Selection Procedure
for communicating sensor nodes’ data to MS. Following subsections describe AR-
MAC from its Channel Selection, Time Slots Assignment, Synchronization and
Frame Format.
4.1.1 Channel Selection
Initially, CN starts scanning for available free Radio Frequency (RF) channels.
If the current RF Channel is busy, CN switches to another RF Channel. CN selects
a free RF channel for communication. After successful selection of RF Channel,
CN broadcasts the Channel Packet with address and channel information to sensor
nodes. On the other side end nodes scan RF channels for Channel Packet from
CN. Sensor node scans the RF channel if it is free it switches to another RF
channel. If the channel is busy it waits for time TCP to listen Channel Packet. If
sensor node does not receive the Channel Packet, it switches again to next channel.
After successful reception of Channel Packet, node starts transmission and sends
an acknowledgment (ACK) packet to CN, as shown in Figure 2.2.
4.1.2 Time Slot Assignment
Once sensor node selects a proper RF Channel after receiving Channel Packet
from CN, sensor node sends out a Time Slot Request (TSR) packet to CN. TSR
packet includes sensor node’s data rate and required time slot information. Au-
thors in [6], propose time slots of fixed length with fixed guard band time. Biomed-
ical sensors in-body and on-body have different data rates and sampling intervals
with different clock drifts. Assigning time slots of equal length for unequal require-
ments is wastage of resources. AR-MAC uses an adaptive scheme for Time Slot
(TS) and GBT time. Based on traffic pattern of nodes, CN assigns time slot and
38
sends Time Slot Request Reply (TSRR). These time slots are of variable length
depend upon the requirements of sensor nodes. Assigned time slot can easily ac-
commodate the transmission of data packet, reception of ACK packet and some
acceptable delay, based on the communication model. A guard band time TGB
is inserted between the two successive time slots to avoid the interference due to
clock drift of node and CN as shown in Figure 2.3.
TS1 TSn------------------ CAP TMS
Guard Time Contention Access Period
Communication with MSGuaranteed Time Slot
Figure 4.3: Time Slots Assignment with Guard-band Time
Value of TGB depends upon length of the successive time slots. Adaptive guard
band avoids possibilities of collision and interference due to clock drift. We use
the following equation to calculate TGB .
TGB
n,n+1 =F
100 ×1
2[T Sn+T Sn+1] (4.1)
TGB
1=F×T S1
100 (4.2)
TGB
n=F×T Sn
100 (4.3)
where Fis guard band factor, depends upon the average drift value. However
guard band time TGB
1is inserted before first time slot and similarly guard band
time TGB
nis placed after Ntime slot. After successful time slot assignment, sensor
nodes enter into sleep mode and wakeup only to send data to CN in allocated
time slots. Periodic sleep reduces energy consumption due to idle listening and
Overhearing. Allocated time slots in CFP are completely collision free thus reduce
the energy consumption and make the communication reliable.
39
4.1.3 Synchronization
TDMA schemes require extra energy cost for periodic synchronization [8]. Syn-
chronization of nodes after N number of cycles is energy consuming process. AR-
MAC uses a novel synchronization mechanism to avoid collision and energy con-
sumption. After successful assignment of time slots to nodes, CN listens to data
packet within expected time slot. Upon arrival of data packet, CN compares cur-
rent arrival time of the packet and expected arrival time with acceptable delay
(D). Based on the difference of current arrival time and expected arrival time a
Drift Value (DV ) is calculated. This DV is transmitted to node within ACKnowl-
edgment (ACK) packet to adjust time slot for future communication. However,
this value depends upon acceptable delay and F. If the difference between the
expected arrival time and current arrival is greater than D, CN sends DV with
in SYNChronization ACKnowledgment (SYNC-ACK) packet for future synchro-
nization to sensor nodes otherwise, CN sends simple ACK packet for received data
packet. For communication of data in future, sensor node adjusts its wakeup time
schedule according to DV . Using this scheme of synchronization a node can go into
sleep mode without loosing synchronization for N number of cycles. Acceptable
delay Dis linked with guard band factor Fas under:
D=Min(T S1......T Sn)×F
100 (4.4)
For future synchronization, decision of sending DV to end nodes is based upon
the difference of current arrival time and expected arrival time of data packet.
△Trepresents this difference.
△T=ExpectedAr rivalT ime −CurrentArriv alT ime (4.5)
DV ={0 if | △T|< D
△Tif | △T|> D (4.6)
40
4.1.4 Frame Formate
Proposed AR-MAC uses two types of packets: Data Packets and Control Pack-
ets. In data packet sensor node sends its periodic data in allocated time slot. For
emergency data, node uses CAP. Control Packets are:
1. Channel Packet: After channel selection Central Node advertises channel
information and its unique address in Channel Packet.
2. Time Slot Request (TSR) Packet: Sensor node TSR packet to Central Node
for Guaranteed Time Slot (GTS) assignment for data communication
3. Time Slot Request Reply (TSRR) Packet: Central node sends Guaranteed
Time Slot information with CAP information to node in Time Slot Request
Reply packet
4. Synchronization-Acknowledgment (SYNC-ACK) Packet: For synchroniza-
tion, Central Node sends the required Drift Value to end node with ACK
of previously received data packet in Synchronization Packet to compensate
the clock drift and maintain the synchronization
5. Data Request (DR) Packet: For on demand traffic/information, Central
Node sends Data Request Packet to end nodes
6. Acknowledgment (ACK) Packet: Each data packet is acknowledged using
Acknowledgment Packet
Preamble Sync Frame Len MPDU
Control Address Other Payload (Variable ) CRC
Packet Type ACK ODT
Figure 4.4: MAC Layer Frame Formate
The MAC Protocol Data Unit (MPDU) stars with 3 octets of overhead. First
octet carries information about packet type. Preceding two octets carry address
41
information. In case of emergency traffic sensor node waits for CAP. Upon suc-
cessful Clear Channel Assessment (CCA), sensor node starts communication with
CN. However, in case of on-demand traffic CN sets On-Demand Traffic (ODT)
field in ACK or SYNC-ACK packet to 1. After receiving this packet sensor node
wakes up in CAP to listen data request from CN. After receiving data request
from CN, sensor node sends requested data packets and waits for ACK packet.
Sensor node enters into sleep mode after successful reception of ACK.
4.2 Energy Consumption Analysis
In order to model energy consumption, we consider the energy consumption
related to transceiver. In this study, we assume energy consumption of sensing and
processing units to be constant. We assume periodic traffic pattern, i.e., sensor
nodes send periodic data to CN in assigned time slots. Most of the time sensor
nodes remain in sleep mode. During allocated time slots, they wakeup to send
data. We use the following equation to measure the energy consumption for N
number of cycles.
ET otal =
N
∑
k=1
ESleepk+
N
∑
k=1
EActivek(4.7)
Energy consumption is a function of time and current drawing from voltage
source for a specific task. When nodes enter into sleep mode they still consume
energy. The sleep mode duration can be calculated from total time frame length
and time for which the node is in active mode.
TSleep =TF r ame −TActive (4.8)
ESleep =TS leep ×ISleep ×V(4.9)
ISleep is the current drawing from voltage source Vduring sleep mode. In
TActive the nodes receive, transmit and wait for Acknowledgment. Energy is also
consumed in switching, from sleep to active and active to sleep mode. The energy
consumed for all these tasks will be considered as energy in TActive .
42
EActive = 2 ×ESw +ET r ans +ERec +ET O ut (4.10)
where, ESw is Switching energy, ET rans Transmission energy, ERec is Receiving
energy and ET Out is Time-Out energy. We describe these terms in details in the
following subsections.
4.2.1 Switching Energy
Most of the time, sensor nodes remain in sleep mode. Sensor nodes turn
on its transceiver in wakeup mode for communication. Switching energy is the
consumed energy for switching transceiver between states; sleep mode and wakeup
mode. Frequently switching of transceiver between states leads to high energy
consumption. Energy consumed for switching the transceiver is determined by
the following equation.
ESw =TSwitch ×ISwitch ×V(4.11)
where TSwitch is the required for the transceiver to switch between sleep and wakup
mode and ISwitch is the required current.
4.2.2 Transmission Energy
Transmission energy is the energy consumed for transmission of Data or Con-
trol packet of length P. Following equation links the transmission energy with
length of packet P, time required for transmission of single byte TByte, current
draw during transmission IT rans and a voltage source V.
ET rans =P×TB yte ×IT rans ×V(4.12)
4.2.3 Receiving Energy
Receiving Energy is the consumed energy while receiving packets and their
associated overhead. Receiving energy is expressed as:
43
ERec =P×TByte ×IRec ×V(4.13)
where, IRec is the Current during reception, TByte is the time for single byte, Pis
the length of packet and Vis the voltage source.
4.2.4 Time-Out Energy
The energy consumed after transmission and before reception of an ACK
packet is termed as Time-Out energy. For time TT Out, current IT Out and voltage
source V, we used the equation given below to calculate the energy consumption
during Time-Out.
ET−Out =TT Out ×IT Out ×V(4.14)
4.3 Simulation Results
We use MATLAB to measure and compare the energy efficiency of the AR-
MAC with that of IEEE 802.15.4. In energy consumption comparison, we consider
the energy consumption of RF transceiver. We use the energy consumption model
from Crossbow MICAz data sheet as shown in Table II. Packets are dropped
randomly with average Packet Error Rate probability from 1% to 20%. Time
frame size used in simulations is TF rame = 1 Second. We used packets format as
shown in Fig.4. Simulation has been carried out for 10 Sensor Nodes.
Table 4.1: Simulation Parameters Value
Parameter Value
Time frame( TF rame) 1 Second
Voltage Source 3 volts
Current Draw in Receive Mode 19.7 mA
Current Draw in Transmit Mode 17.4 mA
Current Draw in Idle Mode 20.0 mA
Current Draw in Sleep Mode 1 micro-A
Number of Sensor Nodes 10
Number of Cycles N1000
44
2 4 6 8 10 12 14 16 18 20
9.25
9.3
9.35
9.4
9.45
9.5
9.55
9.6
Packet Error Rate[%]
Energy Consumption [Joule]
802.15.4
AR−MAC
Figure 4.5: Energy consumption of AR-MAC and IEEE 802.15.4 for N= 1000
We used Eq. 7 to calculate the energy consumption for N= 1000. Figure 5
shows the energy comparison of the AR-MAC with IEEE 802.15.4. The graph in
Figure 5 shows that energy consumption of IEEE 802.15.4 increases with increase
in probability of Packet Error Rate. This increase in energy consumption is due to
extra energy requirement of CSMA/CA operation in IEEE 802.15.4. The energy
consumption of AR-MAC increases with a minor variation due its adaptive time
allocation and adaptive guard band mechanism. AR-MAC assignees guaranteed
time slots to sensor nodes for communication, to overcome the packet collision
and overhearing.
45
Chapter 5
Conclusion
46
5
Conclusion
5.1 Conclusion
Aim of this research work is to analyze the existing MAC protocols for WBANs
with emphasis on energy minimization. These protocols are being developed to
prolong the lifespan of WBANs, reliable communication, flexibility, fair manage-
ment, and QoS. However, MAC protocols based on random access and LPL are
unable to accommodate the emergency and on-demand traffic. On the other hand,
TDMA is a vital approach for medium access to be used in WBANs. Majority of
the existing MAC protocols based on TDMA approach. Each of them has some
advantages and disadvantages discussed above. Due to diverse application require-
ment and hardware constrains, no one protocol is being accepted as a standard.
A new protocol needs to be developed to achieve requirements of WBANs like
energy efficiency, scalability, fairness, reduced implementation complexity, sup-
port for divers application, interoperability, reduced synchronization overhead,
and QoS.
In this thesis we proposed AR-MAC, a new MAC protocol for WBANs. AR-
MAC assigns guaranteed time slots in adaptive manner and makes the communica-
tion reliable by introducing adaptive guard band to avoid collisions. Synchroniza-
tion is achieved by introducing a smart mechanism to compensate drift of sensor
node’s clock. By simulation, we compared the performance of AR-MAC with that
of IEEE 802.15.4 in terms of energy consumption. Simulation results show supe-
rior performance. Future work will be carried out to implement proposed protocol
including all Control and Data Packets in real time scenario.
47
References
48
References
[1] IEEE Std.802.15.4, Wireless medium accesses control (MAC) and physical
layer (PHY) specif ications for low data rate wireless personal area networks
(WPAN). IEEE: Piscataway, USA, 2006
[2] Changle, L., Huan-Bang, L., and Kohno, R., “Performance evaluation of IEEE
802.15.4 for Wireless Body Area Network (WBAN),” in the Proc. of IEEE In-
ternational Conference on Communications Workshops (ICC Workshops 2009)
pp. 1-5, 2009.
[3] Su, H., and Zhang, X., “Battery-dynamics driven TDMA MAC protocols for
wireless body-area monitoring networks in healthcare applications,” IEEE J.
Sel. Areas Commun. 27(4):424-434, 2009.
[4] Zhang, Y., and Dolmans, G., “A new priority-guaranteed MAC protocol for
emerging body area networks,” in the Proc. of Fifth International Conference
on Wireless and Mobile Communications (ICWMC 2009), pp.140-145, 2009.
[5] Marinkovi, S. J., Popovici, E. M., Spagnol, C., Faul, S., and Marnane, W.
P., “Energy-efficient low duty cycle MAC protocol for wireless body area net-
works,” IEEE Trans. Inf. Technol. Biomed. 13(6):915925, 2009.
[6] Kwak, K. S., Ullah, S., Kwak, D. H., Lee, C. H., and Lee., H. S., “A power-
efficient MAC protocol for WBAN,” Journal of Korean Institute of Intelligent
Transport Systems 8(6):131140, 2009.
[7] A. El-Hoiydi et al., “WiseMAC: An Ultra Low Power MAC Protocol for the
WiseNET Wireless Sensor Network,” Proc. 1st ACM SenSys Conf., ACM
Press, 2003, pp. 302-303.
[8] O. Omeni, A. C. W. Wong, A. J. Burdett, and C. Toumazou, “Energy efficient
medium access protocol for wireless medical body area sensor networks,” IEEE
Transactions on Biomedical Circuits and Systems. vol. 2, no. 4, pp. 251-259,
Dec. 2008.
49
[9] G. Fang and E. Dutkiewicz, “BodyMAC: Energy efficient TDMA-based MAC
protocol for wireless body area networks,” 9th International Symposium on
Communications and Information Technology, ISCIT, pp. 1455-1459, 2009.
[10] N.F. Timmons and W.G. Scanlon, “An adaptive energy efficient MAC pro-
tocol for the medical body area networks,” 1st International Conference on
Wireless communication VITAE, pp. 587-593, 2009.
[11] Huaming L., and Jindong T., “Heartbeat-Driven Medium-Access Control
for Body Sensor Networks,” IEEE TRANSACTIONS ON INFORMATION
TECHNOLOGY IN BIOMEDICINE, VOL. 14, NO. 1, JANUARY 2010
[12] H. Li and J. Tan, “Body sensor network based context aware QRS detection,”
in Proc. 28th IEEE EMBS Annu. Int. Conf., Aug. 2006, pp. 32663269.
[13] M. Aboy, J. McNames, T. Thong, D. Tsunami, M. S. Ellenby, and B. Gold-
stein,“ An automatic beat detection algorithm for pressure signals,” IEEE
Trans. Biomed. Eng., vol. 52, no. 10, pp. 16621670, Oct. 2005.
[14] L. van Hoesel and P. Havinga, “A lightweight medium access protocol
(LMAC) for wireless sensor networks,” presented at the 1st Int. Workshop
Netw. Sens. Syst. (INSS 2004), Tokyo, Japan, 2004.
[15] W. Ye, J. Heidemann, and D. Estrin, “An energy-efficientmac protocol for
wireless sensor networks,” in Proc. IEEE Infocom, 2002, pp. 15671576.
[16] W. L. Lee, A. Datta, andR.Cardell-Oliver, “FlexiMAC:Aflexible
TDMAbasedMAC Protocol for fault-tolerant and energy-efficientwireless
sensor networks,” in Proc. 14th IEEE Int. Conf.Netw. (ICON 2006), Sep., pp.
16.
50
